EPA-R2-72-035a
August 1972
S02 FREE TWO-STAGE COAL COMBUSTION PROCESS
APPENDICES
by
Applied Technology Corporation
135 Delta Drive
Pittsburgh, Pennsylvania 15238
for the
ENVIRONMENTAL PROTECTION AGENCY
-------
EPA Review Notice
This report has been reviewed by the Environmental
Protection Agency, and approved for publication.
Approval does not signify that the contents necessarily
reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or
commercial products constitute endorsements or recom-
mendation for use.
EPA Project Officer
Mr. Douglas A. Kemnitz, of the Control Systems
Laboratory Division was the EPA project officer
for the work discussed in this report. Mr. Stanley
J. Bunas is the EPA project officer for the continuation
of this work.
-------
ABSTRACT
This volume contains the appendices to the report entitled,
"SO Free Two-Stage Coal Combustