Radian Radian Corporation
8500 SHOALCREEK BLVD. • P. O. BOX 9948 • AUSTIN, TEXAS 78758 • TELEPHONE 512 - 454-9535
FINAL REPORT
VOLUME I
APCO Contract No. CPA 70-45
A STUDY OF THE LIMESTONE INJECTION
WET SCRUBBING PROCESS
C. .JMICAL RESEARCH • SYSTEMS ANALYSIS • COMPUTER SCIENCE • CHEMICAL ENGINEERING
-------
Radian Corporation
8500 SHOALCREEK BLVD. . P. O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
FINAL REPORT
VOLUME I
APCO Contract No. CPA 70-45
A STUDY OF THE LIMESTONE INJECTION.
WET SCRUBBING PROCESS
Presented to:
AIR POLLUTION CONTROL OFFICE
DEPARTMENT OF HEALTH, EDUCATION AND WELFARE
411 West Chapel Hill Street
Durham, North Carolina 27701
1 November 1971
Prepared by:
tW~~
Philip S. Lowell Delbert M. Ottmers, Jr.
principal. Scientist c9 tu...... j. ~a:'::':.r Senior Engineer
~.A/? '£ U Thomas I. Strange JVlA' . ! /Q I .
1(1, /' ;/7(/(.£/ Senior Computer Scientist (,VY'V7 /' (jl A;.,1. r
Klaus Schwitzgebel James L. Phillips
Senior Scientist Engineer Scientist
Lk/J:;erf f», ol!m erS Jy
c. ,.:MICAL RESEARCH. SYSTEMS ANALYSIS. COMPUTER SCIENCE. CHEMICAL ENGINEERING
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-9535
ABSTRACT
A computerized process model was developed to simulate
the limestone wet scrubbing system being designed for the TVA
Shawnee power plant. A parameter study was performed using
this model to quantitatively estimate the effect of process
variables on overall process performance. The effects of
limestone composition, sulfite oxidation, and limestone dissolu-
tion rates were found to be significant.
Radian provided technical assistance to APCO during
pilot-scale experiments being conducted at their Cincinnati
laboratory facilities. This system employs a venturi scrubber
to absorb sulfur dioxide. Radian assistance included test
plan design, development of sampling and analytical techniques,
chemical analysis of scrubber liquor samples, and interpreta-
tion of experimental data. The set of experiments which were
designed to evaluate the importance of vapor-liquid mass
transfer rates were performed and their results analyzed
this contract period. The measured SOs removals did not
approach the vapor-liquid equilibrium amount.
during
closely
A computational technique
the enthalpy of process streams in
systems.
was developed for calculating
limestone wet scrubbing
This program was conducted under
the Air Pollution Control Office, Process
Division, Contract No. CPA 70-45.
the sponsorship of
Control Engineering
-------
Radian Corporation
1.0
2.0
2.1
2.1.1
2.1.2
2.1.3
2.1.3.1
2.1.3.2
2.1.3.3
2.2
2.2.1
2.2.2
2.2.3
2.2.4
2.2.5
2.2.6
2.2.7
2.2.8
2.2.9
2.2.10
2.3
3.0
3.1
3.1.1
3.1.2
3.1.3 .
3.2
3.2.1
3.2.2
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
TABLE OF CONTENTS
INTRODUCTION. . . .
. . . . . . . . . . . . . .
PROCESS SIMULATIONS
. . . . .
. . . .
. . . . .
Simulation Basis. . . . . . . . . . . . . . . .
Process Description. . . . . . . . . . . . . .
Model Assumptions and Parameters. . . . . . . .
Computer Program. . . . . . . . . . . . . . . .
Executive System. . . . . . . . . . . . . . . .
Equipment Subroutines. . . . . . . . . . . . .
Convergence. . . . . . . . . . . . . . . . . .
Simulation Results. . . . . . . . . . . . . . .
Sulfite Oxidation. . . . . . . . . . . . . . .
Ionic Strength. . . . . . . . . . . . . . . . .
Limestone Amount and Composition. . . . . . . .
Scrubber Feed Rate. . . . . . . . . . . . . . .
Circulating Liquor Temperature. . . . . . . . .
Solids Precipitation in Scrubber. . . . . . . .
Limestone Reactivity. . . . . . . . . . . . . .
Fraction Solids in FB . . . . . . . . . . . . .
Energy Balance Considerations. . . . . . . . .
Process Flow Diagrams. . . . . . . . . . . . .
Summary and Conclusions. . . . . . . . . . . .
APCO VENTURI TESTS. . . . . .
. . . .
. . . . .
Test Plan and Objectives. . . . . . . . . . . .
Type I Experiments. . . . . . . . . . . . . . .
Type II Experiments. . . . . . . . . . . . . .
. Type III Experiments. . . . . . . . . . . . . .
Analytical Chemistry. . . . . . . . . . . . . .
Requirements. . . . . . . . . . . . . . . . . .
Method s . . . . . . .
. . . . .
. . . . .
. . .
Page
1
3
3
6
10
14
15
17
19
20
24
28
30
33
35
38
40
42
44
47
50
53
53
56
62
64
65
65
68
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X '/948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
TABLE OF CONTENTS (continued)
Page
3.2.2.1
3.2.2.2
3.2.2.3
3.2.2.4
3.3
3.3.1
3.3.2
3.3.2.1
3.3.2.2
3.3.2.3
3.3.2.4
3.3.2.5
3.3.2.6
3.3.2.7
3.3.2.8
3.3.3
3.3.3.1
3.3.3.2
3.3.3.3
3.3.3.4
3.3.4
3.3.4.1
3.3.4.2
3.3.4.3
3.3.3.4
3.4
4.0
4.1
4.2
4.3
Liquid and Solid Sampling from Slurries. . . . . 68
Sulfite Determination. . . . . . . . . . . . . . 72
Sulfate Determination. . . . . . . . . . . . . . 74
Total Nitrogen Determination. . . . . . . . . . 75
Experimental Results. . . . . . . . . . . . . . 79
Experimental Equipment and Procedures. . . . . . 79
Treatment of Data. . . . . . . . . . . . . . . . 83
Water Evaporation in the Scrubber. . . . . . . . 83
Flue Gas Sampling Correction. . . . . . . . . . 89
Total Sulfur Material Balances. . . . . . . . . 90
Total Nitrogen Material Blanace. . . . . . . . . 91
Extent of SOs Oxidation. . . . . . . . . . . . . 91
Equilibrium Partial Pressure Calculation. . . . 92
Number of Overall Gas-Phase Transfer Units. . . 95
Re 1 at i v e Kg a . . . . . . . . . . . . . . . . . . 96
Experimental Error Propagation. . . . . . . . . 97
Material Balance Error. . . . . . . . . . . . .100
Number of Transfer Units. . . . . . . . . . . .101
Re lati ve Kga . . . . . . . . . . . . . . . . . .103
Limits of Error for Oxidation. . . . . . . . . .103
Experimental Results and Discussion. . . . . . .104
SO; Absorption. . . . . . . . . . . . . . . . .106
NOx Absorption. . . . . . . . . . . . . . . . .110
CO; Absorption. . . . . . . . . . . . . . . . .111
Sulfite Oxidation Rate. . . . . . . . . . . . .111
General Comments on Experimental Results. . . .112
ENERGY CALCULATIONS. . . . . . . . . . . . . . .113
Ob j ec ti ve s . . . . . . . . . . . . . . . . . . . 113
Theory. . . . . . . . . . . . . . . . . . . . .113
Comparison of Experimental and Calculated
Re suI t s. . . . . . . . . . . . . . . . . . . . . 116
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
TABLE OF CONTENTS (continued)
Page
5.0
SUMMARY AND CONCLUSIONS.
. . . . . . . . . . . 122
6.0
6.1
6.2
6.3
NOMENCLATURE. . . . . . . . . . . . . . . . . 129
Section 2 Nomenclature. . . . . . . . . . . . 129
Section 3 Nomenclature. . . . . . . . . . . . 131
Section 4 Nomenclature. . . . . . . . . . . . 132
7.0
BIBLIOGRAPHY. . . . .
. . . . . . . . . . . . 133
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
1.0
INTRODUCTION
The Air Pollution Control Office (APCO) has been
active in developing several methods for controlling S02
emissions. Limestone based scrubbing processes are presently
the most promising (from a technical feasibility point of view)
for control of fossil fuel power plant emissions. APCO involve-
ment in limestone scrubbing processes has covered areas such
as basic chemistry, feasibility,. and pilot plant studies. A
focal point of APCO effort is the prototype units that will
b'e installed at TV A , s Shawnee generating plant in Paducah,
Kentucky.
Radian Corporation, under previous contract to APCO
(CPA 22-69-138) has developed a theoretical interpretation of
the complex chemistry and chemical engineering aspects of the
limestone injection - wet scrubbing processes. The results of
this work have now been put in the form of a computerized pro-
cess model. This model quantitatively estimates the effect of
a number of important process variables on overall process
performance. The work described in the present report illus-
trates two means of using such a theoretical 'model to gain
valuable process insight. It was intended to (1) provide interim
results based on up-to-date information that will be of value to
organizations involved in design or investigation of lime/lime-
stone wet scrubbing processes and (2) provide tested procedures
for obtaining and interpreting data from experimental units
such as the prototype facility at Shawnee.
In fulfillment of these first two objectives this
report covers the following areas.
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
Process Simulations - In Section 2.0
a series of process simulations is
described. These results demonstrate
how the chemistry and performance of
the limestone injection - wet scrubbing
process vary with important parameters.
Experimental Studies Section 3.0
gives the results of a pilot scale test
series conducted at APCO's Cincinnati
laboratory facilities. Analytical
methods and data analysis techniques
were developed and demonstrated.
In addition, a computational technique was developed
for making enthalpy balances in limestone scrubbing liquors.
This represents a significant improvement to the process model
and will be useful in process design calculations.
-2-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
2.0
PROCESS SIMULATIONS
Radian developed the theoretical basis for simulating
the limestone injection - wet scrubbing (LIWS) process under
APCO Contract No. CPA 22-69-138. Using the Radian developed
chemical equilibrium program and several process assumptions,
preliminary simulations were performed for three processing
schemes at "typical" operating conditions. Although limited in
scope, these initial simulations showed promise in estimating
(1) the technical feasibility of various processing schemes
and (2) the compositions and flow rates for the process units.
The objectives of this task area in the present
contract were twofold: (1) develop a computerized process model
to carry out simulations of the prototype system being designed
for the TVA Shawnee power plant and (2) simulate the prototype
system for a number of processing conditions. In addition to
better defining the technical feasibility of several processing
alternatives, these simulations are useful in predicting the
relative importance of model parameters. Information of this
nature is valuable in process design and development, particu-
larly in the areas of design of testing programs, correlation
of experimental data, and optimization of processing conditions.'
2.1
Simulation Basis
The limestone injection - wet scrubbing (LIWS)
process involves injecting limestone into the power plant boiler
and catching it in a wet scrubber after the air heater (see
Figure 2-1). Injection of limestone into the boiler removes a
portion of the sulfur dioxide ahead of the scrubber, provides
protection from corrosion by sulfur trioxide and alkali salts,
and converts the limestone into quicklime; The major components
-3-
-------
c
LIMESTONE
COAL
MILL
I
,J::-o
I
AiR
FIGURE 2-1
FURNACE
STACK GAS
REHEATER
AIR
HEATER
COOLER
SCRUBBER
EFFLUENT
HOLD
TANK
TEE
SCHEMATIC OF LIMESTONE INJECTION - WET SCRUBBING PROCESS
WATER
MAKE -UP
RECYCLE
HOLD
TANK
CLARIFIER
FI L TER
SOLI DS WASTE
~
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 4S4-9S35
of the LIWS process are the scrubber and the liquid slurry
handling system.
The overall reaction in the scrubber is gaseous SOg
and COg reacting with solid CaO to give solid CaSOs (perhaps
hydrated) and CaCOs. Oxidation will give' rise to sulfates.
The actual reaction path is probably rather complex and is not
presently known. Most probably the gases dissolve into the
scrubber liquid. Many ionic and nonionic reactions then take
place in this liquid. The liquid also reacts through some
mechanism with the solids to form new solids. Resistances to
mass transfer are represented by both sides of the vapor-liquid
film, liquid-phase reactions, the liquid-solid film, diffusion
through a solid layer, and reaction with the solid.
Oxidation of su1fites to sulfates will influence
both the vapor-liquid equilibrium and liquid-solid equilibrium.
The ionic strength of the circulating solution will be increased
by soluble components in the fly ash and limestone (e.g., Na, K,
C1) as well as other flue gas components (e.g., nitrogen oxides).
and the make-up water. Provision was made to include these
components in the process model.
Radian developed a computerized process model for
the limestone injection - wet scrubbing (LIWS) process that is
based upon (1) the ability to predict vapor-liquid-solid
equilibria for the CaO-MgO-NagO-SOg-COg-SOs-NgOs-HC1-HgO system
and (2) a number of process assumptions with regard to equip-
ment characterization and the extent of various chemical
reactions.
Much of the theoretical basis for this process
model was developed under APCO Contract CPA 22-69-138 and is
-5-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
described in the Radian Final Report "A Theoretical Description
of the Limestone Injection - Wet Scrubbing Process". Under the
present contract, a computerized process model was developed
which (1) employs a modular approach to performing process
material aOd energy balances, (2) specifies the extent of var-
ious reaction steps by appropriate equilibrium assumptions or
model inputs, (3) characterizes the various process vessels in
terms of practical equipment descriptions, and (4) converges the
model's. iterative parameters with suitable numerical techniques.
This section presents a description of the Radian
process model. A brief description of the LIWS process with
regard to its relationship to the process model will be followed
by listing of the model assumptions and parameters. In addition,
a brief description of the computer program will be given.
2.1.1
Process Description
The wet scrubbing scheme that was simulated is shown
in Figure 2-2. This flow arrangement is one of the major operat-
ing schemes that has been proposed for the LIWS prototype system
being planned for TVA's Shawnee Plant at Paducah, Kentucky.
The scrubber serves as a gas-liquid contacting
device with some dissolution of limestone solids. In the
present model, the scrubber (S) is assumed to trap all of the
solids in the flue gas stream so that no solids leave with the
gas. For material balance calculations the inlet gas stream
to the scrubber has been divided into two fractions, a flue
gas stream (FG) containing only gas~ous components and a lime-
stone-fly ash stream (LA) containing only solid components.
In this model the stack gas (SG) leaving the scrubber is assumed
to be in equilibrium with the scrubber liquor (and thus with the
scrubber bottoms).
-6-
-------
GAS SPECIES
FG, SG
STACK GAS
SG
WATER
MAKEUP
WM
,. S02 '.
2. C02
3. NOx
4. H20 ,
5. 02
6. CO SCRUBBER ~CRUBBER FEED PROCESS
7. N2 WATER
S SF HOLD TANK
P
j
SCRUBBER
FLUE GAS - BOTTOMS
FG SB SLURRY RECYCLE SR -
.
.
CLARIFIER,
LIQUID
. CV
LIMESTONE
FLY ASH SCRUBBER CLARIFIER
SOLI OS EFFLUENT FEED CLARIFIER -
LA HOLD TANK CF C
I. CoO E
2. MgO CLARIFIER
3. CoS04 BOTTOMS
4. MgS04 CB
5. COS03
6. MgS03 FILTER
7. Co C03 -
F FILTER
8. MgC03 LIQUID
9. FLY ASH FL
10. SOLUBLE No FILTER
II. SOLUBLE CI BOTTOMS
FB
PROCESS SOLID SPECIES
(CF) SRI CB) FB) SF)
I. Co(OH)2 5. Mg (OH)2
2. COC03 6. MgC03
3. Co 503 7. FLY ASH
4. Co S04
FIGURE 2-2
WET SCRUBBING SCHEME
-7-
PROCESS
LIQUID
SPECIES
SB, CF) SR, CB)
FB, CLlL, SF
I. H+
2.0H-
3. HS03
4. S03
5. S04
6. HC03
7. C03
8. HS04'
9. H2S03
10. H2C03
II. Co++
12. Co OH+
13. COS03
14. COC03
15. CoHCO;
16. COS04
17. CoNOt
18. N03
19. Mg ++
20. MgOH +
21 . MgS04
22. MgHCO;
23. MgS03
24. MgC03
25. No+
26. NoOH
27. NoC03
28. NoHC03
29. NoS04
30. NoN03
31. CI-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 994B . AUSTIN, TEXAS 7875B . TELEPHONE 512 - 454-9535
The scrubber effluent hold tank (E) serves mainly
as a solids dissolution and precipitation vessel. This stirred
tank should approximate idealized backmix flow and in practice
should be designed to provide enough residence time and
agitation to hydrate and dissolve the major portion of the CaO
and MgO that dissolves in the system. In this model some CaO
and MgO entering the system is taken to be unreactive. This
amount of CaO and MgO leaves the system unreacted in the filter
bottoms (FB) stream. The amount of "unreacted" CaO and MgO
circulating in the system is determined by various process
parameters, i.e., the amount of slurry recycle (SR), the weight
percent solids in SR, and the composition of the solid phase
in FB. All o.f the remaining CaO and MgO which enters the
effluent hold tank (E) is assumed to be available for reactions.
The liquid leaving the effluent hold tank (E) is assumed to be
in equilibrium with respect to the liquid and solid phases.
This implies that several species will dissolve or precipitate.
A major portion of the precipitation in an actual process will
probably occur in the effluent hold tank.
The clarifier and filter vessels serve mainly as
solid-liquid separators. In the model presented, the clarifier
liquid (CL) was assumed to be in equilibrium with the solids
in the clarifier. The liquid in the clarifier bottoms (CB) was
assumed to have the same composition as the overflow (CL) stream.
This model also assumes that the compositions of the solid and
liquid phases in streams leaving the filter (FB and FL) are the
same as the compositions of the solid and liquid in equilibrium
in the clarifier and filter. This situation would exist if
the clarifier and filter were operated at the same temperature,
or if the solid-liquid "shift" in the filter were insignificant.
The CL and FL streams can contain a small amount of solids due
to clarifier and filter inefficiencies which are specified as
model inputs.
-8-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
Although some solid dissolution and precipitation
could take place in the clarifier due to its long retention
time, the stagnant nature of the liquid in this vessel should
retard significant solid-liquid transfer.
The process water hold tank (p) is a stirred tank
which serves as a liquid mixer. Here the clarifier and filter
liquids (CL ~nd FL) and slurry recycle (SR) are diluted with
water make-up (WM) to produce the scrubber feed (SF) for the
scrubber. In this model, this tank is assumed to be an ideally
backmixed vessel that has attained solid-liquid equilibrium.
Figure 2-2 also shows the chemical species possibly
present in various process streams. Seven gaseous species are
listed for the flue and stack gases. The nitrogen oxides in the
gas phase are designated NOx since not enough information con-
cerning (1) NOx composition in the gas phase, (2) oxidation
rates, and (3) chemical reactions in liquid phase has been found
in the literature. Due to this uncertainty, a realistic frac-
tion of the NOx in the flue gas will be taken as absorbing and
this NOx will show up in the process liquids as "equivalent
nitrates". A refinement of this approximation can be made when
enough information concerning the NOx system becomes available.
The limestone-fly ash stream (LA) could contain
eleven solid species, CaO and MgO being the significant ones. The
fly ash component is not specified in detail because fly ash
compositions vary to some extent and because the soluble nature
of the fly ash is not fully known at this point. For the
purposes of this model, the fly ash will be assumed to contain
a certain fraction of soluble Na+, ~, and Cl-.
-9-
-------
Radian Corporation
2.1.2
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
Model Assumptions and Parameters
assumptions:
The present Radian model is based upon the following
(1)
(2)
(3)
(4)
The partial pressures of S02' CO2 and H20 in
the stack gas leaving the scrubber are in
equilibrium with the scrubber liquid at the
scrubber temperature (vapor-liquid equilibrium
is achieved and the gas leaving the scrubber
is last in contact with liquid having the
composition of the scrubber exit liquor).
After withholding a portion of the input
lime (CaO and MgO) as being chemically
unreacted in the system, solid-liquid
equilibrium is achieved in the scrubber-
effluent and process-water hold tanks.
The clarifier and filter are solid-liquid
separators only (no chemical change occurs).
Ionic reactions taking place in the liquid
phase are rapid and thus at equilibrium.
The first assumption is based on the premise that the gas-liquid
absorption step is not rate controlling. If v-~ equilibrium is
approached, this assumption would apply to countercurrent contactors
whereby the liquid-phase is ideally backmixed or to concurrent con-
tactors. The turbulent contact ~bsorbed (TCA) and the marble bed
(Hydro-filter) scrubber are examples of such countercurrent contactors.
The venturi scrubber is an example of a concurrent contactor. Each
of these three contactors are to be tested at Shawnee.
The second assumption is that the hold tanks are large
enough to permit solid-liquid equilibrium except for a portion of
unreactive lime. This unreacted lime (CaO and MgO) is fixed by
model input.
-10-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
The compositions for the system inlet and exit streams
[limestone-fly ash (LS+FA = LA), flue gas (FG), stack gas (SG),
and water make up (WM)] and operating conditions for the process
are specified by model parameters. Other model parameters are
used to set the "conversion" in the process due to various rate
steps. Examples of these are the fraction of precipitating
solids, limestone in the flue gas entering the scrubber and
dissolving, sulfite oxidizing, and NOx absorbed in the scrubber.
A list of the model parameters along with their
typical units is given below:
1.
Theoretical Limestone
[moles CaO + MgO in LS]
moles SOg in FG
2 .
Lime Hydrating in System
[1 - moles CaO+MgO in FBJ
moles CaO+MgO in LS
3.
Lime Hydrating in Scrubber
[moles CaO+MgO from LS hydrating]
moles CaO+MgO in LS
4.
LS left in SB]
rom LS
5 .
6.
Limestone Composition
[ moles MgO in LS ]
moles OaO+MgO in LS
-11-
-------
Radian Corporation
10.
11.
12.
13.
14.
15.
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
7 .
Fly Ash Level
[ wt. fly ash ]
wt. sulfur in coal
8.
Soluble Sodium
[wt. fro Na in FA]
9.
Soluble Chlorine
[wt. fro Cl in FA]
SO~ Level in Flue Gas
[mole fro So~J
SO~ Absorbed
[moles SO~ abSorbed]
mole SO~ in FG
Sulfite Oxidized
[moles sulfite oXidized]
mole SO~ absorbed
CO; Level in Flue Gas
[mole fro COg]
NO Level in Flue Gas [mole fro NO]
NOg Level in Flue Gas [mole fro NO~]
NO Ab b d [moles NO absorbed]
sor e mole NO in FG
NO Ab b d [moles NO~ absorbed]
g sor e mole NOg in FG
-12-
-------
Radian Corporation
16.
17 .
18.
19.
20.
21.
22.
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-9535
Solids in Filter Bottoms
[wt. fro solids in FBJ
Solids in Clarifier Bottoms
[wt. fro solids in CBJ
Filter Efficiency
[ wt. solids in FB l
wt. solids in filter inlet]
Clarifier Efficiency
[ wt. solids in CB inlet]
wt. solids in clarifier
Solids Precipitated in Scrubber
[yes or no]
Circulating Liquor Temperature
[OF]
o
[Gal. of SF ]
Scrubber Feed Rate 1000 ACF of FG .
Usually, the temperature of the scrubber liquor is determined
by an energy balance about the scrubber. (Here, the humidity
and temperature of the flue gas are the dominant factors.)
Then the temperature of the circulating liquor is set equal to
that of the scrubber liquor.
-13-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE 512 - 454-9535
Process simulations are conducted by specifying the
amount of sag to be removed from a flue gas of known .composition
and adjusting the amount of slurry recycle (SR) (or the amount
of scrubber feed (SF) in some cases) to obtain this S09 removal.
The fractions of limestone solids (LS) reacting in the scrubber
are specified as a model input. In the same way, the fractions
of solid species in the scrubber feed (SF) stream that are
available for reaction are also model inputs.
All of the liquid-phase species are allowed to react
in the scrubber. Thus, for a fixed scrubber feed (SF) rate, the
amount of slurry recycle (SR) required to obtain the desired
sag removal is a measure of the scrubbing capability of the
process. High slurry recycle rates indicate a poor set of scrub-
bing parameters.
2.1.3
Computer Program
The computerized process model consists of essentially
three parts: (1) an executive system that interconnects the
processing units in appropriate fashion and controls the sequen-
cing of computer operations, (2) equipment subroutines that model
each process unit, and (3) convergence subroutines that force
convergence of the model's iterative parameters.
The design of any on~ of these parts is dependent
upon the design of the other parts. For example, a generalized
scheme for performing material and energy balance calculations
is attractive from the standpoint of model flexibility. However,
this approach suffers from the need of greater programming
sophistication and from less specific (and thus slower) convergence
routines.
-14-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN. TEXAS 78758 . TELEPHONE 512 . 454-9535
The computerized model used was based upon a
compromise between generalization for greater system flexibility
and specialization for improved computational efficiency. The
modular concept for equipment subroutines was used. This con-
~ept allows greater flexibility in simulating various processing
schemes. These equipment subroutines are specialized only to
the extent that more rapid convergence can be obtained by using
the engineer's knowledge of the system. Wherever possible the
equipment subroutines were formulated to allow for extension
to models which give a more sophisticated description of the
processing unit.
Volume III of this report is a description and
listing of the computer program used.
2.1.3.1
Executive System
The function of the executive system is to interconnect
the various process units in the appropriate fashion and control
the sequencing of the computer operations.
The process units are interconnected by means of a
"process matrix" during an initial phase of computer operations.
In this phase, model input data are read into the machine, the
process matrix is used to define the processing scheme, and
each equipment "box" is initialized. The interconnection of
equipment boxes for the prototype system is shown in Figure 2-3.
Here each processing unit is labeled by a number, its name, and
the subroutine designation. For example, the subroutine used
for the scrubber [equipment number (4)J is designated SCRUBR.
In addition, the various process streams are labeled by letter
abbreviation and a stream number.
,:;
-15-
-------
SG3
5
(5) (13)
Reheater SGa I.D. Fan
(CLRHTR) 4 (PMPFAN)
WM
14
3 SGl
(3) (4) (9) Process
Cooler GSa Scrubber SF Water
(CLRHTR) 2 (SCRUBR) 15 Hold Tank
(EQMIXR)
6 SB SR
1 GS1 8
I (6)
.... Scrubber
0\ Tee
I Effluent EB CF
Hold Tank 7 (DIVDER) 9
(None)
I I
GSa (10) SG1 -
2 System 3 .
WM (OVALMB) FB
-
14 >- 13
FIGURE 2-3 SIMULATION OF PRO' ~( ~)E SYSTEM
(7)
Clarifier CL
(CLRFYR) 10
11 CB
(8)
Filter
. (FILTER)
FL
12
FB
13
<8
-------
- ,-----
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE SI2 . 454.9535
The "process matrix" used to define this processing
scheme is given on the first page of computer printout for
each simulation case. (A simulation case is given in T.N. 200-
. .
004-]9 in Volume II of this report.) The "process description"
information defines the interconnection of process units. For
example, equipment number 4 is the scrubber (subroutine SCRUBR).
It has input streams numbered 2 and 15 and output streams
numbered 3 and 6.
which the
indicated
printout.
The executive system also must know the order in
process calculations are to be made. This order is
in the lower half of the first page of computer
After the initialization phase has been completed,
the executive routine transfers control of computer operations
to the appropriate subroutines until the process calculations
have been converged. At this point, the calculated results
are printed. The last ten pages of computer printout show these
results in the form of stream vector information. Stream vectors
are the means by which information is transferred from one
equipment routine to another.
2.1.3.2
Equipment Subroutines
Simulation of the prototype system involved the use
of nine equipment subroutines.
(1)
Cooler (CLRHTR) - determines the heat
exchange required to cool the flue gas
stream to a specified scrubber inlet
temperature.
-17-
-------
Radian Corporation
(2)
(3)
(4)
(5)
(6)
(7 )
(8)
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN. TEXAS 78758 . TELEPHONE 512 - 454-9535
Scrubber (SCRUBR) - determines if the scrubber
feed stream contains enough basic species to
scrub the specified amount of SO~ from the
flue gas and forces iteration until convergence
is obtained.
Reheater (CLRHTR) - determines the heat
exchange required to reheat the stack gas
stream to a specified outlet temperature.
I.D. Fan (PMPFAN) - determines the energy
required to pass the stack gas out of the
system based upon pressure losses in the.
cooler, scrubber, and reheater.
Process Water Hold Tank (EQMIXR) - combines
the process streams returning to the scrubber
(sl~rry recycle, clarifier liquid, and
filter liquid) with the water make-up and
allows the exiting scrubber feed stream
to reach solid-liquid equilibrium.
Filter (FILTER) - provides for solid-liquid
separation based upon specified weight
fraction solids in the filter bottoms stream.
Clarifier (CLRFYR) - provides for solid-
liquid separation based upon specified
amount of slurry recycle.
Tee (DIVDER) - divides the exit stream from
the effluent hold tank into the slurry
recycle and clarifier feed streams.
-18-
-------
--
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN. TEXAS 78758 . TELEPHONE 512 . 454.9535
(9)
System (OVALMB) - determines the amount of
water vaporized in the scrubber and the
scrubber temperature using an approximate
scrubber energy balance and then determines
amount and composition of the filter
bottoms stream using a system material
balance.
Actually, the ninth process unit is the effluent hold tank. However,
the System calculational routine eliminates the need for
effluent hold tank calculation. The same subroutine (CLRHTR)
can be used for the cooler and the reheater.
In addition to these equipment subroutines, four
"source" or "sink" type routines were required. Subroutines
for the gas-solid stream (FLUGAS) and for the water make-up
stream (Wl'RMKP) are used to "generate" the process input stream.
Likewise, the exit stack gas (STKGAS) and filter bottoms (FLTRBM)
stream are "discharged" as process exit streams. These last
four subroutines essentially convert units for input/output
purposes and provide a source or sink for the terminal process
streams.
Details of the equipment subroutines are given in
Technical Note 200-004-19, Volume II of this report.
2.1.3.3
Convergence
Two ~ajor iteration loops are required to converge
the simulation cases for the prototype LIWS system. First, a
composition of the scrubber bottoms (SB) stream must be obtained
such that the equilibrium vapor pressure of S02 for this stream
is equal to the desired partial pressure of the stack gas (SG)
-19-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454.9535
stream. In this case, the amount of slurry recycle (SR) or
scrubber feed (SF) is used as the iterative parameter for
"SO:a convergence". Secondly, the fraction of CO;;! absorbed by
process [XA(COa)] is also assumed to be controlled by the
equilibrium vapor pressure of COa above the SB stream. The
Radian process model assumes an initial value for XA(CO:a) and
then adjust this parameter until the molality of total carbonate
in the SB stream coincides with the molality required to give
the CO:a partial pressure of the stack gas.
The approach used to converge the prototype simulation
cases is to iterate on SOa via a "loop" internal to the CO:a
convergence loop. Details are given in Volume II, Technical
Note 200-004-19.
2.2
Simulation Results
A number of simulation cases have been conducted for
the LIWS process shown in Figure 2-2. The design strategy used
in determining the set of simulation cases involved selecting
a set of "base conditions" and then examining the effect of
various process variables when these are varied from the base
level. The base conditions used for these prototype simulations
are listed in Table 2-1. A log of the simulation cases is given
in Table 2-2.
Computer printouts presenting the detailed results
for one simulation case are included in Volume II of this
report, Technical Note 200-004-19. These printouts give'
predicted stream flow rates, compositions, and properties.
Here, the process streams are designated by the stream numbers
shown on the schematic in Figure 2-2. Using a "process matrix"
the Radian executive routine interconnects the various equipment
-20-
-------
Radian Corporation
.
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN. TEXAS 78758 . TELEPHONE 512 - 454-9535
TABLE 2-1 -
BASE CONDITIONS FOR PROTOTYPE SIMULATIONS
Process Variable
Theoretical Limestone
Time Hydrating in System
Time Hydrating in Scrubber
Timestone Solids Dissolving
Recycled Solids Dissolving
Limestone Composition (m.f. MgO)
Fly Ash Level (lbs.FA/1bs.S)
Soluble Na (w.f. of Fly Ash)
Soluble C1 (w.f. of Fly Ash)
SOg in Flue Gas (m.f.)
S02 Absorbed
S02 Oxidized
CO2 in Flue Ga.s
NOx in Flue Gas
NOx Absorbed
Wt. Fraction Solids in FB
Solids Ppt. in Scrubber
Scrubber Liquor Temp. (OF)
( Gal. of SF )
Scrubber Feed Rate \1000 ACF of FG
(m. f. )
(m. f . )
Abbreviation
TL
LU
LH
LD
SD
LC
FA
XWFA(Na)
XWFA(C1)
YFG ( SOg )
XA( SO:;l )
XO
YFG(COg)
YFG(NOx)
XA(NOx)
XWSFB
NSP
Ts
SFTFG
-21-
Base Value
1.50
0.75
0.20
0.40
0.35
0.00
3.35
0.0038
0.0045
0.002
0.90
0.50
0.145
0.0005
0.20
0.60
No (NSP=1.0)
120
14
-------
.
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
TABLE 2-2
LOG OF PROTOTYPE SIMULATION CASES
Case
Designation
Process Variables Deviating from Base Value
PSN1A
1B
Ts = 112.0F, NSP = Yes
2A
2B
NSP = Yes
Ts = 112°F
3A
3B
(Base Conditions)
LC = 0.20 m.f. MgO, NSP = Yes, Ts = 112°F
. 4A
4B
LC = 0.20 m.f. MgO, NSP = Yes
LC = 0.20 m.f. MgO, Ts = 112°F
7
8
LC = 0.20 m.f. MgO
XWSFB = 0.45 w.f. Solids
9
10
XWSFB = 0.45 w.f. Solids, LC = 0.20 m.f. MgO
LH = 0.00 hydrated
11
12
LH = 0.00 hydrated,
Ls, = 2000* GPM
LC = 0.20 m.f. MgO
13
14
LSF = 2100 GPM, LC = 0.20 m.f. MgO
XO = 0.10 oxidized
15
16
XO = 0.10 oxidized, LC = 0.20 m.f. MgO
XO = 0.90 oxidized
17
18
XO = 0.90 oxidized, LC = 0.20 m.f. MgO
XWSFB = 0.75 w.f. solids
XO = 0.30 oxidized
19
XO = 0.00 oxidized
*
Case converged with LSR = 0 and Ls, = 2000.
-22-
-------
Radian Corporation
TABLE 2-2
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-9535
LOG OF PROTOTYPE SIMULATION CASES (cont.)
Case
Designation
20
21
22
23
26
Process Variables Deviating from Base Value
XWFA(Na)= 0.0212 w.f. Na, XWFA(C1) = 0.0311 w.f. C1
LU = 0.90 utilized
LU = 0.90 utilized, LC = 0.20 m.f. MgO
XWFA(Na) = 0.0125w.f. Na, XWFA(C1) == Or0181 w.f. C1
LH = 0.20 hydrated, LD = 0.60 dissolved, SD = 0.60 dissolved
-23-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
units for the prototype system in the manner depicted by Figure
2-2. Stream information for the fifteen process streams in
the prototype system is given in the cgs unit system. Some of
the flow rates and stream properties have been converted to the
American engineering system of units anc;J are included as the
last part of the stream information. Values for the model input
parameters are given as the first page of the computer printout
for each case.
Simulation results will now be discussed as they
reflect the importance of a number of process variables. These
variables include: (1) sulfite oxidation, (2) ionic strength
of the process liquor, (3) limestone amount and composition,
(4) scrubber feed rate, (5) scrubber liquor temperature, (6)
solids precipitation in the scrubber, (7) extent of solids
dissolution, and (8) the fraction of solids in the filter bottoms
stream. General remarks concerning the process stream composi-
tions and unit energy balances will also be presented.
2.2.1
Sulfite Oxidation
The effect of sulfite oxidation on the operation of
the limestone wet scrubbing process is shown in Tables 2-3 and
2-4. For the cases in Table 2-3, the percent oxidation was
varied from 0 to 90% for a system utilizing a pure CaO lime
solid. The scrubber liquid to gas ratio was maintained at 14
gallons of scrubber feed per 1000 ACF of flue gas, except for
the 0% oxidation case. In this case, only 13.4 ga1/l000 ACF
were needed for 90% SOa removal from a 2000 ppm S02 flue gas.
As the percent oxidation increases from 10 to 90% the volume frac-
tion of slurry recycle required in the scrubber feed progressively
increases from 0.208 to 0.819. This loss of scrubbing capability
with increased oxidation is evidently caused by the formation
-24-
-------
Radian Corporation
TABLE 2-3
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
- EFFECT OF SULFITE OXIDATION IN SYSTEM
WITH PURE CaO ADDITIVE
QUANTITY
Sulfite Oxidation, %
Scrubber L/G (Ga1/1000 ACF)
Slurry Recycle/Feed Ratio
(pH) S B
BASIC-SPECIES MOLALITIES
Ins F (OH-)
InU (SO~)
Insr [CaSOs (,e) J
In $ F [MgSO,3 (1,) ]
In.s F (total basic
species~")
FB SOLID-PHASE MOLALITIES
mfS[CaS03° ~H20(S)J
InF B [CaSO.:, (S) J
~r LIQUID-PHASE MOLALITIES
m~F[total SOa(t)J
ffisp[tota1 S03(1,)J
P SN19
o
13.4
o
4.37
1.6x10-2
1.7x10-p
2.6x10-4
o
2.7x10-a
5.34
0.00
2.8x10-'1
0000
PSN13
10
14.0
0.208
4.30
1.6x10-2
1.7x10-s
2.6x10-4
o
2.8x10-e
4.80
0.52
208x10-4
1.0x10-z
PSN18
30
14.0
0.577
4.20
1.6x10-2
1.7x10-c
2.6x10-4
o
2.8x10-:;
3.73
1.59
2.8x10-4
1 . Ox 10 -:3
PSN2B
50
14.0
0.710
4.06
-2
1.6x10
1.7x10-5
2.6x10-4
o
2.8x10-a
2.65
2.65
2.8x10-4
1.0x10-z
PSN15
90
14.0
0.819
3.39
1.6x10-Z
1.7x10-=
2.6x10-4
o
2.8x10-2
0.52
4.70
2.8x10-4
1 . Ox 1 0 - 2
.l~
Basic species include OH-, SO~, CO~, CaOH+, CaS03(1,), MgOH+, MgSOs(t),
and NaOH(1,).
-25-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
TABLE 2-4
- EFFECT OF SULFITE OXIDATION IN SYSTEM
WITH 20 MOLE/o MgO ADDITIVE
QUANTITY P SN14 P SN4B P SN16
Sulfite Oxidation, % 10 50 90
Scrubber L/G (Gal/1000ACF) 14.0 14.0 14.0
Slurry Recycle/Scrubber
Feed Ratio 0.837 0.877 0.900
(pH) ~ e 4.32 4.08 3.52
BASIC-SPECIES MOLALITIES
ID~ F (OH-)
m 0 ; (SO ~)
IDSF [CaS03(.e)]
mH l!'1g S03 (1,) ]
rn 0 F (total basic
1.8xlO-::;
1.8xlO-'"
2.5xlO-4
1.4x10-3
2.3xlO-3
species';'()
FB SOLID-PHASE MOLALITIES
m,~[CaS08' ~H20(s)J
ID,~[CaSO-),(s)J
5.04
0.41
SF LIQUID-PHASE MOLALITIES
ms,[total S02(t)J
rns,Ltotal S03(t)J
1.8xlO-S
1.5xlO-l
1.9xlO-c:
1.5xlO-""
2 . 5 Ox 1 0 - .;
1.lx10-3
2.0xlO-:;
2.76
2.63
1<5xlO-~S
1.2xlO-l
2.0xlO-;;
1 . 3 x lO -."
2.5xlO-';
9.4x10-'=
1.8xlO-s
0.54
4.79
1 . 3 x 1 0 - ~,
1 . lx 1 0 - i
;',
Basic species include OH-, SO~, CO~, CaOH+, CaS03(t),
MgOH+, MgS03 (t), and NaOHO,).
-26-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-9535
of the strongly acidic sulfuric acid (the pH of the scrubber
bottoms stream is lowered).
A similar increase in the fraction of slurry recycle
with increased sulfite oxidation is observed in the simulation
cases for a limestone additive containing 20 mole% MgO and
80 mole% CaO. However, the change in slurry recycle with sul-
fite oxidation is not as significant for the MgO cases. As the
percent oxidation increases from 10 to 90% the fraction of
slurry recycle only increases from 0.837 to 0.900. This observa-
tion can be explained in terms of the nature of the basic species
entering the scrubber with the scrubber feed stream. Basic
species are defined as those species capable of liberating a
hydrogen ion due to the first ionization of sulfurous acid, i.e.,
basic with respect to sulfurous acid.
In the calcium-based liquors, a major portion
(approximately 60%) of the basic species entering the scrubber
in the scrubber feed liquid is the hydroxyl ion. In the
magnesium-based liquors, sulfite species [SO~, CaS03(.t), MgS03(.t)]
make up a major portion (75-80%) of the basic species. Thus,
the calcium-based scrubbing systems are much more sensitive to
factors that influence the concentration of hydroxyl ion. Since
sulfite oxidation gives rise to a stronger acid (sulfuric,
instead of sulfurous), less hydroxyl ions (aH+ x aOH- = KW) will
be available.
-27-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
Of course oxidation changes the type of solids
leaving the system. As expected, the amount of calcium sulfite
solid leaving the system in the filter bottoms (FB) stream
progressively decreased with oxidation while the amount of
calcium sulfate increased. The concentrations of sulfites and
sulfates in solution do not follow this trend, however, since
they are controlled by solubility relationships. These observa-
tions are reflected by the molalities presented in Table 2-3 and
2-4.
2.2.2
Ionic Strength
The effect of ionic strength for systems injecting
a pure CaO limestone is given in Table 2-5. The ionic strength
of the scrubbing solution was varied from 0.48 up to 1.00 by the
addition of sodium chloride to the m0gel. This would correspond
to different degrees of fly ash solubility in the actual process.
For the cases presented here, the effect of ionic strength on
scrubbing efficiency was fairly minor. As the ionic strength
increased from 0.48 to 1.00, the fraction of slurry recycle
required in the scrubber feed only increased from 0.710 to
o
0.733. This iricrease can be explained in terms of the total
concentration of basic species dissolved in the scrubber feed
liquid. Here again basic species are defined as all species
capable of liberating a hydrogen ion due to the first ionization
of sulfurous acid. For the cases presented in Table 2-5, the
sum of the basic-species molalities in the scrubber feed decreases
from 2.78xlo-a to 2.62xlo-a g-moles/kilogram HaO as the ionic
strength increases.
-28-
-------
TABLE 2-5
EFFECT OF IONIC STRENGTH
CASE NUMBER
PSN23
QUANTITY
IONIC STRENGTH
SCRUBBER L/G (GAL/1000 ACF)
SLURRY RECYCLE/SCRUBBER
FEED RATIO
I
I'V
\.0
I
LIQUID-PHASE MOLALITIES OF SF
OH-
CaOH+
CaS03(t)
NaOH (.~)
All Basic Species
Ca++
ACTIVITY COEFFICIENT OF SF
SPECIES
Ca++
OH-
S03
PSN2B
0.48
14.0
0.710
-2
1.56x10
-2
1.17x10
2.59x10-4
-4
1.95x10
-2
2.78x10
1.14x10-1
0.238
0.721
0.188
0.75
14.0
0.717
-2
1 . 5 Ox 10
1.14x10-2
2.48x10-4
-4
6.39x10
-2
2.73x10
1. 21x10-1
0.226
0.744
0.157
PSN20
1.00
14.0
0.733
-2
1.38x10
-2
1.11x10
2.38x10-4
1.04x10-3
-2
2.62x10
1. 27x10-1
0.228
0.784
0.139
'0
~
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
Although the sum of the basic-species molalities does0
not change much in this instance, the individual molalities and
activity coefficients do vary significantly over the ionic
strength range shown with OH-, CaOH+, and CaS03(t) decreasing
and NaOH(t) increasing 0 The deviation of activity coefficients
from the ideal value of unity demonstrates their need in these
aqueous equilibrium calculations and suggests their consideration
in rate correlation. The significant variation of individual
molalities and activity coefficients with ionic strength suggests
that other regions of system operation may exist in which ionic
strength will have a significant effect on scrubbing efficiency.
2.2.3
Limestone Amount and Composition
The effect of limestone amount and composition is
shown in Table 2-6. In cases PSN2B and PSN4B, 112.5% of the
limestone required for 100% SO~ removal was considered available
for reaction in the system. (The rate of limestone addition [theo-
retical limestone] was set equal to 150% of the stoichiometric amount
and the limestone availability was 75%. Thus,~since 1.5xO.75 = 1.125,
the limestone available for reaction was 112.5%.) Case PSN2B cor-
responds to a process injecting limestone containing a negligible
amount of MgO, whereas Case PSN4B corresponds to a process injecting
limestone containing 20 mole% of reactive MgO. To remove 90% of the
S02 in the flue gas containing 2000 ppm SO~, 71.0% of the scrubber
feed was slurry recycle in Case PSN2B. On the other hand, Case
PSN4B required the scrubber feed to be 87.7% slurry recycle.
An explanation of this observation is possible when
the nature of the scrubbing liquid is examined for Cases PSN2B
and PSN4B. Enough calcium oxide has been added to the system in
PSN2B so that the scrubbing liquor becomes a saturated solution
with respect to calcium hydroxide (pH=11.3) and Ca(OH)~ leaves
-30-
-------
TA~LE 2-6
EFFECT OF LIMESTONE ~IOUNT AND COMPOSITION
CASE NUMBER
QUANTITY PSN2B PSN4B PSN21 PSN22
LIMESTONE AVAILABLE,
% THEORETICAL 112.5 112.5 135 135
LIMESTONE COMPOSITION,
% MgO 0 20 0 20
SCRUBBER L/G, GAL/
1000 ACF 14 14 14 14
SLURRY RECYCLE/SCRUBBER
FEED RATIO 0.710 0.877 0.510 0.510
I FILTER SOLIDS STREAM
VJ
t-I pH 11.3 8.4 11.3 11.3
I
IONIC STRENGTH 0.48 0.70 0.48 0.48
mFBL(Tota1 CaO) 1.5x10-1 -2 -1 1.5x10-1
2.2x10 1.5x10
.mFBL (Total MgO) 0 2.4x10-1 0 9.1x10-7
mFBS[Ca(OH)2] 9.6x10-1 0 2.16 6.1x10-1
mFBS[Mg(OH)2] 0 1.14 0 1.59
% S03 AS LIQUID SULFATE 0.36 4.54 0.37 0.36
% MgO AS Mg(OH)2 -------- 82.69 -------- 99.99+
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE 512 . 454-9535
the system as a solid in the waste solid. However, all of the
calcium is utilized in PSN4B [no Ca(OH)~ in the waste solids] so
that the scrubbing liquor becomes saturated with respect to
magnesium hydroxide (pH=8.4). The preliminary conclusion here
is that a calcium-based scrubbing liquor is capable of more
scrubbing than the calcium plus magnesium-based liquor. This
is under the assumption of equal solution rate kinetics.
If the amount of limestone available for reaction is
increased (refer to Case PSN2l for a 100% CaO system and Case
PSN22 for an 80% CaO-20% MgO system), the amount of slurry re-
cycle required to remove 90% of the SOa in a 2000 ppm flue gas
is reduced. In fact, for both cases PSN2l and PSN22, only 51.0%
slurry recycle is required. The reduction in amount of slurry
recycle for both cases is expected since with more limestone
available the solid contains more calcium and magnesium hydroxide.
[The simulation cases are run with a specified fraction of
Ca(OH)a and Mg(OH)a dissolving in the scrubber.] The fact that
Cases PSN2l and PSN22 require the same amount of slurry recycle
can be explained again by examining the composition of the
scrubbing liquor. It turns out that enough CaO is added in
PSN22 so that the scrub~ing liquor is saturated with respect to
Ca(OH)a' Thus the liquid phase of Cases PSN2B, PSN2l, and PSN22
are almost the same. The molality of total magnesium species
in the liquid phase for Case PSN22 is only 9.lxlO-? g-moles
per kilogram of liquid water. Almost all of the magnesium (99.99+%)
leaves the system as Mg(OH)a solid in Case PSN22, whereas only
82.7/0 leaves as Mg(OH)a solid in Case PSN4B.
For the calcium plus magnesium-type scrubber liquid
(PSN4B), the ionic strength of process liquor is 0.70 g-moles
per kilogram H20 as compared to 0.48 for calcium-type liquids
-32-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXA~ 78758 . TELEPHONE 512 - 454-9535
(PSN2B, PSN2l, PSN22). Of course, this would be expected since
magnesium species are more soluble. Along the same line, 4.54%
of the sulfate leaves the system in the liquid phase for Case
PSN4B. Less than 1% of the sulfate leaves in the liquid phase
for the calcium-type liquids. The amount of sulfate leaving the
system as a liquid would continue to increase as more magnesium
and less calcium is made available to the process.
2.2.4
Scrubber Feed Rate
The effect of scrubber feed rate is shown in Table
2-7. Here the cases shown in the first two data columns are the
same as those shown in Table 2-6, i.e., PSN2B and PSN4B. Cases
PSNll and PSN12 were made based upon high scrubber feed rates.
in the system using a pure CaO limestone (PSNll), 20 gal/1000
ACF of scrubber feed containing no slurry was sufficient to
remove 90% of the SOg from a 2000 ppm flue gas. In a system
using an 80 mole% CaO - 20 mole% MgO limestone (PSN12), 87.2%
slurry recycle was required at a scrubber feed rate of 21 gall
1000 ACF. This can be explained in terms of the amount of basic
species dissolved in the calcium-based and calcium plus magnesium-
based scrubbing solutions.
Comparison of Cases PSN2B and PSNll indicates that
the liquid-phase of the scrubber feed has made a significant
contribution toward absorbing SOa in the scrubber. On the other
hand, comparison of PSN4B and PSN12 indicates that the liquid-
phase for Ca plus Mg-based scrubbing solutions is not as potent.
In the Ca plus Mg-based cases, dissolution of the SF solids in
the scrubber account for a major portion of the scrubbing.
-33-
-------
TABLE 2-7
EFFECT OF SCRUBBER FEED RATE
I
W
.p-
I
CASE NUMBER
OUANTITY PSN2B PSN4B PSN11 PSN12
,
LI}lliSTONE COMPOSITION,
% HgO 0 20 0 20
SCRUBBER L/G, GAL/
1000 ACF 14 14 20 21
SLURRY RECYCLE/SCRUBBER
FEED PATIO 0.710 0.877 0.00 0.872
Wt.% SOLIDS IN SF 2.1 6.0 0.1 3.8
Wt.% SOLIDS IN SR 2.9 6.9 0.0 4.3
mSF (BASIC SPECIES IN 2.8x10_2 -3 -2 2.6x10-3
LIQUID) 2.0x10 2.7x10
mFBS [Ca(OH)2J 0.96 0 0.57 0
mFBS [~ig (OH) 2J 0 1.14 0 1.10
FRACTION C02 ABSORBED -4 -4 -3 3.5x10-4
5.2x10 2.4x10 1 . 4x 1 0
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-9535
These remarks are supported by (see Table 2-7): (1) the amount
of solids in the scrubber feed stream (weight per unit time) is
about the same for Cases PSN4B and PSN12 and (2) the concentra-
tion of basic species in solution is approximately 10 to 15 times
as much for the calcium-based solutions as for the Ca plus magnesium-
based solutions (refer to the molalities of basic species in the
SF liquid shown in Table 2-7).
2.2.5
Circulating Liquor Temperature
The temperature of the scrubbing liquor was varied
by changing the temperature of the gas-solid stream entering
the scrubber. In cases PSN1A through PSN4A, a gas inlet tempera-
ture of 150°F was used with a resulting scrubbing liquor temperature
of 112°F. In cases PSNlB through PSN4B,a gas inlet temperature
of 225°F was used with a resulting scrubbing liquor temperature of
120°F. The "A" cases correspond to a flue gas containing 8 mo1e%
water vapor which has been precooled to 150°F. The "B" cases
correspond to the same flue gas with less precooling. As mentioned
earlier, the temperature of the other process liquors was set
equal to the temperature of the scrubber liquor.
The effect of changing the circulating liquor
temperature on the scrubbing capability of the system is shown
by the results in Tables 2-8 and 2-9. The cases in Table 2-8
are based upon a system feeding a pure CaO additive, whereas the
cases in Table 2-9 correspond to a system with a 20 mo1e% MgO -
80 mo1e% CaO additive. Comparison of the slurry recycle to scrubber
feed rate (SR/SF) ratio for the A and B cases shows that changing
the circulating liquid temperature from 112 to 120°F has little
influence on the scrubbing capability of the process. For the
-35-
-------
Radian Corporation 8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
TABLE 2-8 - EFFECT OF CIRCULATING LIQUOR TEMPERATURE AND
SCRUBBER PRECIPITATION OF SOLIDS IN SYSTEM
WITH PURE CaO ADDITIVE
QUANTITY
Scrubbing Liquor Temp. (~F)
Solids Ppt. in Scrubber
WM/SF Ratio
SR/SF Ratio
Fraction of CO2 Absorbed
SF CHARACTERISTICS
Wtc% Solids
pH
Ionic Streng th
Molalities in
Key-Species
Total SO:2
Total CO2
Total SO,>
Total CaO
Total MgO
SB CHARACTERISTICS
Wt.% Solids
pH
Ionic Strength
Key-Species
Total SO;;
Total CO~
""
Total S03
Total CaO
Total MgO
Molalities in
Ion Molalities
++
Ca++
Mg -
HSO:;
SO=
"'-
HC03
PSN1A
112
Yes
0.97x10-;;
0.729
.5.8x10-4
2.25
11.5
0.485
Liquid
2.62x10-4
4.66x10-s
1. 04x10-;;
1.52x10-1
0.00
2.98
3.89
0.487
Liquid
1.27x10-;;
2.70x10 :.;
1.03x10-:;
1.45xlO 1
0.00
0.121
0.000
1.25xlO-:;
5.1lx10-3
1.76x10-s
-36-
CASE NUMBER'
PSN1B PSN2A
120 112
Yes No
1.75x10-~ 0.97x10-;
0.752 0.696
4.7x10-4 6.2x10-4
2.59
11.3
0.476
2.75x10-4
4.15x10~6
0.97x10-2
1.50x10-}
0.00
3.35
3.96
0.481
1.18x10 -2
2.39x10-3
0.95x10-:;
1.44xlO-1
0.00
0.119
0.000
1.15x10-:;
4.66x10 :;
1.85xlO-;o
1. 91
11.5
0.484
2.62x10-4
4.66x10-"
1.04x10-;;
1.52x10 -
0.00
2.44
3.95
0.519
1.48x10-2
2.69xlO-s
2.50x10-;;
1.6lx10-1
0.00
0.129
0.000
1.43x10-2
1.24xlO-2
2.03xlO-G
PSN2B
120
No
1.75x10-3
0.709
5.2x10-'"
2.10
11.3
0.475
2.7f)xlO-tg
4.15x10-a
0.97x10-;;
1. 50x10-:'
0.00
2.64
4.06
o . 514
1.49x10-2
2.38x10-,3
2.44x10-:;
1 . 61x 10 -
0.00
0.127
0.000
1.44x10-2
1. 19x1O -:;
2.32xlO-s
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-9535
TABLE 2-9
- EFFECT OF CIRCULATING LIQUOR TEMPERATURE
~~D SCRUBBER PRECIPITATION OF SOLIDS IN
SYSTEM WITH 20 MOLE % MgO - 80 MOLE % CaO
ADDITIVE
QUANTITY
Scrubbing Liquor Temp. (OF)
Solids Ppt. in Scrubber
WM/sF Ratio
SRI SF Ratio
Fraction of CO2 Absorbed
SF CHARACTERISTICS
Wt.% Solids
pH
Ionic Strength
Key-Species Molalities in
Total SO~
lotal CO;
Total S03
Total CaO
Total MgO
SB CHARACTERISTICS
Wto% Solids
pH
Ionic Strength
Key- Species
Total S0:G
Total CO2
Total S03
Total CaO
Total MgO
Molalities in
Ion Molalities
++
Cart
Mg -
HSOs
SO:=
HCO;
PSN3A
112
Yes
O. 9lxlo-a
0.655
5.6xlO-';'
1.49
8.48
0.892
Liquid
2.93xlO-3
3.5lxlO-4
2.65xlO-1
1.60xlO-2
3.88xlO-1
2.05
4.41
0.927
Liquid
4.63xlo-a
2057xlO-3
2.50xlO-1
1.74xlo-a
3.94xlO-l
1.13xlo-a
2.35xlO-1
4.34xlO-2
8.43xlo-a
6.l6xlO-5
-37-
CASE NUMBER
PSN3B PSN4A
120 112
Yes No
1.70xlo-a 0.93xlO-2
o . 7 5 2 '0 . 883
4.lxlO-4 2.7xlO-4
2.47
8.39
0.787
2.36xlO-3
2.57xlO-4
1.96xlO 1
1.74xlO-2
3.l5xlO-1
3.02
4.46
0.830
40 04xlo-a
203lxlO-3
1.88xlO-1
1.89xlo-a
3.25xlO-1
1. 26xlO-2
2.09xlO-l
3.8lxlo-a
604lxlo-a
6.l2xlO-s
5.99
8.56
0.721
1.55xlO-3
1.73xlO-'"
1.39xlO-1
2.2lxlo-a
2.52xlO-1
6.34
3.97
0.766
1.6lxlo-a
2.57xlO-3
1.54xlO-l
2.62xlo-a
2.70xlO-1
1.8lxlO-:;;
1.83xlO-l
1.57xlo-a
5. 83xlO-2
2.23xlO-5
P SN4 B
120
No
1.72xlo-a
0.877
2.4xlO-4
5.98
8.45
0.686
1.52xlO-3
1.56xlO-4
1.24xlO-l
2.23xlo-a
2.35xlO-1
6.43
4.06
0.735
1.62xlo-a
2.29xlO-3
1.40xlO -1
2.64xlo-a
2.55xlO-1
1. 82xlo-a
1.75xlO-1
1.57xlO-:G
5 .13xlo-a
2.54xlO-s
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
pure CaO cases, an increase in temperature lowers slightly the
scrubbing capability of the process. For the 20 mo1e% MgO cases,
the temperature effect is dependent upon whether precipitation
is allowed in the scrubber.
It should be emphasized here that these conclusions
are based upon the model process assumptions. More specifically,
the effect of temperature on reaction rates has not been con-
sidered and the temperature in each process vessel has been taken
to be same.
Since the scrubbing liquor temperature is determined
by adiabatic humidification of the flue gas stream, the amount
of water makeup (WM) required for the process changes for the A
and B cases. As expected, approximately twice as much water
makeup is required for the higher temperature cases.
In general, the differences between the results of
the lower and higher temperature cases (A vs. B) are relatively
minor. More COa is absorbed by the process at the lower tempera-
ture operation. The weight percent solids in the process streams
is slightly less for the lower temperature operation. Most
key-species and ion molalities are within 15% of each other for
the comparable A and B cases.
2.2.6
Solids Precipitation in Scrubber
The effect of solids precipitation in the scrubber is
also shown in Tables 2-8 and 2-9. The first two cases in each
table correspond to processes in which complete precipitation of
solids occurs in the scrubber, i.e., the liquid phase is saturated
with respect to the "equilibrium-type" solids (CaS03' xHaO,
-38-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN. TEXAS 7875B . TELEPHONE 512 . 454.9535
CaS04' xHa 0, CaC03, MgS03' xHa 0, and MgCO,3' xHa 0) . In the
last two cases in each table, no precipitation of solids is
allowed to occur in the scrubber, i.e., the scrubbing liquor
becomes supersaturated. The two extremes s~ould then bound
the effect of solids precipitation in the scrubber.
The effect of solid precipitation on scrubbing
capability of the process (refer to SR/SF values) is not signifi-
cant for the calcium or the calcium-magnesium based systems.
For the calcium-based system, solids precipitation results in
an approximately 5% increase in the SR/SF ratio. On the other
hand, solids precipitation in the calcium-magnesium based system
results in a 15% decrease in the SR/SF ratio.
The following rather general observations can be made
regarding the effect of solids precipitation.
(1)
(2)
(3)
(4 )
More CO2 is absorbed when solids
precipitation occurs.
In the calcium system, the weight
percent solids of the circulating
streams is higher when precipitation
occurs.
Surprisingly, the weight percent
solids of the circulating streams
is lower when precipitation occurs
for the calcium-magnesium system.
As expected, the scrubbing liquor
(SB) contains less total SOg, S03,
and CaO when precipitation occurs
in the calcium based system.
-39-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
(5)
In the calcium-magnesium based system,
the total MgO in circulating liquors is
greater when precipitation occurs in
the scrubber. As a result, concentra-
tion of total S02' COa, and S03 is also
greater in these cases.
2.2.7
Limestone Reactivity
The effect of limestone reactivity was examined by
varying the fractions of lime hydration and hydroxide dissolution
in the scrubber. The re'sults of these simulation cases are
presented in Table 2-10. The first three cases (PSN9, 2B, and
26) show the effect of changing limestone reactivity in a
system using pure CaO additive. The last two cases (PSN10 and
4B) show this effect in a 20 mole% MgO - 80 mole/o CaO sys tern.
In the first case (PSN9), the lime hydration step is
considered to be very slow so that essentially none of the CaO
entering with the gas-solid stream hydrates in the scrubber.
Thus, for this case, none of the lime in the gas-solid inlet
stream is available for reaction. Only the hydrated solids in
the scrubber feed stream are allowed to react. For PSN9, 35%
of the Ca(OH)~ in the scrubber feed stream has been taken as
reacting in the scrubber. In PSN2B, 20% of the lime (CaO)
entering the scrubber in the gas-solid stream is allowed to
hydrate and 40% of the resulting hydroxide is taken as reacting
with the scrubber liquor.
-40-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
TABLE 2-10
- EFFECT OF LIMESTONE REACTIVITY
QUANTITY PSN9 P SN2 B P SN26 PSN10 P SN4B
Limestone Composition, % MgO 0 0 0 20 20
Lime Hydrating in Scrubber, % 0 20 20 0 20
Limestone Solids Dissolving,% ------ 40 60 ------ 40
Recycled Solids Dissolving, io 35 35 60 35 35
SR/SF Ratio 0.783 0.710 0.523 0.891 0.877
SF CHARACTERISTICS
Wt.%.Solids 3.09 2.10 0.95 6.88 5.98
pH 11.3 11.3 11.3 8.45 8.45
Ionic Strength 0.475 0.475 0.475 0.672 0.686
SB CHARACTERISTICS
Wt.% Solids 3.62 2.64 1.49 7.32 6.43
pH 4.06 4.06 4.06 4.08 4.08
Ionic Strength 0 . 514 0 . 514 0 . 514 0.721 0.735
Note:
Liquid-phase compositions of pure CaO
identical. Liquid-phase compositions
differ by less than 5%.
cases are essentially
of 20 mole% MgO cases
,-
-41-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-9535
Comparison of cases PSN9 and PSN2B shows the sensitivity
of the process operation to additional reactivity. The SR/SF
ratio is decreased by about 10% with this increase in additive
reactivity. If the reactivity of the hydroxides (both in the
gas-solid inlet and scrubber feed streams) is increased to 60%
(refer to case PSN26) the SR/SF ratio is reduced even further.
A similar but less dramatic effect is observed in the 20 mole%
MgO cases (compare PSNIO and PSN4B).
Aside from the change in SR/SF ratio, the major
difference between these simulation cases is in the weight per-
cent solids in the circulating process streams. For the pure
CaO cases listed, the solids loading ranges between 1.0 and
3.6 wt.% solids. For the 20 mole% MgO cases, the solids load-
ing is in the 6.0 to 7.3 wt.% solids range. The compositions
of the liquid phases of the SF and SB streams for each of these
cases is very similar. The liquid-phase compositions for the
pure CaO cases are identical, whereas these compositions for
the two 20 mole% MgO cases differ by less than 5%.
2.2.8
Fraction Solids in FB
The major effect of changing the weight percent
solids in the filter bottoms stream is to change the system's
purge rate with respect to the solids. Simulation results
showing this effect on scrubbing capability for the calcium
and calcium-magnesium based systems are given in Table 2-11.
In both systems, the effect of changing the weight percent
solids is not very significant. For the calcium-based cases
(PSN7, 2B, and 17), the average deviation of the SR/SF ratio is
less than 2.3%. For the calcium-magnesium cases (PSN8 and 4B),
the average deviation is less than 1. 0%.
-42-
-------
TABLE 2-11 - EFFECT OF WEIGHT PERCENT SOLIDS IN FILTER BOTTOMS STREAM
CASE NUMBER
QUANTITY PSN7 PSN2B P SN17 PSN8 P SN4B
Limestone Composition, % MgO 0 0 0 20 20
Wt.% Solids in FB 45 60 75 45 60
SR/SF Ratio 0.691 0.710 0.736 0.891 0.877
WM/ SF Ratio x 103 22.3 17.5 14.4 22.0 17.3
FB/SF Ratio x 103 10.0 5.5 2.7 9.8 5.3
SF CHARACTERISTICS
Wt.% Solids 1.95 2.10 2.32 6.17 5.98
I pH 11.4 11.3 11.2 8.59 8.45
.J:'-
W Ionic Strength 0.285 0.475 0.905 0.406 0.686
I
SB CHARACTERISTICS
Wt. % Solids 2.50 2.64 2.83 6.63 6.43
pH 4.08 4.06 4.03 4.09 4.08
Ionic Strength 0.317 0.514 0.953 0.457 0.735
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-9535
As the weight percent solids in the filter bottoms
stream is varied from 45 to 75%, the liquid ratio of FB/SF
varies from 10.0 to 2.7 for the calcium-type cases. (Note that
the WM/SF changes merely reflect the FB/SF changes.) This change
in liquid purge rate has a major influence on two items. First
of all, the ionic strength of the circulating liquors changes
by a factor of three. Thus, the effect of changing the weight
percent solids in FB should contain the influence of changing
ionic strength on scrubbing capability (SR/SF ratio). As
discussed earlier, this effect was fairly minor for the process
operating near the conditions of Case PSN2B.
The second item being changed by varying the purge
rate is the amount of sulfite and sulfate leaving in the process
in the liquid phase. This effect on process scrubbing capability
should be relatively minor. The predicted concentrations of
total SO;! and S03 in FB differ by about 15% for cases PSN7 and
17.
Results from the calcium-magnesium cases (PSN8 and
4B) shown in Table 2-11 reflect essentially the same trends.
Trends in the solids loading for the SF and SB streams are also
shown in Table 2-11.
2.2.9
Energy Balance Considerations
The enthalpy of each process stream is calculated
by the Radian process model. At this point, energy balances
around each process unit are not calculated. As mentioned
earlier, the temperature of the scrubbing liquor is determined
-44-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE 512 . 454-9535
by the adiabatic humidification of the flue gas stream. This
calculation neglects (1) heat losses from the scrubber and
(2) heats of reaction and mixing for the scrubber. The tempera-
ture of the other process vessels was set equal to the scrubber
liquor temperature.
This simplified approach is justified from the
standpoint that reaction rate data are presently not available
so that this version of the model does not warrant a more
sophisticated energy balance calculation. Later, as rate data
become available and equipment sizes are more predictable,
energy balances including heat losses to the surroundings can
be incorporated into the model. The ability to calculate the
enthalpy of a process stream is a maj or step in the direction of
being able to calculate energy balances.
Based upon the temperature determined for the
scrubbing system (using the adiaba.tic humidification method),
it is interesting to examine the energy "imbalance" for each
process unit. This can be done conveniently by computing the
adiabatic temperature difference 6TA' i.e., the temperature
increase of the vessel's exit streams that would be required to
satisfy the exact adiabatic energy balance. The calculation
involves the following equations:
6TA
Exit Streams
!;.H.
~ ~
Inlet Streams
I:.H.
] J
Exit Streams
I:.w.Cp,
~ ~ ~
(2-1)
=
where
Hi, Hj
=
enthalpies of the exit and inlet
streams in calories per second
-45-
J
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 7B75B . TELEPHONE 512 - 454-9535
w.
~
=
mass flow rate of the exit stream
i in grams per second
Cpo
~
=
average heat capacity of the exit
stream i in calories per gram - ce.
For the base simulation case (PSN2B), the adiabatic
temperature differences for the scrubber, effluent hold tank,
and the process-water hold tank are 1.1, 0.78, and less than
O.lloC, respectively. That is to say, the temperatures of the
stack gas (SG) and scrubber bottoms (SB) streams leaving the
scrubber should be increased by about 1°C to place the scrubber
in energy balance if the scrubber were operated adiabatically.
Likewise, the slurry stream leaving the effluent hold tank (E)
should be approximately 0.8°C higher if this tank operates
adiabatically.
Generally, the adiabatic temperature differences for
the other simulation cases do not significantly differ from
PSN2B. The most significant variation is observed for the cases
in which the sulfite oxidation parameter is varied. For PSN19
with no sulfite oxidation, the scrubber 6TA is O.SloC compared
to 1. 1°C for PSN2B (SO% oxidation). For PSNlS with 90% oxida-
tion, the scrubber 6TA differ from PSN2B by 10% or less. A
similar effect is observed in the effluent hold tank 6TA' For
PSN19 (XO=O), the effluent hold tank 6TA is 1. 12°C compared to
0.78ce in PSN2B (XO = 0.50). For PSN15 (XO = 0.90), the 6TA
is 0.45ce.
of a degree
mation used
Since these values for 6TA are all in the neighborhood
or so centigrade, the error involved by the approxi-
should be relatively minor if the vessels operate
-46-
-------
Radian Corporation
8409 RESEARCH BLVD, . P,O, BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
adiabatically. Actually, each process vessel will lose heat
to the surroundings, especially if they are uninsulated. Since
all of the 6TA calculated represent temperature rises, heat
losses will tend to counterbalance this error.
As rate data become available, the process
can be used to size equipment vessels. At this point,
rigorous energy balance including the determination of
losses to the surroundings should be incorporated into
model.
model
a more
heat
the
2.2.10
Process Flow Diagrams
The simulation results obtained from the Radian
process model contain data pertinent to the construction of
process flow diagrams. Figures 2-4 and 2-5 are process flow
diagrams for simulation cases PSN2B and PSN4B. Case PSN2B is
the base conditions simulation for a system with a pure calcium
limestone, whereas PSN4B is the base conditions simulation for
a system with 20 mole% MgO - 80% mole% CaO solids additive.
These figures summarize much of the pertinent pro~ess
information. Comparison of the liquid molalities for PSN2B
and PSN4B shows the difference in liquid compositions for the
pure calcium and the calcium-magnesium systems. The concen-
tration of MgO can vary markedly depending upon the amount and
reactivity of the MgO content in the solid additive. Likewise,
some of the other concentrations vary significantly as the
solution change from the calcium to the calcium-magnesium system.
-47-
-------
COOLER DUTY
70,'1306"tU/M\M.
I
.s::--
00
I
flUE GAS
COOLER
SC£UBBER
EFFLUEI-1T
HOLD T AMK.
CLA£IFIE~
FILTE~
PR.OC.ESS WA,ER
HOLD ,II.MK.
PROCESS
COt-!DITIOIJ5
7":0% LIME HYD~ATES IN :SYSTE.M
ZO % LIME H'/DR.ATES It.! SC.RU88ER.
40% LIMESTONE SOLIDS DISSOLVE II\J SCRUB8ER
3"''-. ~ECYCLED SOLIDS DISSOLVE IW SC.WBBER
MO SOLIDS PRECIPITATE It-! SCRUBBER
IW.F CIRCULATlI-1G liqUOR TEMPERII.TURE
LI MESI0t-1E
~
CoO
M90
CONlPOSITIOt-!
~
1.000
0,000
F"L Y ASH
COMPo
I-IOtO
t-!oCl
IIJSOLU8LE
FL'C ASH
COMPOSrTtOt-l
WT. FIC!:.
0.0011
0.0075
0.~914
90010 :50Z REMOVAL
'50% SO~ OXIOATIOr-J
20% r-JOx REI<\OVAL
PURE CALCIUM LlMESTOME
3.3S LB5 FLY ASH PER LB SULFUR It.! COAL
1':>0% IHlOlttT1CAL LIMESTONE
~ GA.'5 -50LID 5"t~EI>.M.5 ~ LIQuID - SOLI D S,.R.EAMS
QUA.I.ITIT'( 0 (2) (~) (4) (5) QUA.>JTITV (0;) (7) .~ (II) 0 @ @ ~
TOr!l.L n.ow eATE (LBS/MIIJ.) 5.~13 S,0 LIQUID OEt-I:')ITY (lBS/GAL) 5.0:.2 6.S0 e.50 8.50 e.':>o a.so 8.so 6.S0 e:z.s a.so
PRE~sue: (ATM) 1.000 0.990 0.971 0.9/01 1.000 pH 4.0/D 11.~4 I\. ~4 (1."34 11.34 11.~4 \\.34 11.~4 Go. Go5 1\.34
GAS CO/-o'.POSITlOM(MOLE Fit.) IONIC STI0 0.14'2>0 0.13910 0.13'='10 0.13'3<0 TOTII.L SOz 1.4'3'(IO"~ 2.7& "0"4 '2.1'>'(10"4 '2.7<;> < 10"4 ,.70;,,10'4 '2.10;-10'4 Z.7,,"<10-4 2.70; x 10"4 0 2.7<'<10"4
t-IO~ 0.000"" 0.0000; 0.0004 0.0004 0.0004 TOI",L CO~ '2. 38 .IO-~ 4.14 <10'''' 4.\4 < 10'~ 4.14 '(10-1. 4.14<10-1. 4.14110;' 4.14'(10-1. 4.14110'<' 0 4.15c10""
Oz 0.0300 0.0300 0.OZ84 0.OZ84 0.0284 10TII-L SO~ 2.44 x 10-z 9."4 cIO-~ ~.G.4.,0'" 9.(.4<10"~ 9.104.10'3 9.<'4dO'3 9.~4 .10' 9.<..4 '(10-3 0 19.':'''<10'
M2 0.7425 0.7425 0.7151 0.7151 0.71'51 TOTII-L M200; 1.40'(10-1 1.40<10-1 1.40 (/0-1 1.40. \0"1 1.40 x 10-1 \.40 '(10-1 1.40110'1 1.<10.10"1 0 1."38 '(10-1
HC.l 0.0000 0.0000 0.0000 0.0000 0.0000 TOTII.L Co 0 1.(.1'( 10-1 I.,,"Z-IO-I \.'>2 ('0-1 I.';Z'(IO"' 1::''2'(\0-1 1.':>2<10-1 1.':>2<10-\ ISZ'(IO-I 0 ISO-IO"I
HzO 0.0600 0.0800 0.IIID3 0.11<0'3 0.11/03 TO,.II-L MqO 0 0 0 0 0 0 0 0 0 0
TOTAL Mo-;:O 5.1"<10"2 5.18 <,o-z :,.\8<\Oz ':>.18 -10-2 S.l8<\O'2 ':>.18 <'0-2 5.18-/0-2 5.18.10'2 0 :>.09.10'2
TOTII-L HC.l 8.04
-------
~
-v
I
+'
I.D
I
FLUE GAS
COOLEg
FILTEg
P~OCESS WA,EI!.
HOLD TMJK
SCIW55E12
EFFLUEt-lT
HOLD TA/J\('
ClARll=lEI!.
PROCE.SS
COIJD\ T lOr-!:::'
7'=>% LIME H'(D£J)ITES It.! S'1'STEM
20 -;. LIME HYDRATES 1t.1 SC.RUBBE.~
40";' LIMESTOfJE SOLIDS DISSOLVE. IN SCIW~BE~
:I~0l. RECYCLED SOLIDS DISSOLVE IN SCI2UBBE.g
NO ':.OLlDS PI2ECIPITATE IN SCIWBBE~
120"F Cll2CUlATHJG liQUOR TEMPERATUI2E
LlMESTQ~E COM?OSITIOt-.1
COMPo WT. FR.
CaO 0.848
MqO 0.152
FLY ASH
COMPo
t.!a1!O
No C.l
ItJSOlUBLE
FL"( ASH
COMPOSI 'T lOt-!
'NT. F~.
0.0011
0.007S
0.9914
90.1. $OZ ~EMOVAL .
50010 SOz OXIDAilOtJ
20% NO)!. RE.MOVA.L
20 MOLE -I. 1'1190 LlMESTOME.
3.35 Las FlY AsH PER LB SUlFU2 It.! COAL
\Soil. iHEORETICAL LlMESTO/JE
~ GAS-SOLID STI2.Et>.MS =E~ UQUID - SOUD STI2Et>.MS
qUl!I.lJTITY (I) (2) ~) (4) (~ au At-J"CrT '( (,,) (7) (8) (9) 4.0) ~~ S/MIN.) 5,"\1 5.,"'" =-.(,.5<: 5,,,"0:,2: <;;,(,.<;; 2 TOTAL FLOW I2I1-,E (lBS!MltJ.) 12,828 12,828 11,387 \,441 1,1'20 321 I~Z 159 199 1'2,8"-9
$OLlOS CO'-1,ENT (G/NM3) \Go.l:>7 IGo'<;;7 0 0 0 SOLIDS COt-JiEt-lT (Wi ".) (,..43 ".74 ".74 (,..74 0.09 30.0 0.59 ,"0.0 0 ':) .~8
GM FLOW eAiE (SCFM) "".925 ",,",925 "'" 459 ""',459 iD9,459 LIQUID FLOW gATE (GPM) 1,~87.1 1,3B(.,.1 1,2~0.9 155.8 1'2<:'-7 2G,.1 18.7 7.4 '24.1 \ 403.4
1EMPERII-TUI2E (oF) 21'S 22S 120 'Z50 'ZSO LIQUID DEIJS\T'1' (LBS/GAL) 8.(,,3 8.100 8.G.0 8.iDO 8.(,,0 8.100 8.<00 8.~0 8.25 8.5~
PRESSURE (AiM) \.000 o.~~o 0.971 O.~10 O....,=>IO O.iD~" O.IO'='iD 0.G.94.10"3 1.54 <10.5 1.'54,10.5 1.'=4 tlO'5 1.'54 <10"3 0 1.'5<: dc"
t-JO" 0 .0005 0.0005 0.0004 0.0004 0.0004 10ill-L co;: 2:2'" rIO"~ 1.'3'hI0"4 15';1" 10"4 IS'" <,0,4 1.""".10"4 1.5':),10'4 1.s~,10-4 1.5~ .10"4 0 \.5<..10'4
0;: 0.0300 O.O~OO 0.0284 0.0'284 0.0284 101AL SO~ 1.40,10"' 1.25.10"' 1.2S,\0" 1.2",10"' 1.2S,,10"' 1.2,:>,,10"' I.<:~"IO"I 1.2<;,10-1 0 1.<:'hIO"
!-.Iz 0.74"25 0.74'2.5 0.7154 O.71'S4 0.71 'S4 10T AL t-Jc: 05 1.4".10" 1.4",10"' 1.4",10.1 1.4"<10.' 1.4"tI0-' 1.4".10"1 \,4",,10'\ 1.4" <10.1 0 1.43.10.'
H c..l 0.0000 0.0000 0.0000 0.0000 0.0000 TOiAL CaO 2.G.4 riD" 2:n.10.t 2.23. 10'Z 2.23r10"Z 2:23 <'O,z 2.23<10.Z (.23,10"2 2:23,10'?' 0 2.Z~.IO'1
1-1,0 0.0800 0.0800 0.1I'S9 0.1159 O.IIS~ i01AL Mq 0 '(.55,10" 2:3~.10' 2.39,10" 2.39<10" '2.3':>.10"' '2.39<10"' 2.39<10'" ?3~.10'1 0 Z.~S.:o'r
TOill.L !JazO S.37dO.t '5. ~'9"IO'l S.~" .10" S.,,':>.IO'7. 5.3':3.10" 5.39,10-' S.~9.10'Z 5.3':1, 10" 0 5.ZEl'IO.l
TOill.L HC.l 8.%
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN. TEXAS 7B758 . TELEPHONE 512 - 454-9535
As process rate data become available, the
computerized process model can be modified to include rate
correlations. This version of the process model will be
invaluable in process engineering S02 control systems.
2.3
Summary and Conclusions
Radian has developed a computerized process model
for the limestone injection - wet scrubbing process. Based upon
the flow arrangement of the "prototype" system, twenty-six
simulation cases have been run. The simulation results are
dependent upon the validity of the stated process assumptions.
The following conclusions have been drawn based
upon these simulation results.
(1)
(2)
(3)
Scrubbing solutions originating from high
calcium limestones are more efficient than
those originating from dolomites. The
magnesium content of the limestone additive
is an important process variable.
Sulfite oxidation reduces the SOa scrub-
bing capability of the wet scrubbing
process. Sulfite oxidation is an important
process variable.
For these simulation cases, the effect
of ionic strength, circulating liquor
temperature, solids precipitation in the
scrubber, and fraction solids in the
filter bottoms stream on scrubbing capa-
bility were relatively minor. The effect
-50-
-------
Radian Corporation
(4 )
(5)
(6)
(7)
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE 512 - 454-9535
of circulating liquor temperature
could be more significant if (1) the
temperature of each process vessel
were varied independently and (2)
reaction rates were considered.
The ionic strengths of process liquors
can vary over a wide range. The result-
ing deviations from ideality must be
considered in the prediction of v-t-s
equilibria and correlation of reaction
rates.
The extent of lime hydration and
dissolution has a major influence on
the scrubbing capability of the
process.
The effect of varying scrubber feed
rate is much more dramatic in the
pure calcium system than in the calcium-
magnesium system.
The liquid-phase compositions of the
various process streams are determined
to a large extent by the s-i, equilibria.
These compositions remain relatively
constant as most model parameters
are varied. The major change occurs as
the amounts of CaG and MgO available
for reaction are varied.
-51-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9S35
(8)
If the wet scrubbing processes were
operated adiabatically, the temperature
variations between process vessels would
be only about ZOF. The effect of the exo-
thermic reactions occurring within the
scrubber and effluent hold tanks would
tend to be counterbalanced by heat losses
from these vessels.
It should be emphasized that these conclusions are dependent
upon the validity of the process assumptions and the number
of simulation cases run. In its present form, the process
model is a powerful tool in predicting the effects of process
variables, estimating process flows and compositions, and
determining the technical feasibility of various processing
alternatives. The model needs to be extended to include rate
correlations and comprehensive energy balances.
-52-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
3.0
APC 0 VENTURI TESTS
Radian Corporation is providing technical assistance
to APCO personnel during pilot-scale SO::) scrubbing experiments
being conducted at the Division of Process Control Engineering
laboratory facilities in Cincinnati. The Radian contribution
has included test plan design, development of sampling and analy-
tical techniques, chemical analysis of scrubber liquor samples,
and interpretation of experimental data. Overall experimental
strategy and requirements are first discussed briefly. Specific
results of the first test series are then presented.
3.1
Test Plan and Obiectives
Several variations of lime/limestone scrubbing
processes are under development and are described in detail
elsewhere (TE-OOS). All of these designs involve sorption of
SOa into an aqueous slurry of lime or limestone. Present
knowledge of pertinent process chemistry and kinetics suggests
that the overall degree of SOa removal may be limited by any of
the following process rate steps:
transfer of SO::) from the vapor to the
liquid phase
hydration or dissolution of the solid
alkaline reactant
oxidation of sulfites to sulfates in
the liquid phase
.
.
precipitation of CaS03/S04. nHaO
-S3-
-------
I~--.
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 7B75B . TELEPHONE 512 - 454-9535
Evidence that any of these steps is extremely fast or slow
relative to the others would greatly simplify design and
interpretation of future scrubbing experiments as well as
design of a commercial unit. Thus, the experimental program
described here was intended to look semiquantitatively at the
relative rates of these important process steps.
It should be emphasized from the outset that the
present test plan was formulated to make use of an existing
pilot scale scrubbing system. Consequently, experimental
results are not necessarily directly applicable to those
commercial scrubbers generally considered as prime candidates
for use in lime/limestone S02 control processes. As long as
appropriate restrictions are recognized, however, some valuable
process conclusion may be drawn.
A simplified flow sheet for this experimental
system is shown in Figure 3.1-1. A more detailed description
of how the apparatus was operated during the Type I experiments
is given in Section 3.3.1.
The test plan as it is presently being carried out
remains essentially the same in scope and objectives as that
described in Radian Technical Note 200-002-5 (Contract CPA 22-
69-138, Jan. 16, 1970). Emphasis in this report is placed
upon experimental and data analysis procedures for the completed
experiments (Type I). Later test plans will undoubtedly be
modified in consideration of preliminary results. Major program
objectives are listed below:
1.
Determine the extent of vapor-liquid
and solid-liquid mass transfer in the
-54-
-------
~
I
V1
V1
I
To I.D. ~~ n ~n( Atmosphere
Scrubbing
Venturi
Water Makeup
or
Clear Liquid Feed
Effluent Hold Tanks
Lime Fly Ash
Solids SO:;! NO NOIii
Dry Dispersion
Venturi
watjrIMake;:arijierS
,
FLOW SHEET -
FIGURE 3.1-1
APCO INHOUSE PILOT SCALE SCRUBBING SYSTEM
~F1ue
'Gas
-------
Radian Corporation
.8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN. TEXAS 78758 . TELEPHONE 512 . 454-9535
venturi scrubbing section, i.e., the
approach to equilibrium.
2 .
Study factors influencing the important
mass transfer steps in the scrubbing
section.
3.
Measure the hydration and dissolution
rate of calcium oxide (and magnesium
oxide) in the effluent hold tanks.
4.
Measure the precipitation
calcium sulfite, sulfate,
in the hold tanks.
rates of
and carbonate
5 .
Gather preliminary information on the
absorption of NOx, build-up of Na+, Cl-,
NO;, and liquid phase reactions of NOx.
6.
Provide realistic conditions for field
evaluation of sampling and analytical
methods.
For convenience of operation and to make maximum use of
information flow during the tests, the program was divided
into three segments or experimental types.
3.1.1
Type I Experiments
The Type I pilot-scale experiments were designed
to evaluate the importance of vapor-liquid mass transfer rates
under conditions that can be related to lime/limestone scrubbing
processes for SO:;a absorption.
-56-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
Application of the two-film theory of mass transfer
to a countercurrent or cocurrent gas absorber leads to the
familiar expression for the height (or length) of the contactor
in terms of the height and number of overall transfer units:
z
=
(HogT.U.)(NogT.U.)
y.
(K ~ a)[J~n
og Yout
(3.l-la)
=
(l-y)log
(l-y)
mean
dy ]
(y-y'k)
(3.l-lb)
""--.
where z is the axial dimension of the contactor, ~ the, molar
gas flux, K a the product of an overall gas phase mass transfer
og
coefficient and an interfacial area per unit of volume, and
y the mole fraction of the component being absorbed. For dilute
gas mixtures as in the system of interest Equation 3.1-1 becomes
3.1-2.
z
=
(K ~ a)
og
Yin
J
Yout
~
y-y"J(
(3.1-2)
For a particular scrubber of length z with a gas flux G, values
of K a can be calculated from experimental measurements of
og
inlet and outlet concentrations of 802 provided equilibrium
data (y';'() are available.
The overa.ll transfer coefficient, Kog can also be
written in terms of individual liquid and gas film coefficients.
Kog
=
1
l/kg+ m/k.e
(3.1-3)
-57-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
where m is a Henry's Law-type constant. These individual
coefficients kg and kJ. are functions of the compositions and
flow rates of the gas and liquid phases. If for example, a
change in the scrubber liquor composition (all other experimental
conditions remaining the same) results in the marked change in
the K (or K a), then the m/k~ term in Equation 3.1-3 is a
og og ""
significant portion of the resistance to mass transfer in the
scrubber. The effects of a change in gas flow rate on k or
g
liquid flow rate on kJ. are not readily observed in most commercial
scrubbers since the interfacial area, a, changes in an unknown
manner. The change that is observed is thus that of the product,
K a, and the effect usually cannot be assigned specifically to
og
the gas or liquid film.
Knowledge of the variation of K a with gas and
og
liquid flow rates and variation of K with phase composition
og
are both useful in scrubber design if vapor-liquid mass transfer
is sufficiently slow to be process rate limiting. If, on the
other hand, another process step, say reactant dissolution is
rate limiting while vapor-liquid transfer is rapid, this knowledge
is of much less importance.
A series of ten Type I experiments was designed
to investigate the effects of gas ra.te, gas tempera.ture, liquor
rate, and liquor composition on vapor-liquid mass transfer in
the pilot scale venturi scrubber system. The specified experi-
mental conditions are outlined in Table 3.1-1. Actual experimental
conditions varied somewhat and are given in Section 3.3.
Water and dilute NaOH solutions were selected as
scrubbing liquors to avoid the problem of precipitation as a
possible rate-limiting step. The highest specified level of
-58-
-------
TABLE 3.1-1
SPECIFIED EXPERIMENTAL CONDITIONS
APCO INHOUSE SCRUBBING EXPERIMENTS - TYPE I
Inlet Gas Comp. (mo1e%)
,Run No. SO~ CO", NOx Inlet Gas Rate (ACFM at T,P) Scrubber Feed Compo Scrubber Feed Rate (gpm)
1 .2 .45 900 at 300°F, 1 atm Tap Water 10
2 . .2 .45 .05 900 at 300°F, 1 atm Tap Water 10
3 .2 .45 .05 1100 at 300°F, 1 atm Tap Water 10
4 .2 .45 .05 900 at 300°F, 1 atm Tap Water 15
5 .2 .45 .05 900 at 250°F, 1 atm Tap Water 10
6 .2 .45 .05 900 at 3000F, 1 atm .12 wt. % NaOH 10
7 .2 .45 .05 900 at 3000F, 1 atm .12 wt.% NaOH 15
8 .2 .45 .05 1100 at 3000F, 1 atm . 12 wt. % NaOH 10
9 .2 .45 .05 900 at 2500F, 1 atm .12 wt.%.NaOH 10
10 .2 .45 .05 900 at 3000F, 1 atm .045 wt.% NaOH 10
I
VI
\0
I
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-9535
NaOH concentration results in a liquor with a pH corresponding
to that expected for the scrubber feed in a limestone injection -
wet scrubbing system operated with a clarifier and filter.
The first experiment shown in Table 3-1 is one at
"base" conditions .except that no NOx is added to the system.
This is essentially a warm-up experiment in which the S02
removal can be measured in the absence of NOx' The second
experiment establishes base conditions for runs with water
as the scrubbing medium. The first two experiments can be
compared to observe the NOx effect if any. The third and fourth
experiments are designed to determine the approach to vapor-
liquid equilibrium at higher levels of liquid and gas flow rate.
The fifth experiment is designed to observe the effect of a
lower inlet flue gas temperature.
The next series of five experiments uses NaOH
solutions as the scrubbing medium. The difference between the
pure water and the NaOH solutions should give an indication as
to whether the liquid or vapor film is controlling in case
vapor-liquid equilibrium is not attained. Gas and liquid flow
rates and inlet gas temperature are varied as in runs 1-5. In
runs 6-9, sodium hydroxide solutions of basicity approximately
equal to that of saturated calcium hydroxide solutions (~ 0.12
wt.% NaOH) are indicated. Run 10 was conducted with a more
dilute solution again to observe the effect of liquor concentra-
tion on vapor-liquid mass transfer rate.
For each Type I experiment, comparison of the
observed value for the outlet SOg partial pressure in the flue
gas with an equilibrium value calculated')'( from the effluent
liquor analyses indicates the degree of approach to vapor-liquid
;'~
All equilibrium calculations were done with a digital computer
using the Radian Corporation routine essentially as described
in the final report for APCO Contract CPA 22-69~138.
-60-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-9535
equilibrium. If the data warrant, a more accurate
of scrubber performance can be made by calculating
transfer units from the definition:
characterization
the number of
Nog
=
Yin
J~
y-y,,\
Yout
(3.1-4)
Analytical requirements for the Type I experiments
include gas phase analyses of S02' NOx, CO2, H20, and O~.
Liquid phase requirements were for determinations of total Ca,
Mg, Na, Cl, K, SOa, CO::!, S03 and N. These requirements and
applicable methods are discussed further in Section 3.2.
The vapor-liquid mass transfer rate data resulting
from the Type I experiments will be of qualitative use only in
predicting behavior of other scrubber systems. Thus, no attempt
to correlate mass transfer coefficients with process variables
is planned. Some of the Type I results which may hopefully be
extended to other venturi scrubber systems are:
approximate ranges of vapor-
liquid mass transfer coefficients,
qualitative effects of gas and liquid
flow rates on these transfer coefficients,
order of magnitude effects of liquor
composition on observed mass transfer
coefficients.
With regard to vapor-liquid mass transfer in scrubber systems
other than the venturi, it is likely that no reliable generaliza-
tion can be made.
-61-
-------
Radian Corporation
8409 RESEARCH 8LVD. ~ P.O. 80X ~~48 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454.~535
The Type I experimental series has been completed
and results are given in Section 3.3.
3.1.2
Type II Experiments
The Type II experiments have not yet been completed.
The test objectives will be outlined here. Type II experiments
were designed primarily to measure the extent of s-j m~ss
transfer in the scrubber section. Here, the objectives are to
determine 1) if a significant amount of lime hydrates in the
scrubber and 2) if a significant amount of hydrated lime (from
the hold tanks) dissolves in the scrubber. For these experiments,
the scrubber system (Figure 3-1) is operated with continuous
solids feed, waste slurry clarification, and clarifier liquid
recycle.
To accomplish these objectives, two kinds of tests
are proposed. Initially a boiler calcined lime/fly ash mixture
is to be added continuously to the flue gas upstream from the
venturi. Saturated liquid effluent from the clarifiers (plus
necessary water makeup) that contains no solids is used as the
scrubber feed. After the system has reached steady state, the
entering and exit liquid and gas streams are sampled and analyzed.
Then, with the liquid flow rate held constant, the limestone
addition is stopped and the gas phase is again analyzed before
the ~iquid composition from the ciarifiers can change significantly.
If there is no increase in S02 concentration downstream of the
venturi throat, it may be concluded that a negligible amount of
CaO (or MgO) from the lime/fly ash additive was hydrated and
dissolved in the scrubber.
-62-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 994B . AUSTIN, TEXAS 7B758 . TELEPHONE 512 - 454-9535
Since the liquid leaving the scrubbing section must
pass through five 55 gallon drums in series before leaving the
clarifying system, the clarified liquid should maintain its
composition for a minimum of about five minutes. This should
enable ample time to make the necessary gas phase analyses.
The major advantage of this type of test for extent of s-~
mass transfer is that only relatively simple gas phase analyses
are required.
The second kind of test involves estimating the
extent of scrubber dissolution of solids that have hydrated in
the effluent hold tanks. To accomplish this objective, a
similar technique is employed. The system is returned to its
original state, i.e., lime/fly ash mixture is added continuously
to the flue gas before the venturi and clear liquid from the
clarifier fed to the scrubber. When the system has reached
steady state, the limestone addition is again stopped and instead
of clear liquid from the clarifier a slurry from the effluent
hold tanks at the same flow rate is added to the scrubber. The
gas phase is then analyzed quickly before the liquid composition
can change significantly. Here, the timing is more critical
since the liquid only must pass through two 55 gallon tanks in
series. However, enough time should be available to perform the
gas phase analyses and the liquid can be monitored by pH probe
to determine if significant changes in liquid composition have
occurred. If a significant increase in SO:;l absorption is obtained
for the slurry recycle case compared to the clear liquid case,
this would indicate that a significant amount of hydrated lime
[Ca(OH):;! and Mg(OH)aJ from the effluent hold tanks has dissolved
in the venturi scrubbing section.
-63-
-------
Radian Corporation'
840'1 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-9535
D This qualitative information rega.rding dissolution
of solid reactants in the scrubber should be useful in planning
future studies. That is, if no reactants dissolve, one would
not expect any significa.nt amount to dissolve in other short
residence time scrubbers. If, on the other hand, some lime
reacts in the scrubber, it will be of value to investiga.te in
more detail variables affecting this rate. Again, however, the
information will be of little or no value as far as application
to widely different scrubber types such as the marble bed or
turbulent contact absorber.
3.1.3
Type III Experiments
The Type III experiments have not been started.
The primary goal of the Type III experiments is investigation
of solid-liquid mass transfer rate~ in the effluent hold tanks.
In addition, vapor-liquid and solid-liquid mass transfer in
the scrubber can be further investigated if this is thought to
be desirable based on results of the Types I and II experiments.
Also of interest in this series is the extent and effect of
buildup of soluble species in the system when the unit is operated
with continuous solids removal and liquor recycle.
With regard to the general applicability of solid-
liquid rate information from this system, valuable information
concerning order of magnitude rates and ranges of rates should
result. More precise correlations that can be extrapolated
with confidence to other systems are not expected. Extremely
difficult analytical problems of separating and characterizing
the individual solid species need to be overcome to extract
such data from a. system using a complex slurry.
-64-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS .78758 . TELEPHONE 512 . 454-9535
3.2
Analytical Chemistry
The experimental plan summarized above requires
sampling and analytical procedures applicable to the complex
liquid and slurry environments encountered in the limestone
scrubbing processes. This section discusses problems associated
with these techniques and preliminary methods chosen for the
analysis of some of the key species, namely S03' SO:;!, CO:;! and
total N. The methods were chosen on the basis of reliability
and low cost instrumentation rather than speed. An in-depth
study of methods more suitable for routine analyses is presently
being carried out at Radian under Contract CPA ?0-143.
3.2.1
Requirements
The requirements for a successful analytical
program are suitable sampling techniques and analytical methods
free from interferences.
o
The process unit can be roughly divided into two
sections, namely into acidic and basic streams. The environment
should be acidic in the scrubber itself and in the lines between
scrubber and effluent hold tank. It is expected that the
solutions circulated in the rest of the system are alkaline.
From these streams representative solid and liquid samples must
be taken. It can be expected that the aqueous phase, at least
in some parts of the unit, is not in t~ermodynamic equilibrium
with the solids. Sampling techniques must therefore be chosen
for which the following are at a minimum:
loss or gain of acidic gases (Sag, COg)
solid-liquid mass transfer during the
sampling procedure
-65-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN. TEXAS 78758 . TELEPHONE 512 . 454-9535
oxygen pick up from the air.
In addition, the samples must
analysis. This is especially
in presence of nitrites.
not undergo any change before the
important for the sulfite analysis
Nitrites can be formed in the scrubber by absorption
of NO and NOa from the flue gas according to the reaction
NO + N02 + H20
~
2HN02
The extent of this reaction as well as the resulting concentration
of nitrites is presently unknown. The nitrites formed can react in
an acidic medium with sulfites. The extent and course of these
reactions is also presently unknown. Nevertheless provisions have
to be made to quench any reaction between sulfite and potentially
present nitrites instantaneously. Analytical methods chosen for the
sulfite determination must be free from interferences from nitrites.
o
After the sampling, sample handling, and storage problems
are solved the actual analysis of the individual species can be
attacked. Here another difficulty is encountered. A quantitative
analytical scheme must be developed for a system for which even quali-
tative analyses are unavailable. The only basis presently avail-
able is the process simulations discussed in Section 2.0. From
the cases described in this section the process simulations PSN2B,
PSN4B and PSN22 have been selected. PSN2B represents the case of
a 100% pure limestone. PSN4B and PSN22 reflect the expected liquor
compositions for a 20% dolomite in different stoichiometries. In
PSN22 enough CaO is available to precipitate Mg(OH)2 in the hold
tanks. PSN4B reflects a case with a buildup in magnesium salts.
The expected concentration ranges for the key species in the
scrubber and filter bottoms are presented in Table 3.2-1. The
simulation PSN2B represents a pure limestone case. The simula-
tions PSN4B and PSN22 represent cases where a limestone containing
-66-
-------
TABLE 3.2-1 ~x Jec :ec S :ream ComJosi :ions ~or ~x :reme S:-.mu.a :ion
. .
::ases
] PSN 2B PSN 4B PSN 22
SCRUBBER FILTER SCRUBBER FILTER SCRUBBER FILTER
BOTTOMS BOTTOMS BOTTOMS BOTTOMS BOTTOMS BOTTOMS
moles/kg mg/kg moles/kg mg/kg moles/kg mg/kg moles/kg mg/kg moles/kg mg/kg moles/kg mg/kg
-a _4 -2 -3 -2 . -4
SO:a 1. 49x10 955 2.75x10 18 1.62x10 1040 1.54xl0 99 1.49x10 955 2.75x10 18
-3 _6 -3 -4 -3 -6
CO:a 2.38xl0 105 4.14xl0 0.2 2.29xl0 101 1059xl0 7 2.38xl0 105 4.14x10 0.21.
-2 -3 772 -1 11200 -1 10000 -2 -3
803 2.44x10 1950 9.64x10 1.40x10 1.25x10 2.44x10 1950 9.63x10 771
-1 -1 15100 -1 -1 -1 -1
N:aOs 1.40x10 15100 1.40xl0 1.46xl0 15800 1046x10 15800 1.40xl0 15100 1040xl0 15100
-1 -1 -2 -01 -1 -1
CaO 1. 61xl0 9030 1.52xl0 8520 2.64xl0 1480 2.23xl0 1250 1.55xl0 8690 1.53xl0 8580
-1 -1 _3 -7
MgO 2.55xl0 10300 2039xl0 9640 4.15xl0 167 9.11xl0 0.04
-2 -2 -2 -3 3340 -2 3200 -2 3210
Na:aO 5.16xl0 3200 5018xl0 3210 5.37xl0 3330 5.39xl0 5.16xl0 5018xl0
HCl -2 -2 -2 -a -2 -2 2950
8.04xl0 2930 8.07xl0 2940 8.36xl0 3050 8.39xl0 3060 8.04xl0 2930 8.08xl0
I
~pH 4.06 11.3 4.08 8044 4.06 11.3
I
I
~
"
I
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE 512 - 454-9535
20% MgO is used as the sorbent. All three cases are based on
the use of 1.5 times the stoichiometric amount of sorbent. In
case PSN2B and PSN4B, 75% of the sorbent was assumed to be
utilized in the system. In case PSN22 90% was assumed to be
available for reaction in the system. Soluble sodium, compounds
of nitrogen, and the chlorides do not show much variation in
concentration. The concentrations of species which form insoluble
salts vary more extensively. SO~ varies from 18 to 1040 mg/kg
water, 'COs from 0.2 to 105 mg/kg H~O, S03 from 771 to 11200 mg/kg
H~O, CaO from 1250 to 9030 mg/kg HaO and MgO from 0.04 to 10300
mg/kg H20. The analytical methods ultimately chosen for the
analysis of the process must be applicable to these extreme limits.
3.2.2
Methods
In this section preliminary results for sampling
using porous frits and a continuous centrifuge are described.
Both methods were applied to saturated and supersaturated solu-
tions without crystallization nuclei. The methods have not yet
been tested on actual supersaturated solutions containing solids.
Preliminary analytical methods for the analysis of SO~, S03,
total N, and CO2 are also discussed.
3.2.2.1
Liquid and Solid Sampling from Slurries
A continuous centrifuge (Sorvall, Model SS4) was
chosen for a quick solid-liquid separation. This instrument is
based on the principle developed by Szent-Gyorgyi and Blum.
The centrifuge allows the collection of a sediment from a slurry
with the supernatant liquid being continuously exhausted from
the system. The principle of the centrifuge is shown schematically
in Figure 3.2-1. The shaded part of the centrifuge does not
-68-
-------
r--- ----
CLEAR SOLUTION""
- SLURRY
I l '
+
~
THE SHADED PARTS DO NOT ROTATE
FIGURE 2-1 - PRINCIPLE OF CENTRIFUGE (SCHEMATIC) SORVALL MODEL SS4
-69-
~
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
rotate. The slurry enters on the right side of the top, descends in
the inner tubing and is finally distributed into the rotating sam-
pling tubes. In the tubes the solids and liquids are separated by
centrifugal force. The clear liquid ascends and is pumped out of
the system.
The SS4 model can be
With the SS34 rotor fields up to
velocities for particles greater
10 em/see under these conditions.
operated at speeds up to 17500 rpm.
37000 g are developed. The settling
than 1 micron are greater than
The sampling tUDes can be filled from 1/2 to 2/3 full
with an unreactive organic liquid which is immiscible with water.
Carbon tetrachloride, CC14, is a suitable compound. The specific
gravity is 1.6 g/cm3 which is between the specific gravity of water
and of the solids. Figure 3.2-1 shows the principle of this separa-
tion technique. This method has a unique advantage in that the
particles are no longer in contact with the slurry once they have
entered the organic phase.
The flow rates through the whole system depend upon the
speed of rotation and the pressure of the liquor. For example, flow
rates of 500 ml/min at 16" head and 11000 rpm and 1000 ml/min at 43"
head and 17000 rpm were measured. When two tubes two-thirds filled
with CC1, are used, the residence time of the fluid is one and two
seconds for the described cases.
In order to determine whether supersaturated solutions
would tend to form crystallization nuclei due to vibrations in the
centrifuge, supersaturated sulfite and sulfate solutions were pre-
pared. Samples were analyzed for calcium before and after centrifu-
gations using an EDTA titration and for sulfite using an iodometric
technique. The results of the analyses were well within the accuracy
limits of the methods of analyses employed. The consumption of 0.1
-70-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
molar EDTA was
sample (100 ml
in the sulfite
was 0.7%.
22.7 ml for a filtrated and 22.65 ml for a centrifuged
aliquots) in the sulfate,;experiments. The difference
content of a filtered and' centrifuged sulfite sample
Next, an equilibrated calcium carbonate slurry (1%)
was prepared from very fine powdered CaC03. Samples were taken
(1) by filtration through Number 5, Geniune Watman filter paper
in a closed system, (2) by filtration through a 7 micron stainless
steel filter probe, and (3) by centrifugation. The following
table shows the results.
Sample
mg CO!:l/liter
.
1.
2.
3.
4.
5.
Filtered through
Centrifuged
Filter probe 7~
Filter probe 7iJ
Centrifuged
filter paper
7.4
7.8
8.0
8.1
8.2
The experiments were conducted in the same time sequence as
reported above. The slurry was stirred with an electric motor
and a stirring rod during the time the samples two through five
were taken. CO\;! upt.ake from the air can well be responsible for
the slight increase of the carbonate content.
In summary, the centrifuge shows the following features:
1.
Flow rates vary between 500 to 1000 ml/min.
2.
Residence times are in the order of a few
seconds for the case in which two tubes
filled two-thirds full with CC1, are used.
I-
\
I: ,
.,
-71-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE 512 . 454-9535
3 .
No sP9Rtaneous nucleation was
observ~a when supersaturated
calcium sulfite and calcium sul-
fate samples were centrifuged.
4.
Operating at 17,000 rpm, extremely
fine particles are separated.
5.
Representative separated solid and
1jquid samples can be obtained using
the CC14 technique.
The separation technique described above has not
yet been tested in the field.
3.2.2.2
Sulfite Determination
The method outlined here gives accurate results in
the presence of nitrites. The expected concentration of nitrites
has not yet been established, however, the presence of nitrites
must be anticipated.
Nitrites and su1fites are known to react at pH
values of about four or less. This experimental fact prohibits
the use of the iodine-thiosulfate method. Interference problems
from nitrite can be avoided by titration of the sulfite with
iodine at a pH 6.0-6.2. The iodine solution is generated as
needed for each determination using standard iodate solution
and excess iodide ion at low pH (1-2). This method is more
convenient and more reliable than using standard iodine s~lu-
tions. The sys tern is buffered to pH 6.0-6.2 after this /~t;:~p
wi th an acetic a.cid - sodium acetate buffer. The reaction
rates for the sulfite-nitrite and nitrite-iodine reactions are
-72-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE 512 . 454.9535
slow in this pH-range,
oxidation by iodine is
with arsenite solution
detection.
while the reaction rate for sulfite
fast. The excess iodine is back-titrated
using a dead-stop technique for end point
The dead-stop apparatus consists of two platinum
electrodes (1 cm x 1 cm, 1 cm apart), a 1.5 volt battery, a
voltage divider (1.5V to O.lV) and a microammeter with a range
0-20 ~A. One electrode is connected to one terminal of the
voltage divider. The meter is placed in series with the second
electrode and the other voltage divider terminal. Thecurrent
stays fairly constant at the beginning of the titration and
shows a sharp drop at the end point.
For samples containing between 5xlO-s and 5xlO-4
moles sulfite the following convenient procedure was developed.
To 20 ml KI solution (50 g KI/ltr), two ml of a
0.0833 M KIOs and 2 ml of N HCl are added. The iodine genera-
tion is complete after 15 seconds. Then 175 ml of a pH 6.0
buffer solution are added (1 mole/liter sodium acetate, 0.05
mole/ltr acetic acid). Next, the sample is introduced and the
excess iodine back-titrated with 0.01 mole/ltr sodium arsenite
solution using a 50 ml buret and the dead-stop apparatus.
The concentration of total SOa in the sample can
be calculated using the following equation:
C
=
(B-S)M [mole/ltr]
V
Where:
C
=
concentration of total SO~ (mole/liter)
-73-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
B = volume in milliliters of arsenite
solution needed to titrate the blank
S = volume in milliliters of arsenite
solution needed to titrate the
sample
M = molarity of the arsenite solution,
mole/liter (normally 0.0100)
V = volume of sample used, milliliters.
Using a NO;:SOa mole ratio of 50:1, sulfite
determinations with a 1-3% error have been conducted using
this procedure. Five determinations of KgS03 without nitrite
added gave a relative deviation of 0.25%.
3.2.2.3
Sulfate Determination
The method described here is similar to the ASTM
referee method D 516 (1968) based on the precipitation of sulfate
as BaS04. In order to avoid the time-consuming gravimetric
determination of the amount of BaS04 precipitated, a back titra-
tion method was chosen.
The sulfite (SOa) present in the sample solution is
oxidized to sulfate with hydrogen peroxide. Interfering cations
(including Fe3+ and A13+ as well as Ca++ and Mg++) are removed
by means of a cation exchange resin (Amber1ite, E.G. 120 A.R.
100-200 mesh). The sample containing 0.15 to 0.40 m moles total
sulfur is acidified and a known excess of barium chloride
-74-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
(10 ml 0.06 M BaCIa) is added. The barium sulfate precipitate
is allowed to digest overnight. The excess barium is titrated
with EDTA (0.01 M) using a 50 ml buret in a buffered solution
containing magnesium - EDTA complex and using Eriochrome Black
T as the indicator.
The molarity of total SOa+ S03 as S03 is
so L-m?lesJ
3 l~ter
=
MBVp,- MEVE
VA
Me
Ve
ME
V£
VA
=
molarity of BaC12 solution (moles/ltr)
volume BaCIa solution used (ml)
molarity EDTA solution (moles/ltr)
volume EDTA solution (ml)
volume of aliquot taken (ml)
=
=
=
=
The sulfate content in the original sample (before sulfite
oxidation) is calculated by difference from the value found
above and the sulfite content determined by the method described
under 3.2.2.2. The accuracy of the method is within one to two
percent for samples containing about 0.4 m moles sulfate.
Supporting data are given in Table 3.2-2.
302.2.4
Total Nitrogen Determination
aqueous
are, as
The amount and kind of nitrogen oxides in the
phase of the limestone injection wet scrubbing slurry
mentioned earlier in the report, presently unknown.
-75-
-------
TABLE 3.2-2
SULFATE ANALYSIS ERROR
m moles m moles m moles
m moles m moles m moles m moles SO: Added SO:' Equivalent !ota1 m moles
SAMPLE Not- Ca* Mg++ NO; C1- as ~ S04, Added as ~ 503 SO; Added SO: Found ERROR %
8-4-1 4.0 .20 10 8.4 .250 .190 .440 .440 0.0
8-5-2 4.0 .20 10 8.4 .250 .123 .373 .378 + 1.3
8-7-1 4.0 .20 10 8.4 .250 .160 .410 .414 .98
8-7-2 4.0 .20 10 8.4 .250 .147 .397 .398 .25
8-11-1* 4.0 .20 10 8.4 .250 .170 .420 .422 .48
8-11-2** 4.0 .20 10 8.4 .250 .133 .383 .381 + .52
* Sample # 8-11-1 also contained 8.11 m moles NO; and .06 m moles CO;
** Sample # 8-11-2 also contained 4.29 m moles NO; and .05 m moles CO;
+ Original sample volume approximately 20 m1
I
......
0\
I
-------
Radian Corporation
84~ RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
A first step for gaining insight into this problem
area was the choice and testing of a method to determine total
nitrogen in solution. Then measurements could be made to
determine if the nitrogen oxide loss from the gas phase could
be accounted for in the liquid phase. The next step will then
be the qualitative and quantitative determination of the
nitrogen compounds.
The method chosen was a reduction procedure in the
alkaline region using Devarda's alloy (50% Cu, 45% AI, 5% Zn).
The end product of the reduction is NH3 which is distilled
(with some water) into a solution of known HCl content. The
excess HCl is back titrated with standard NaOH using methyl
red as color end point indicator.
'"
The apparatus consists of an evolution flask with
thermometer, a cyclone for eliminating the carry-over of fine
alkaline mist, a condenser, absorption flask, burner and a 50 ml
burette. To a sample containing 5 to 25 mg total nitrogen,
6 g Devarda's alloy and 100 ml 2N NaOH are added. Distillation
is begun after the first vigorous reaction ceases. The amount
of Devarda's alloy needed to carry out a complete reduction is
greater than the amount needed for normal determinations, since
part of the atomic hydrogen reacts with the sulfite to form HgS.
Tht total N in solution is calculated by
t tIN [moles]
o a liter
=
(VHCl' NHCL - VNaOH' NNaOH)' 1
Vsample
-77-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
VHCl
=
ml HCl used in the absorption flask
VNaOH
=
ml NaOH used in the back titration
Vsample =
ml sample analyzed
NHCl
=
normality of the hydrochloric acid
NNaOH
=
normality of the sodium hydroxide
solutions
VHCl
is preferably equal to 50 mland NNaOH
= NHCl = 0.05N.
The accuracy of the above method is
samples containing 15-25 mg total N and up to 1
The following list gives supporting data.
within 1% for
g sulfite.
Exp. ~ 803 (g) NaN03actua1 NaNOs found % Error
1 1 112.8 113.0 -0.2
2 1 152.6 153.4 -0.5
3 1 150.8 151.1 -0.2
4 1 163.5 164.9 -0.9
-78-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN. TEXAS 78758 . TELEPHONE SI2 . 454.9535
3.3
Experimental Results
During the period covered by this
Type I experimental series was completed.
discussed below.
report only the
This test series is
3.3.1
Experimental Equipment and Procedures
A schematic diagram showing the essential features
of the pilot-scale scrubbing equipment as it was operated for
the Type I runs appears in Figure 3.3-1. A detailed sketch of
the venturi scrubber itself is shown in Figure 3.3-2.
Flue gas was produced by burning natural gas in an
adjacent incinerator. In order to supply sufficient volume to
the scrubber, however, this stream was diluted with a large
portion of air. The CO~, HaO, and O~ content of the resulting
gas are thus not representative of a power plant stack gas.
The significance of this will be discussed later. The tempera-
ture and humidity of the gas could be adjusted by prespraying
with water at this point.
SO~, NO, and NOa were added to the scrubber inlet
gas just after the incinerator from three separate cylinders.
Approximately 2000 ppm SO; was used in all runs, and 500 ppm
of 90% NO, 10% NOa in all except 1 and I-A. The gas stream
then flowed through a dry mixing venturi (not shown) and into
the scrubber where it was contacted with the feed liquor.
Downstream from the scrubber, the gas entered a knockout drum
where most of the entrained liquor was removed. The gas was
then discharged to the atmosphere by a centrifugal fan. A flow
(ASTM) venturi (FI-l) was located in a straight run of duct
following the I.D. fan. Gas flow rates were calculated from the
observed pressure drop within an estimated error of ! 5%.
-79-
-------
~
T\V
I.D.Fan
I
00
o
I
Knockout Tank
.'TI"
7:Y
Scrubber Effluent
~c15c1?~
il'~T'o'r20"T34'
8
o
.~
3 '
.....V
Panel Mounted Instrument
for sample point
Local Instrument or analysis
TR = Temp. Recorder
PI = Pressure Indicator
FI 1::1 Flow Indicator
~S~
Flue
Venturi Scrubber
FI
2
Scrubber Feed
FIGURE 3.3-1 PILOT SCALE SCRUBBING EQUIPMENT
APCO SCRUBBING EXPERIMENTS - TYPE
I
Gas
NO
N02.
-------
I
(X)
.....
I
T
..
~
l
I ~
10°
-L
-//1"
8
Material:
~~
)"
1/8
14 Ga. 304 S.S.
30°
_L
,
~
30° Cone
~ (Adjrstable) - -
If-~~ -I I
I I
I I
II'
I I
1'1
i I
I ,
! I
5"
--4"
i
I ~ 3"
I I 4
Flue Gas
E::
Liquid Feed Pipe
FIGURE 3.3-2 Pilot Scale Venturi Scrubber
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
.The scrubbing liquor was pumped from a well mixed
55 gallon drum into which the proper amount of concentrated
NaOH solution was metered and continuously diluted with tap
water. The scrubber feed flow rate was measured by a calibrated
venturi within + 5%. A sample of the inlet liquid for each
run was analyzed for total amounts of S03, COg, Ca, Ca+Mg,
K, Na, Cl, and N. Corresponding effluent liquor samples were
analyzed for total SO;, S03' N and COg and in some runs also
for Ca, Mg; Na, and K. Analytical methods used were discussed
in Section 3.2.
The temperature of the liquor in contact with the
exit gas stream in the knockout drum was at first assumed to be
equal to that of the gas at sample point 7 (slightly above the
knockout drum) 0 After it was noted that the accuracy of this
arrangement was inadequate for purposes of equilibrium calcula-
tions a thermometer (TI 7-L in Figure 3.3-1) was placed in the
knockout drum.
Gas sampling points were located as shown in Figure
3.3-1. Whittaker SOg and NOx analyzers were manifolded such
that each of the sample points could be selected and analyzed
consecutively after the desired experimental conditions were
achieved. NOx is taken to mean NO+NO~. Manufacturer's specifi-
cations for the SOg analyzer call for an accuracy of + 2% of
full-scale (+ 60 ppm). The instrument was calibrated against
a standard gas (SOg in nitrogen) once a week and appeared to
operate well. Some additional error was probably introduce9 by
uncertainties in the calibration gas composition so that + 80
ppm is a more reasonable estimate of accuracy. The specified
-82-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN. TEXAS 78758 . TELEPHONE 512 . 454.9535
accuracy of the NOx analyzer is also + 2% of full-scale. Inlet
09 and COg concentrations were determined by gas chromatograph.
Temperature and pressure were also recorded for each sample
point.
3.3.2
Treatment of Data
Table 3.3-1 summarizes actual experimental conditions
for the Type I test series. Tables 3.3-2 and 3.3-3 give the
results of chemical analyses of the scrubber liquors. Using
these data a number of calculations were made for each run.
Sample calculations for Run I-4 are given below:
3.3.201
Water Evaporation in the Scrubber
The amount of water evaporated in the scrubber must
be estimated in order to calculate exit gas and liquid flow rates
(G and L) and concentrations. This was done by assuming that
the flue gas stream was saturated with water in the scrubber.
The humidity of the inlet flue gas was taken to be the same as
that of the ambient air since the flue gas was diluted by a
large factor. For purposes of calculation, the ambient humidity
can be assumed constant at ~ .01 1b HgO/1b dry gas (.02 for
Runs 5, 9 and 9A because of prespraying) without introducing
significant error.
First the experimental value of inlet gas flow rate
(Gin) must be converted to units consistent with the humidity
chart used (lb dry gas/min). Converting to lb. total gas/min,
Gin(lb/min)
=
492°R 1 lb. mole
Gin(ft3/min) x T. x 359 ft3 x mole wt. of gas
~n
(3.3-1)
-83-
-------
TABLE 3.3-1
EXPERIMENTAL CONDITIONS
APCO INHOUSE SCRUBBING EXPERIMENTS - TYPE I
Outlet
Scrubber Scrubber Indicated Corrected Indicated Corrected Flue
Inlet Gas Rate Scrubber Feed Effluent Outlet Outlet Outlet Outlet Gas
Inlet Gas Compo (mole '7.) Feed Rate /lP("HaO) Temp. SOa SOa NO, NO, T e:::p .
Run No. ~ ~ ~ ..£L. !!.a£ (ACFM at T.1 atm) Compo (GPM) Scrubber (TOF) (mo 1e '7.) (mole %) (mo Ie %) (mole %) -C:U
1-1 .198 .85 ---- 19.6 2.4 1040 at 300°F Water 10 8.75 92 .174 .1645 110
1A .200 .6 ---- 20.6 2.3 1090 at 320°F Water 10 8.6 92 .172 .163 110
2 .211 1.2 .054 17.7 4.0 1020 at 310°F Water 10 7.8 92 .190 .178 .052 .049 115
3 .205 .9 .056 19.6 2.5 1230 at 315°F Water 10 12.0 92 .159 .150 .054 .051 110
3A .199 .4 .051 18.8 2.3 1210 at 300°F Wat:er 10 11.4 92 .161 .156 .049 .047 100
3B .290 1210 at 300°F Water 10 12.6 84 .2475 .246 85
I 4 .206 1.3 .044 18.8 2.5 1010 at 310°F Water 15 9.0 92 .166 .157 .044 .042 110
00 5 .197 .6 .039 20.1 4.2 1010 at 260°F Water 10 8.5 92 .168 .159 .037 .035 110
.p-
I 6 .200 1.1 .044 19.E 2.5 1070 at 310°F .0282M NaOH 10 8.4 92 .090 .085 .039 .037 110
6A .196 1l:!0 at 300°F .0416M NaOH 9.9 9.3 86 .0855 .085 90
7 .198 .7 .051 19.2 2.7 1050 at 310°F .0365M NaOH 15 9.2 95 .049 .046 .047 .044 115
8 .246 .5 .049 20.1 1.6 1270 at 325°F .0370M NaOH 10 12.4 92 .ll8 .12,+ .048 .047 100
SA .202 1255 at 300°F .0349M NaOH 9.9 12.4 94 .100 .098 94
9 .239 .3 .048 19.8 1.2 1060 at 260°F .0393M NaOH 10 10.6 90 .126 .124 .048 .047 95
9A .200 1115 at 250°F .0280M NaOH 10 9.5 72 .110 .110 80
10 .239 .10 .038 19.2 .6 1090 at 315°F .0190M NaOH 10 8.6 92 .139 .133 .038 .037 105
lOA .2075 1100 at 300°F .0116M NaOH 9.7 8.7 84 .142 .142 84
-------
r ~A ~ ~ ~ 3.3-2
CHEMICAL ANALYSES OF SCRUBBER FEED
-------
----- ---------- - ------- --- ----
TABLE 3.3-3
CHEMICAL ANALYSES OF SCRUBBER EFFLUENT
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
The molecular weight of the flue gas is given by
Mole Wt.
= L Yi
i
mole wt.i
(3.3-2)
.th f h
where y. is the mole fraction of the ~- component 0 t e gas.
~
Thus, for Run 1-4,
G. (lb total gas/min)
~n
=
1010 f~3 x 492°R
m~n 770uR
x
1 lb. mole
359 ft3
28.7 lb.
x 1b.mo1e
=
51.6
(3.3-3)
Since the inlet humidity was assumed to be .01 lb.
H:;IO/1b dry gas, the flow rate of dry gas, Gd is now
ry
Gd (lb. dry gas/min)
ry
=
Gin(lb. total gas/min)/1.01
=
51.6/1.01
=
51.1 lb. dry gas/min
(3.3-4)
Now, the amount of water evaporated in the scrubber
can be calculated using a humidity chart. The humidity (H) of
saturated flue gas at the scrubber outlet temperature (110°F)
is found to be .06 lb. H90/1b'dry gas so that
-87-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
Wa.ter evaporated
=
(Hout- Hin) x Gdry
=
(.06 - .01) x 51.1
=
2.55 lb. H:aO/min
=
1.2 liter/min
(3.3-5)
Finally this amount
the inlet liquor rate and added
obtain the corresponding outlet
of water is subtracted from
to the inlet gas rate to
rates.
G
out
=
G. + 2.55 lb. H:aO/min
~n
=
51.6+2.55
= 54.2 lb. gas/min (3.3-6)
and
L = L. - 1.2 liter/min
out ~n
= 56.7 - 1.2
= 55.5 liter/min (3.3-7)
Concentrations for chemical species measured in the
feed liquor only (for this run, Ca, Mg, Na, K, and Cl) can now
be corrected by a factor of 56.7/55.5 if effluent liquor con-
centrations are desired.
-88-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
3.3.2.2
Flue Gas Sampling Correction
A small correction based on water condensation
must be made for the S02 and NOx gas analyses. Since the flue
gas instruments analyze the gas at ambient temperature, some
water condenses in the sample lines. Assuming that the
analyzers operate at 80°F, the amount of water condensed in
the sample lines would be
Rin - Rout
=
.06 - . 023
=
.037 lb. Rg07lb. dry gas
(3.3-8)
The appropriate correction factor must be based on molar
quantities. The molar gas flow rate in Run 1-4 is 1.89 lb.
mole/min. Cooling to 80°F would decrease this by the water
condensed which from 3.3-8 (after suitable conversion) is .10
lb. mole/min. Corrected values are thus
sag t
au
=
.1660 x ~:~~
=
.1570 mole %
(3.3-9)
NOXout
=
044 1.79
. x 1:89
=
.042 mole%
(3.3-10)
-89-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE 512 - 454-9535
3.3.2.3
Total Sulfur Material Balances
Using measured liquid and gas flow rates and
compositions, a total sulfur material balance was made for each
run. That is, the indicated amount of sulfur removed from the
flue gas is compared with that measured in the scrubber liquor
as in (3.3-11).
S(g) in - S(g) out
1
S(~) out - S(~) in
(3.3-11)
The amount of sulfur removed from the gas is the difference of
the products of inlet and outlet compositions and flow rates.
For Run 1-4,
(1.79 lb. mole gas/min x 2.06xlO-s mole fracto SOg)in
- (1.89 lb. mole gas/min x'1.57xlO-S mole fracto SOa)out
= 7.2xlO-4 lb. mole total S/min removed from the gas
(3.3-12)
Similarly for the scrubber liquid,
(55.7 liter/min x 5.57xlO-s g mole/liter)out
(56.7 liter/min x 1.05xlO-s g mole/liter)in
=
0250 g mole total S/min appearing in the liquor
=
5.5x10-4 lb. mole total S/min
(3.3-13)
-90-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN. TEXAS 787S8 . TELEPHONE 512 . 454.9535
These are compared by converting to similar units and calculating
a material balance closure error using Equation 3.3-14.
Material Balance Error
=
6S(g) - 6S( i,)
6S ( i,)
=
7.2x10-4 - 5.5x10-4
5.5x10-4
=
+ 31%
(3.3-14)
3.3.2.4
Total Nitrogen Material Balance
Because of the small difference in inlet and outlet
NOx and insufficient sensitivity with both the gas and liquid
analyses, a total N material balance was not calculated. Indi-
cated NOx removal based on both gas and liquid phase analyses
was quite small (see Section 3.3.4.2).
3.3.2.5
Extent of SOg Oxidation
The fraction of absorbed S02 that
calculated for each run from the liquid phase
total S analyses. In Run 1-4 for example:
was oxidized was
total S02 and
Fraction Oxidized
=
,
1 - SOto~~~n~ ;gs~~6~~r
=
1 - .004 mole SOg/liter x 55.5 1iter/min
.250 g mole total S/min
=
.11
(3.3-15)
-91-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 99~8 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . ~5
-------
O~ OCT 1C
TABLE 3.3-4
T EMPC"RATURE
.3:56:5..99&
IN PU T
MCLES
S02
C02
S03
NA20
APCO INHOUSE EXPERIMENTS
TYP E I, NO.4
= 4.00 DC 0- 03
;; 1.43000- 03
= 1.51000- 03
=&.73000- 04
N2.05 =1.43000-04
HCL =2.04000-03
=1.34000-03
=5.550G2+Ql
MGO
=4.43rOO-04
CAO
H20
~aUE OUS S 0 LU T ION EaUILIFSRIA
c c ~., p 0 N E NT MOLALITY ~ CT IV IT Y A CT I V IT Y f"EFFIClfNT
H20 '3.997-:)1
H+ 3. 4 35- C 3 3 . 1 15- 0 3 3.07C-r,1
OH- 6. h 7 S- 1 2 ::.989-12 3.372-01
H5 0 3- 3.131-03 2.858-03 8.982-;j1
S03-- 7.[,72-08 ~.394-0B S.51C-f'l
5014-- 1.09<)-03 7.011-Cu t,.381-n1
H CO 3- 2.4S2-07 2.2 3C- 07 a.9F!2-G1
C03-- 5 . 9 97- 1 S 3.<304-15 6.510-01
NO 3- 2.8c:;3-04 2.534-04 8.883-1)1
HS 0 (j- 3.007-04 2.598-04 8.372-(')1
I H 2$ 0 :3 8. 1 .C3 8- 04 8.203-04 l.ob2+00
\0
u,) 42C03 1 . 4 3 0- 0 3 1 .4 33- G 3 1.002+80
I
C A+ + 1.214-03 7.94S-C4 E..543-fn
C AOH+ 1 . :n c- 1 3 1.193-13 8.172-01
CAS 0 3 1.101-07 1.1 D3-07 1.C\J2+CO
CAC03 5 . 4 3 C- 1 5 5.448-15 1.002+1]0
CAHC03... 3.82~:-09 3.431-09 '3.972-01
CAS 04 1.?52-Q4 1.2SS-04 1.'J02+GO
C~N03+ e.. 0 1 0- 07 5.393-07 8.'372-01
MG++ 4.0 ;? 2- 04 2.[,10-04 6.l4eC'-Ol
'''GOt-! + 7 . 4 I) 'S- 1:3 6.701-1~ 8.972-('1
M GS 0 3 1 . 1 38- 03 1.190.-08 1.002+00
MGHC03+ 6. 1 ') 4- 1 0 1:,.557-10 i3,<;'!72-n1
i1 GS a &4 If . 0 ? f,- C 5 4.034-05 1.002+(10
r'lGC03 2.827-15 2. 8 3 2 ~ 1 5 1.002+[10
Nh 1.341-03 1 .208- 0 3 '3.0C3-fJ1
N~OH 1.343-15 1.'347-15 1.1J02+GO
~ 6 C 03- 1. D 37- 16 9.305-17 13.972-01
NAHC03 1.51 Z- 10 1 .514- 10 1.002+00
N AS 0 4- 5.219-06 t4.€82-0G 8.972-01
33.330
DE G. C
-------
TABLE 3.3-4
N 4NO 3
Cl-
PS02 =~.g9802-C4
PC02 =5.18249-02
PH =
I
\0
~
I
(continued)
1.21r..-07
2 . 0 4 0- 0 3
4T M.
HM.
2 .506
1.21?-07
1 .3 3 Q- a !:
MOL E CU L A R ~r it T E R = 9. 99 <:11 5- 8 1 K G S .
IONIC STRENGTH = 1.812~4-02
1.OG2+!lO
a.9'=-8-nl
Page 2
TYPE I, NO.4
RES. E.N. =
-1.307-0'3
-------
I -- - --
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE 512 - 454-9535
3.3.2.7
Number of Overall Gas-Phase Transfer Units
A quantitative comparison of the results of these
experiments requires that a number of transfer units be
calculated from the data for each run using the definition
Nog
=
Yin
J
Yout
~
y-Y';'(
(3.3-16)
If the equilibrium partial pressure (y>;'() does not vary linearly
with the actual partial pressure (y) over the range from scrubber
inlet to outlet, this integral cannot be evaluated analytically.
Since y* can be calculated for any liquid composition using the
Radian subroutine, however, integration using the trapezoidel
approximation is easily done by computer.
For purposes, of the calculation, the scrubber was
divided into a number of increments of equal S02 transfer. Over
each increment, the vapor and liquid concentrations of total S02
were calculated exactly by material balance from the initial
(inlet) concentrations. Next, assumptions must be made concern-
ing a) the relative transfer rates of NOx, COa and H20, and
b) the rate of oxidation of S02 to S03. These affect the liquor
composi tion and thus, y>;'~. In the calculations presented here,
it was assumed that
all water is transferred (evaporated)
in the first increment. This assumption
is supported by experimental data show-
ing very rapid drop in the gas temperature
between the first two gas sample points.
-95-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN. TEXAS 787S8 . TELEPHONE SI2 . 4S4.9535
NOx and CO~ are transferred at the same
relati ve rate as 8°9'. That is, after
a given fraction of the total absorp~ion
of 80~ has occurred, an equal fraction
of the total changes in CO~ and NOx con-
centration are also assumed to occur.
the oxidation rate of SO~ to 803 is
proportional to the concentration of
of 80~ in the liquor. This assumption
is somewhat arbitrary, but the result-
ing values of N should still be a valid
og
comparison between runs for this scrubber.
The program calculates the appropriate incremental changes in
the liquor and gas compositions and a value of l/y-y* after
each increment. The incremental areas, 6Y/(y-y*) , are then
avg
summed to yield Nog.
3.3.2.8
Relative Kga
From the integrated expression for a gas absorber
(3.1-2) it can be seen that for a given scrubber length,
a/Kga is inversely proportional to the number of transfer units.
Comparisons of K a must be adjusted for ~, the molal gas flux.
g
Relative values of Kga were calculated from this relationship.
That is
-96-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
(Kga)relative
-
(K a)
(Kga )ref
=
Nog x (;'
~ ref ~ref
(3.3-17)
The reference value of Kga was arbitrarily taken to be that of
Run 1-1.
3.3.3
Experimental Error Propagation
Errors associated with measured and calculated
quantities are an important aspect of any experimental study.
Where possible, the statistical significance of results and
conclusions should be examined. One approach to this error pro-
pagation problem is to estimate the variance, a2, of key
experimental measurements (assuming errors are normally distri-
buted). Then, the variance of calculated quantities can be
expressed as a function of these individual estimates. For
example, given a function of a number of variables (measurements)
f
f(Xl,X2,X3, ... XN)
(3.3-18)
if S~ is the estimated variance for measurement x. then the
~ ~
estimated variance of the function will be
sa
f
=
n
\'
~
i=l
(~)2 S~
ox. ~
~
(3.3-19)
-97-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
Use of Equation 3.3-19 of course implies knowledge
of the distribution of the random error for each measurement,
xi. Since the Type I experiments were run under "shake down"
conditions, detailed information regarding the precision of
experimental measurements is not, in fact, available. A
statistical treatment of error in this case is not justified.
It may even be regarded as dangerous in that rough estimates
of experimental error would seem more rigorous when expressed
in statistical terms.
Some treatment of error propagation is nevertheless
useful in aiding engineering interpretation of the results of
these experiments. For this purpose, "limits of error" will be
somewhat arbitrarily assigned to the experimental measurements
which were made. The propagation of these limits though
calculations can then be expressed by a differential approxi-
mation. Given a function of interest, for example
g
=
g (Xl' X::p X3'''. XN)
(3.3-20)
if b.Xi is the estimated limit of error for measurement xi and
6g the limit of error for g then
n
= I ~~~
i=l ~
dg
dx.
~
(3.3-21)
6g2!
n
\' logi
Lox.
. 1 ~
~=
b.Xil
(3.3-22)
-98-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
It should be recognized that this expression does
not take into account the tendency for randomly distributed
errors to cancel out in calculations. Thus, the resulting
error estimate for g is likely to be greater than a statistically
expected error. Also, the limits of error, Lx., used in this
~
calculation are. not estimated in any rigorous faBhion but merely
represent the engineer's experience and judgment that a certain
measurement "could not have been wrong by more than this much".
The most. important experimental measurements from
the standpoint of data analysis and interpretation of results
the following estimated limits of error:
have
inlet and outlet flue gas SOa analyses
- + 80 ppm
inlet and outlet flue gas flow rates -
+ 5%
inlet and outlet liquid rates - + 5%
inlet and outlet total SO~ (liq.)
analyses - + 2%
inlet and outlet total S (liq.) analyses
- + 2%
These estimates may be used to establish limits of
error for calculations involving these data.
-99-
-------
.
.~~~ - ".-
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE 512 . 454-9535
3.3.3.1
Material Balance Error
Material balance calculations were made
to check the consistency of the data for each run.
phase calculation,
(see 3.3.,2.2)
The gas
S(gas)
=
Gi Yin - Go Yout
(3.3-23)
can be approximated by
S(gas)
'"
G(Yin- Yout)
(3.3-24)
for the purpose of estimating an overall error. (Gi differs
from Go by only a small correction for water evapora,ted in
the scrubber.) The error in S(gas) in terms of experimental
errors in G, Yin, and Yout is then given by Equation 3.3-22 or
6S(gas) '" ~ 6G +~ 6y. + oS
= 6Yout
,aG oYin ~n oYout
'" (Yin - Yout) 6G + G6Yin- G6y au t
=
(Yin - Yout) 6G I \
..... + G6\ Y in - Yout) (3.3-25)
Expressing this error as a percentage,
6S(gas) x 100
S(gas)
=
[6g +
6(Yin-
(Yin-
Yout)
Yout)
] x 100
(3.3-26)
-100-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN. TEXAS 78758 . TELEPHONE 512 . 454.9535
The most serious uncertainty i~ that introduced by experimental
error in Yin- Yout. This ranges from + 160 ppm/330 ~ + 48%
for Run 1-2 where y. - Y tis small to only + 160/1520 = + 11%
~n au --
for Run 7 where y.-y is large. The 6.G/G term accounts for an
~ a
additional + 5% uncertainty in the limits of error for the gas-
phase material balance.
The liquid phase total S balance is more accurate.
Limits of error are estimated in a similar manner, giving
approximately + 10% for all runs.
Comparison of calculated errors in.sulfur material
balances with estimated limits of error shows that only two
runs, 3 and 3A, lie outside these limits.
3.3.3.2
Number of Transfer Units
Limits of error for calculated values of Nag can
be readily estimated for the NaOH runs (6-10) since y* = 0
and
Nag
=
tn(Yin/Yout)
(3.3-27)
Again using Equation 3.3-22
6.Nog
oN
~ I~
oYin
oNog
6.Yin!+loy 6.Youtl
out
'"
6.y in 6.y au t
+
Yin Yout
(3.3-28)
-101-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
By substituting appropriate values from Run 6, for example,
6Nog
,..., =*:80 + =*:80
= 2000 trnr
~
=*:.13
(3.3-29)
Limits of error for the remaining runs where y*
calculated and entered in Table 3.3-5.
=
o have been
Error estimates for Runs 1-5 and lOA are more difficult
because an analytical expression for Nog is not available. An
estimate of the error can be made by applying Leibniz's formula
to' equation 3.1-4. For convenience let ~ be the dummy variable
of integration for concentration y. Liquid phase molalities, mi'
will influence the equilibrium term ~*.
~og
=
6Yin /:::.Yout
(Y-Y*)in =*: (Y-Y*)out
I
=*: L 6mi
i=l
SYout l ~
om. t-~*
3-
Yin
(3.3-30)
The third term of Equation 3.3-30 may be approximated
by performing the indicated differentiation and replacing the
integral with a summation. This is done in Equation 303-31.
y.
f3-n 0 d~
omi i=i*
Yout
L
= I
J.=1
/:::.tJ. o~~
(~-~*)~ omi
(3.3-31)
-102-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758. TELEPHONE 512 . 454.9535
The general expression for the error in number of
transfer units, Nog' is given by Equation 3.3-32.
~Nog
=
~Yin
(Y-Y*)in
~Yout
:!: (Y-Y*)out
:!:
I
I
i=l
L
L\mi I
L=l
*
~~L a~L
(~-~*)L ami
(3.3-32)
The special case y* = 0 inp1ies a~*/ami = o.
Equation 3.3-32 then reduces to 3.3-28.
3.3.3.3
Relative Kga
, I
-I
I
I
I
Values of relative Kga have only a small
uncertainty since the factor (~/~ f) (see 3.3-1) is
re
and has limits of error of only + 10%.
additional
near unity
3.3.3.4
Limits of Error for Oxidation
Calculation of the amount of sag oxidized in the
scrubber involves uncertainties in three measurements: the
liquid phase analyses of total SO~ and S, and the liquid flow
rate. Recalling the method of calculation of oxidation
Ox
=
~
1 -
[SOc] x Lav
[SJout x Lav - [SJin x Lav
""
1~
-~
(3.3-33)
-103-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE 512 . 454.9535
where brackets indicate a concentration in moles/liter. This
approximation shows that the error in liquid flow rate cancels
out because it affects the S02 and total S quantities equally.
The error in fraction oxidized is then
6,OX
!!:!
aox [ ] ~ aox 6,[S]
mo:J 6, SOa ~ out-in
.. out-~n
~
~ + [SO~] 6,[S]out-in
[S]out-in ~~ut-in
(3.3-34)
Using data from Run 1-4, for example,
6,OX
e!
~8x10-5 4x10-3
4.52x10-s + (4.52)9x10-e (~9x10-5)
e!
:1= .018 ~ .018
""
~ .04
Similar estim~tes for remaining runs have been made and results
entered in Table 3.3-5.
3.3.4
Experimental Results and Discussion
Results of the Type I Experiments are summarized
in Table 3.3-5. Of primary interest are calculated values of the
overall coefficients for mass transfer of SOa. Also of interest
are observations regarding absorption of NOx and COg and oxida-
tion of SOg to S03 in the scrubber.
-104-
-------
TABLE 3.3~5
Summary of Results
APCO Inhouse Scrubbing Experiments - Type I
50; Out Total
50; in (atm x 10-6) SulfL:r Indicated
10 -G) Number of Overall Relative % Material NO, Venturi Effluent ;:>H
RU!1 No. (at!:) Y. ~ EC!uilib. Gas Transfer Units Koa Oxidation Balance Rer:1.:>val tP(IIH?O) Exp. Calc. C=ents
1 1980 1645 1210 .25 :I: .11 1 :I: .4 9:1:4 - 14% 8.75 2.5 2.5
1A 2000 1630 1260 .29 :I: .13 1.2 :J: .5 16:J:3 + 4% 8.6 2.3
2 2110 1780 970 .20 :J: .09 .8 :I: .4 0::4 - 19% 1% 7.8 2.4 2.6
3 2050 1500 1110 11::4 +128% 2% 12.0 2.5 2.5 P;>or
}~Lcrial
Bal.1.nc~
3A 1990 1560 950 0,,4 +107% 3% 11.4 2.5 II
33 2900 2460 2280 .42 :I: .19 2.0 :J: .9 24::4 7% 12.6 2.2
I 4 2060 1570 900 .34 .15 1.3 .6 11:iA + 31% -3% 9.0 2.4 2.5
t-I ,, :I:
0
l1J 5 1970 1590 860 .27 :I: .12 1.0 % .4 55:1:2 + 18% 5% 8.5 2.5 2.3 Quest:.o:'!-
I able Oxi-
ca.civ:':
6 2000 850 0 .86 :I: .13 3.5 :I: .5 9:1:4 + 23% 9% 8.4 6.1 6.0
6A 1960 850 0 .84 :I: .13 3.6 :I: .5 14 %4 + 3% 9.3 6.9
7 1980 460 0 1.46 :I: .21 5.9 :I: .8 0::4 + 20% 5% 9.2 6.7 7.2
S 2460 1475 0 .51 :I: .09 2.4 :I: .3 l4:iA 6% -1% 12.4 5.8 6.5
8:" 2020 980 0 .72 :I: .13 3.5 :i: .5 4::4 + 2% 12.4 6.2
9 2390 1475 0 .48 :I: .09 2.0 :I: .4 17:1:3 - 12% 10.6 6.1 6.9
9A 2000 1100 0 .60 :I: .11 2.6 :I: .4 0%4 + 1% 9.5 6.3
10 2390 1580 0 .41 :: .08 1.7 :I: .3 25:1:3 +22% -4% 8.6 4.4 6.0
IDA 2075 1420 775 .43 :I: .12 1.8 :i: .5 9:1:4 + 18% 8.7 3.0
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
. 3.3.4.1
SO'" Absorption
Referring to Table 3.3-5, comparison of S02
absorption rates among the Type I experiments supports the
following conclusions:
For this scrubber arrangement, S02
removal does not approach the equilibrium
amount.
Liquor composition has a strong effect
on the rate of SOa absorption. Run 2
the base case using water as a sorbent
results in a relative value of Kga of
.8 ~ .4. Run 6A, where the scrubber
liquor was .04l6M NaOH solution (all
other conditions the same) shows a rela-
time Kga of 3.6 ~ .5. As expected,
Run lOA in which the NaOH solution was
only .0116 molar, yielded a relative
value of Kga (1.8 ~ .5) between those
of Runs 2 and 6A.
These results indicate that the gas film
resistance to SOa mass transfer is not
rate limiting in this particular experi-
mental system. That is, the liquid film
resistance and effect of liquor composition
on this resistance are significant. This
behavior should also be characteristic of 502
removal systems using limestone liquors in a
venturi scrubber. This means that the design
procedure for such a scrubbing system will
have to allow for simultaneous effects of
-106-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9949 . AUSTIN, TEXAS 79759 . TELEPHONE 512 . 454.9535
liquor composition on both equilibrium
80; partial pressure and rate of 80g
mass transfer.
Although it has not been demonstrated
conclusively here, the beneficial effect
of increasing alkalinity on the overall
mass- transfer coefficient obviously has
an upper limit. This occurs when the
liquid film mass transfer resistance
becomes small with respect to the gas
film resistance. There is some evidence
that this point has been exceeded in
these experiments. Runs 6 and 6A and 9
and 9A for example show similar mass
transfer coefficients in spite of signi-
ficant differences in experimental NaOH
concentrations.
An increase in liquid flow rate increases
Kga. This result would be expected strictly
on the basis of an increase in a, the area
available for mass transfer. Run 7 in which
the liquor flow rate was 15 gpm resulted
in a relative Kga of 5.9 ~ .8 while Run 6A,
the base case at 10 gpm, shows a ~ga of only
3.6 ~ .5. It should be noted that this
dramatic increase in mass transfer with
liquor flow rate will not be characteristic
of all scrubber types. In a venturi, assum-
ing that one is in a reasonable range of
operation, additional liquor goes to form
additional droplets and thus, surface area.
-107-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN. TEXAS 78758 . TELEPHONE 512 - 454-9535
In a scrubber such as a fixed-bed packed
tower, however, the area available for
mass transfer is determined more by the
packing type than the liquor flow rate in
normal operating regions. In a packed bed
the influence of liquor rate on the over-
all coefficient will depend on whether the
liquid film resistance is controlling.
For vapor film controlling operations the
influence would be negligible.
The present experiments were not sufficiently
accurate to determine the effect of gas flow
rate on Kga. The change in gas flow rate
between Runs 6 or 6A and Run 8A was only 10-
20%. While the relative values of Kga were
the same for these runs, the effect of such a
small percentage change in gas flow rate may
be hidden by experimental error.
Even though there was no noticeable increase
in K a between Runs 6 or 6A and 8, there
g
was a substantial increase in pressure drop
across the venturi. The increase in pressure
drop caused by increasing the liquid flow
rate was much smaller (see Table 3g3-S). Since
most of the power input to a venturi is from
the gas phase pressure drop, it would appear
that increasing the. liquid flow rate rather
than the gas flow rate is a more suitable
way to increase S02 mass transfer rate in a
venturi scrubber. The result of an in-
crease in liquid flow rate in this case
is an incre~se in "a," the interfacial
-108-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN. TEXAS 78758 . TELEPHONE 512 - 454-9535
area for mass transfer. This effect
should not be confused with that of
increasing "L/G" or the liquid to gas flow
ratio to improve the equilibrium driving
force. The back pressure of Sea is
smaller above a more dilute solution so
that increasing LIG increases the driving
force for mass transfer. The behavior
discussed above was not equilibrium related
since the partial pressure of Sea above
the liquor was near zero in both cases.
A decrease in scrubber liquor temperature
may slow S09 absorption under some condi-
tions. Run 9A was conducted with precooled
flue gaB so that the scrubber liquor tempera-
ture reached only 72°F compared with 86°F
for Run 6A. K a was significantly lower in
g
Run 9A relative to Run 6A. It is possible
that this effect is due to a chemical reac-
tion rate limiting mechanism since such
reaction rates are normally strong functions
of temperature. Under other process condi-
tions; where the equilibrium partial pressure
of Sea is not zero as in these experiments,
one would expect a lower liquor temperature
to be beneficial since it would decrease the
partial pressure of Sea above the solution.
These temperature effects require more study
before firm conclusions can be reached.
-109-
-------
Radian Corporation
8409 ~ESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
An increase in gas phase SO!;! inlet
concentration appea.rs to decrease the
relative value of K a. This is seen
g .
by comparing Runs 8 with 8A and 9 with
9A. This conclusion is at best tenta-
tive in view of experimental difficulties
with gas phase SO!;! measurements during
Runs 8, 9, and 10. Since scrubber liquor
compositions were similar for these runs,
no satisfactory explanation is seen for
this effect.
3.3.4.2
NOx Absorption
There was apparently no significant removal of
NOx from the flue gas. The gas phase monitor indicated NOx
absorption percentages ranging from -4% to +9%. These figures
are on the same level as the limits of error introduced by
instrumental accuracy. Liquid phase total N analyses (see
Table 3-4) show a maximum of '" .46 m mole N/liter while only
10% removal of inlet NOx would result in most cases in '" 1
m mole N/liter of scrubber effluent (in addition to the", .2
m mole N/liter in the scrubber feed).
In view of the questionable sensitivity of both the
gas and liquid phase nitrogen analyses for low levels of NOx
removal, a more accurate determination of NOx removal must await
operation of the equipment with recycle as in the Type II and
III experiments. Here, the buildup of soluble nitrogen salts
should provide a good estimate of the amount of NOx absorbed
from the gas.
-110-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN. TEXAS 78758 . TELEPHONE 512 . 454-9535
3.3.4.3
CO~ Absorption
Carbon dioxide is also absorbed by alkaline
scrubbing liquors. In the water runs (1-5), CO2 is first
absorbed, but then evolves as the scrubbing medium becomes
more acidic. Since S03 is a stronger acid than CO2, the
scrubber effluent ultimately contains less COs than the scrubber
feed (see Tables 3.3-2 and 3.3-3). When the scrubbing liquor
remains alkaline as in some of the NaOH runs, there is a net
sorption of CO2.
Since normal flue gas contains much more CO2 than
S09' the possibility exists that these acid gases will
"compete" for available alkalinity. It is possible, then,
that experiments using 13-14% CO2 rather than - 1% CO2 as in
the flue gas used here, would show some detrimental effect
of CO2 on S03 mass transfer rates. The equilibrium capacity
of the liquor for S02 however is not significantly effected by
higher CO2 partial pressures.
3.3.3.4
Sulfite Oxidation Rate
No attempt was made in the Type I experiments to
investigate effects of process parameters on the rate of sulfite
oxidation. The data in Table 3.3-5 do suggest an interesting
observation, however. Oxidation of sulfite to sulfate averages
10% for seven water runs (excluding Run 5, which appears to be
questionable). Nine runs with NaOH liquors also result in an
average of 10% sulfite oxidation. Since the total sulfur absorb"ed
for the water runs averages .225 g mole/min while that for the
NaOH runs averages .825 g mole/min, it is evident that the rate
of oxidation is approximately four times as great in the latter
case. Gas and liquid contacting parameters are the same between
-111-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN. TEXAS 78758 . TELEPHONE 512 . 454-9535
the two sets of runs so that gas film resistance to 0::1 transfer
seems to be ruled out as a rate-limiting step for oxidation.
It is tempting to conclude that the rate is proportional to the
concentration of SO::!, indicating a first order chemical reaction
rate limiting step, but observed results do not necessarily
eliminate the liquid film mass transfer resistance, since
chemical reaction and thus concentration of the reactant (SO~)
in the liquid is known to favorably effect liquid film mass
transfer coefficients (DA-020). This affect could also account
for more rapid oxidation at higher SO; concentrations.
It should be noted that the concentration of oxygen
in the flue gas used here is five to six times that normally
encountered in a large coal- or oil-fired boiler. Thus the
observed oxidation rate behavior may not apply to scrubbers
operating with more typical flue gas compositions.
3.4
General Comments on Experimental Results
The results discussed above are perhaps of most
value in demonstrating the applicability of the techniques used
to characterize vapor-liquid mass transfer in a complex scrub-
. bing system. In all cases, calculated values of equilibrium S03
partial pressures appear to be reasonable. In addition, agree-
ment of calculated vs. experimental pH values for the scrubber
effluent (see Table 3.3-5) in the acidic range is excellent.
Thus, even though results for this particular scrubber may not
be directly applicable to commercial scrubbers of different
designs, a method is established whereby the vapor-liquid mass
transfer capability of a system of interest can be easily checked.
If this capability is then found to be acceptable under reason-
able operating conditions, attention can be directed exclusively
to problems associated with solid-liquid mass transfer at a
considerable savings of effort.
-112-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-9535
4.0
ENERGY CALCULATIONS
Process design procedures usually require energy
balance calculations. In this section the objectives, theory,
and results of a means of calculating the energy balance for
lime/limestone scrubbing systems are presented.
4.1
Ob;ectives
Limestone scrubbing systems are three phase systems.
Chemical reactions and heats of dilution as well as sensible
heat changes occur. A means of calculating enthalpy balances
for process design or data analysis is necessary. In order to
do this an enthalpy calculation routine was developed for each
of the three phases.
Since chemical reactions and interphase mass
transfer are to be considered it was natural to choose the
heat of formation of the various species as the reference state
enthalpy. In addition to sensible heat and phase transition,
heats of dilution were to be considered.
4.2
Theory
were ideal.
For gases it was assumed that all solutions formed
For solids it was assumed that no solid solutions
were formed. Both of these assumptions lead to
that the enthalpy of the stream is equal to the
enthalpy. This is expressed in Equation 4-1.
the conclusion
reference state
-113-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
T
H
=
L
i
n.H~
~ ~
(4-1a)
=
\" (298
L ni Hf,i +
i
T
S
298
\ \- 298-+T
C dT) + ) n.6H h
p , ~ ~ p ase change
~
(4-1b)
The heat capacities were used in the form of
equation 4-2.
C
P
= crl + crgT + cr3T2- cr4/T
(4-2)
For the aqueous solutions the entha1pies were
calculated from Equation 4-30
H = '\ n.H.
L ~ ~
i
(4-3a)
= I ni[Hi + (Hi - Hi)]
i
(4-3b)
The value for the liquid phase term ~n.H~ is
calculated in the form of Equation 4-1. The term tH~ - Hi)
is calculated from the following thermodynamic relationship.
(Hi - Hi)
=
~£na 2 a!nY~
-RT2 u i = -RT ----~
aT aT
(4-4)
The activity coefficients used in this project
are of the following form:
-114-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
p,nY.
~
a -l~
= (Az. .tn10)[
~ 1 + ~.Bl~
~
where
6
A - 10824x10
- (DT) s/a
k
d 2
o
~
B = 50.29do
(DT)~
+ b.l] + U.l
~ ~
(4-5)
(4-6)
(4-7)
Equation 4-4 may be evaluated for the nonwater species using
Equations 4-5, 6, and 7.
a a
(Hi - Hi) = RT A;itn10 {3( r + 3)bil -
a. ==
~ (~i) = ~/ (~~o)
pop
1 aD 1
r == D aT + T
o k
13. = a.Bl2
~ ~
l~[2eir + 3(~ + t)]
(1 + B.ja
~I
(4-8)
(4-9)
(4-10)
(4-11)
For water the term is given in Equation 4-12.
-115-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 994B . AUSTIN, TEXAS 78758 . TElEPHONE 512 - 454-9535
:2
(Hw - H~)= R~ { 3 (r + ~) [p',naw
J
+ tnlO '\ m. (U; I + ~)l
mw L J 2 tn 10 -
j=l
k
- A(r + ~) I2tnlO
mw
J
I
j=l
2
m.z.
J J
S~
J
[ - 6 tn (1 + SJ.)
Sj
+
2
6 + 9 S. + 2 S . J}
1 1
( 1 + Sj r~
(4 -12)
A more complete derivation is given in T.N. 200-004-12
which is in Volume II of this report.
4.3
Comparison of Experimental and Calculated Results
The enthalpies of the gases and solids of interest
are well documented. For these compounds ,Radian used values
from rel~able compilations such as Kubaschuvski (KU-003),
Landolt-BClrnstein (LA-008), and JANAF (8T-006). As a result
no comparison with experimental values was necessary for gases
and solids.
For key aqueous species the values of the heats of
formation and heat capacities used are given in Table 4-1.
Values of other ionic species were calculated using
the fact that heats of reaction can be calculated from the
variation of equilibria with temperature.
oR-nK -
~-
6HO
~
RT:ia
(4-13)
-116-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN. TEXAS 78758 . TELEPHONE 512 . 454.9535
TABLE 4-1
HEAT OF FORMATION AT 25°C AND MEAN HEAT
CAPACITY BETWEEN 25 AND 60° C
Species H£ KCal/gmo1e . 25-60oC
Cp Ca1/gmo1eOK
H2 0 ( t) -680317 18.04
H+ 0.00 23.
(aq)
SO:(aq) -216.90 -99.
NO; (aq) -49.372 -49.
Ca"H;q) -129.77 45.
S02(g) -70.93 -----
CO:; (g) -94.01 -----
C1(aq) -40.023 -51.
Mg(;q) -110.41 51.
Nataq) -57.279 35.
Heat Capacity Data from Criss and Cobble (CR-007).
Heat of Formation Data from NBS Circular 500 (RO-007).
-117-
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
Values of 6H~ were calculated from the constants
used in the Radian equilibrium program (LO-OO7). For example,
the enthalpy of OH-( ) can be calculated from the reaction
. aq
HaO(t) - H1aq) + OH(aq) + 6HR
( 4 - 14 )
o
25 C
Hence the value of Hf OH- is
, (aq)
o
a5 C
Hf,OH-(aq)
=
o 0
25 C 25 C
H + 6H R
f,H20(1.)
o
25 C
H +
f,H (aq)
=
-68.32 + 13.53 - 0
=
-54.79 Kcal/gmole
(4-15 )
manner.
Values of heat capacity'were obtained in the same
It should be remembered that the second differentiation
of tnK introduces inaccuracies.
The complete tabulation of literature and calculated
heat of formation and mean heat capacity values are given in
Table 4-2.
Using the values in Table 4-2 several systems were
calculated and compared with experimental results. These are
given in Table 4-3.
-118-
-------
Radian Corporation
Species
H20
H+
OH-
HSO;
SO;
SO:
HCO;
CO=
3
NO;
HSO;
H2 S03
H2 C03
Ca++
CaOH+
CaSO~
Ca CO~
CaHCOt
CaSO~
+
CaN03
Mg++
MgOH+
840'1 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454.9535
TABLE 4-2
STANDARD STATE HEATS OF FORMATION AND MEAN
HEAT CAPACITIES OF LIQUID SPECIES
o (kca1) ]800C( ca1 )
Hf,298.16 mole C~ 26 mo1eoK
- 68.317 18.04
0.000 23.
- 54.799 - 47.
-149.376 - 16.
-152.276 -121.
-216.900 - 99.
-164.729 - 27.
-162.126 -132.
- 49.372 - 49.
-211.660 - 13.
-145.516 + 7.
-166.971 91.
-129.770 45.
-183.320 2.
-279.736 - 76.
-289.720 - 87.
-293.118 18.
-344.033 8.
-173.445 4.
-110.410 51.
-162.839 4.
continued
-119-
-------
Radian Corporation
8409. RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-'535
TABLE 4-2 (continued)
0 (kca1) ]SOOC( ca1 )
. Hf,298.16 mole C~ as mo1eoK
Species
MgSO~ -260.707 - 70.
MgHCOt -274.063 24.
MgSO~ -322.469 - 48.
MgCO~ -270.226 - 81.
Na+ - 57.279 35.
NaOHo -112.078 - 12.
NaCO; -218.017 - 97.
NaHCO~ -222.008 8.
NaSO~ -273.076 - 64.
Na NO~ -106.651 - 14.
Cl- - 40.023 - 51.
-120-
-------
~ ~ A ~L ~ 4 - 3
COMPARISON OF EXPERIMENTAL AND CALCULATED RESULTS @ 25°C
Experimental Calculated Experimental Calculated
Molality ~Hso1n 6Hso1n La La
Solute (mole/kg H."O) (cal/mol e) (ca1/mo1e) (ca1/mo1e) (ca1!mo1e)
HCL(g) 0.04 -17,800 -17,895
CaC1a(s) 0.02 -19,560 -19,560 360 339
0.05 ------- ------- 490 443
0.10 ------- ------- 620 517
0.60 ------- ------- 1020 529
Ca(OH):a(s) 0.004 - 4,082 - 3,525
0.01 - 3,819 - 3,400
0.02 - 3,523 - 3,345
NaC1 0.06 ------- ------- 94 112
0.37 ------- ------- 42 181
I
......
N
......
I
~
-------
Radian Corporation
8409 RESEARCH 8LVD. . P.O, 80X 9948 . AUSTIN, TEXAS 78758 .' TELEPHONE 512 . 454-9535
5.0
SUMMARY AND CONCLUSIONS
A computerized process model was developed to
simulate the prototype system being designed for the TV A
Shawnee power planto Using this model, a parameter study was
performed to quantitatively estimate the effect of important
process variables on overall process performance. Radian has
also provided technical assistance to APCO during pilot-scale
SO:a scrubbing experiments being conducted at their Cincinnati
laboratory facilities. This technical assistance has included
test'plan design, development of sampling and analytical tech-
niques, chemical analysis of scrubber liquor samples, and inter-
pretation of experimental data. In addition, a computational
technique was developed for calculating the enthalpy of process
streams in limestone wet scrubbing systems.
The computerized model for the limestone injection-
wet scrubbing process was developed based upon (1) the ability
to predict vapor-liquid-solid equilibria for the CaO-MgO-Na20-
SO:a-CO:a-SOs-N:aOs-HC1-H:aO system and (2) a number of process
assumptions with regard to equipment characterization and the
extent at various chemical reactions. The computerized model
was designed to be a compromise between generalization for
greater system flexibility and specialization for improved
computational efficiency.
The process model was used to perform twenty-six
simulation cases for the "prototype" system. The simulation
results demonstrate how the chemistry and performance of the
limestone injection-wet scrubbing process vary with important
parameters. The following conclusions have been drawn based
upon these simulation cases.
-122-
-------
Radian Corporation
( 1)
(2)
(3)
(4)
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 7875B . TELEPHONE 512 . 454.9535
Scrubbing solutions originating from
high calcium limestones are more
efficient than those originating
from dolomites. The magnesium con-
tent of the limestone additive is
an important process variable 0
Sulfite oxidation reduces the SOa
scrubbing capability of the wet
scrubbing process. Sulfite oxida-
tion is an important process
variable.
For these simulation cases, the
effect of ionic strength, circu-
lating liquor temperature, solids
precipitation in the scrubber, and
fraction solids in the filter
bottoms stream on scrubbing capability
were relatively minor. The effect of
circulating liquor temperature could
be more significant if (1) the tempera-
ture of each process vessel were varied
independently and (2) reaction rates
were considered.
The ionic strengths of process liquors
can vary over a wide range. The result.-
ing deviations from ideality must be
considered in the prediction of V-J,-S
equilibria and correlation of reaction
rates.
-123-
-------
Radian Corporation
(5)
(6)
(7)
(8)
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE 512 - 454-9535
The extent of lime hydration and
dissolution has a major influence
on the scrubbing capability of
the process.
The effect of varying scrubber feed
rate is much more dramatic in the
pure calcium system than in the
calcium-magnesium system.
The liquid-phase compositions of the
various process streams are determined
to a large extent by the s-~ equilibria.
These compositions remain relatively
constant as most model parameters are
varied. The major change occurs as
the amounts of CaO and MgO available
for reaction are varied.
If the wet scrubbing processes were
operated adiabatically, the temperature
variations between process vessels would
be only about 2°F. The effect of the
exothermic reactions occurring within the
scrubber and effluent hold tanks would
tend to be counterbalanced by heat losses
from these vessels.
It should be emphasized that these conclusions are dependent
upon the validity of the process assumptions and the number of
simulation cases runo
-124-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 994B . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-9535
Under a previous APCO contract, Ra.dian devised a
test program for the pilot venturi scrubbing system located in
APCO's Cincinnati laboratories. The major program objectives
include:
( 1)
(2)
(3)
(4)
(5)
(6)
Determine the extent of vapor-liquid
and solid-liquid mass transfer in
the venturi scrubbing section, i.e.,
the approach to equilibrium.
Study factors influencing the important
mass transfer steps in the scrubbing
section.
Measure the hydration and dissolution
rate of calcium oxide (and magnesium
oxide) in the effluent hold tanks.
Measure the precipitation
calcium sulfite, sulfate,
in the hold tanks.
rates of
and carbonate
Gather preliminary information on the
absorption of NOx, build-up of Na+, Cl-,
NO;, and liquid phase reactions of NOx.
Provide realistic conditions for field
evaluation of sampling and analytical
methods.
The first set of experiments (Type I) were designed
to evaluate the importance of vapor-liquid mass transfer rates
under conditions that can be related to lime/limestone scrubbing
-125-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE SI2 . 454-9535
processes for S02 absorption. The Type I experiments were
performed and their results were analyzed during the contract
period. The limitations of this experimental program have been
noted. A computational method has been developed for engineer-
ing analysis of these data. Chemical methods of analysis for
CO2, N, and SOa were developed so that pert inent experimental
data could be obtained.
Based upon the results from the Type I experiments,
the following conclusions were drawn regarding SOa absorption:
SO a removals obtained in this scrubber
arrangement did not closely approach
th: vapor-liquid equilibrium amount.
Liquor composition has a strong effect
on the rate of S02 absorption. This
would indicate that the gas film
resistance is not rate limiting in
this experimental system. That is,
the liquid film resistance and effect
of liquor composition on this resistance
are significant.
An increase in liquid flow rate increases
K a. This result would be expected based
g
upon an increase in interfacial area for
venturi scrubbers.
The present experiments were not sufficiently
accurate to determine the effect of gas flow
rate 0'£ K a.
g
-126-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9949 . AUSTIN, TEXAS 7975B . TELEPHONE 512 . 454.9535
An increase in gas flow rate caused a
significant increase in pressure drop
across the venturi. The effect of
liquid flow rate on pressure drop was
much less. Thus, increasing the liquid
flow rate to increase SO; mass transfer
rate would be more suitable than increas-
ing gas flow rate.
A decrease in scrubber liquor temperature
may slow S02 absorption under some con-
dit ions.
The NOx absorption was rather low (less than 10%). The concen-
tration CO" in the flue gas stream used in these experiments
was abnormally low (~l% instead of the 13-14% expected from
from coal-fired boilers). Higher CO2 concentrations should
result in more COa absorption in the NaOH scrubbing runs which
could show some detrimental effect on SOs removal. The amount
of sulfite oxidation averaged 10% oxidation for both the water
and NaOH runs, even though more SOa was absorbed in the NaOH
runs.
Even though these results may not be directly
applicable to commercial scrubbers of different designs, a
method has been established whereby the vapor-liquid mass trans-
fer capability of a system of interest can be easily checked.
Results of the Type I could be extended to other venturi
scrubber systems with regard to approximate ranges of K a,
g
qualitative effects of gas and liquld flow rates on K a, and
g
the order of magnitude effects of liquor composition of K a.
g
-127-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9'/48 . AUSTIN. TEXAS 78758 . TELEPHONE 512 - 454-9535
The next series of experiments planned for the APCD
venturi scrubbing system involve studying solid-liquid mass
transfer in the scrubber and the hold tanks. Hopefully, this
program will furnish reliable data for estimating the solid-
liquid mass transfer rates, will demonstrate (in the field)
chemical methods of analysis for characterizing limestone
scrubbing streams, and will develop satisfactory computational
methods for performing an engineering analysis of this element
of the scrubbing system.
-128-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 : TELEPHONE 512 . 454-9535
6.0
NOMENCLATURE
The nomenclature for this report is
within any given section but not for the entire
are separate nomenclature lists for Sections 2,
consistent
report. There
3, and 4.
6.1
Section 2 Nomenclature
C
CB
CF
CL
Cpo
1.
E
F
FA
designation for clarifier vessel
designation for clarifier bottoms stream
designation for clarifier feed stream
designation for clarifier liquid stream
h . f h . th .
average eat capac1.ty 0 t e 1.- eX1.t stream
designation for scrubber effluent hold tank
designation for filter vessel
designation for fly ash stream; and
fly ash level (lbs. FA/lb. S in coal)
designation for filter bottoms stream
designation for flue gas stream
designation for filter liquid stream
designation for gas-solid stream entering system
entha1pies of the exit and inlet streams
designation for the limestone - fly ash (solid)
stream entering scrubber
limestone composition (mole fraction MgO in
LS stream)
fraction of solids from LS stream dissolving
in scrubber
lime hydrating in scrubber
designation for limestone stream
lime hydrating in system
1 1. f h .th . . h th
mo a 1.ty 0 t e J-- speC1.es 1.n t e N- stream
FB
FG
FL
GS
H. ,H.
1. J
LA
LC
LD
LH
LS
LU
mN (j )
-129-
-------
Radian Corporation
NSP
P
S
SB
SD
SF
SFTFG
SG
SR
TL
Ts
w.
~
WM
XA( COg)
XA(NOx)
XA( SO:; )
XO
XWFA(Cl)
XWFA(Na)
XWSFB
YFG(C02)
YFG(NOx)
YFG(SOg)
8409 RESEARCH 8LVD. . P.O. 80X 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
index for solids precipitation occurring in
scrubber (NSP=l implies No ppt.)
designation for process water hold tank
designation for scrubber v~ssel
designation for scrubber bottoms stream
fraction of solids from SR stream dissolving
in scrubber
designation for scrubber feed stream
scrubber feed rate to flue gas ratio
SF/1000 ACF FG)
(gal.
designation
designation
theoretical
for stack gas stream
for slurry recycle stream
limestone
temperature of the scrubber liquor
mass flow rate of the ith exit stream
designation for water make-up stream
fraction of CO:; in flue gas that is absorbed by
the scrubber system
fraction of NOx in FG that is absorbed in system
fraction of SO:; in FG absorbed by the system
fraction of absorbed SO:; that is oxidized in the
system
wt. fraction of soluble chloride in fly ash
(FA) stream
wt. fraction of soluble sodium in fly ash (FA)
stream
wt. fraction of solids in the filter bottoms
(FB) stream
mole frac ti on of COg in flue gas (FG) stream
mole fraction of NOx in flue gas (FG) stream
mole fraction of SOg in flue gas
-130-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 . 454-9535
GREEK
6.TA
temperature increase of the vessel's
that would be required to satisfy an
adiabatic energy balance.
exit streams
exact
6.2
Section 3 Nomenclature
z
interfacial area/unit volume of contactor, ft-1
molar gas flux, lb. moles ft-g sec-1
gas flow rate, lb. min-l or ft3 min-l
humidity, lb. H20/lb dry gas
height of an overall gas phase transfer unit, ft.
individual liquid and gas phase mass transfer
coefficients, lb. mole ft-a sec-l
overall gas phase mass transfer coefficients,
lb. mole ft-2 sec-l
liquid flow rate, liter min-l
Henry's law constant based on mole fraction
number of overall gas phase transfer units
gas phase concentration, mole fraction
equilibrium gas phase concentration, mole fraction
axial length of an absorber, ft.
a
G
G
H
Hog
k t ' kg
Kog
L
m
Nog
y
y'/(
GREEK
iR
experimental error in a measured or calculated
quantity
dummy mole fraction variable
6.
-131-
-------
Radian Corporation
8409 RESEARCH BLVD. . P.O. BOX 9948 . AUSTIN, TEXAS 787S8 . TELEPHONE 512 . 454-9535
6.3
Section 4 Nomenclature
mi
ni
.. f .th .
actJ.vJ.ty 0 J.- specJ.es
ion size activity coefficient parameter of ith species
Debye-HUckel limiting slope constant
.th .
activity coefficient deviation parameter of J.- specJ.es
Debye-HUckel constant
heat capacity
density of water
dielectric constant of water
enthalpy of mixture
reference state enthalpy of ith species
partial molal enthalpy of ith species
standard state heat of reaction
ionic strength
equilibrium constant
integral heat of dilution
1 1. f. th .
mo a J.ty 0 J.- specJ.es
1 f .th .
mo es 0 J.-- specJ.es
. a.
. oJ.
a.
, J.
A
b.
J.
B
Cp
do
D
H
Hi
Hi
~H~
I
K
L
R
T
Ui
gas constant
absolute temperature
activity coefficient
species
volume
parameter for uncharged ith
v
z.
J.
.th .
charge number of J.- specJ.es
GREEK
a.
term defined in Equation 4-9
term defined in Equation 4-11
activity coefficient of ith species
heat capacity coefficients of Equation 4-2
term defined in Equation 4-10
s.
. J.
y.
J.
cr.
J
1"
-132-
-------
Radian Corporation
700
CR-007
DA-020
KU-003
LA-008
LO-007
RO-007
ST-006
TE-005
8409 RESEARCH 8LVD. . P.O. BOX 9948 . AUSTIN, TEXAS 78758 . TELEPHONE 512 - 454-9535
BIBLIOGRAPHY
Criss, C. M., and J. W. Cobble, "The Thermodynamic
Properties of High Temperature Aqueous Solutions;
The Calculation of Ionic Heat Capacities up to 2000;
Entropies and Heat Capacities Above 2000," JACS, 86,
5390-5393 (1964).
Danckwerts, P. V., Gas Liquid Reactions, McGraw-Hill,
New York, 1970.
Kubaschewski, 0., et al., Metallurgical Thermochemistry,
4th Edition, Pergamon Press, 1967.
Landolt-B8rnstein, "Kalorische Zustandsgr8ssen;' ~(4)
Springer-Verlag, Berlin, 1961.
Lowell, P. S., et al., "A Theoretical Description of
the Limestone Injection - Wet Scrubbing Process,"
Volume I, Final Report for APCO Contract No. CPA-22-
69-138, 9 June 1970; PB 193-029.
Rossini, F. D., et al., Selected Values of Chemical
Thermodynamic Properties, Circular of the NBS 500
(1952).
Stull, D. R., et a1., Janaf Thermochemical Tables,
PB-168370, Dow Chemical Co., Midland, Michigan,
August 1965.
Tennessee Valley Authority, "Sulfur Oxide Removal
from Power Plant Stack Gas - Conceptual Design and
Cost Study - Sorption by Limestone or Lime: Dry
Process," TVA, Knoxville, Tennessee, 1968.
-133-
------- |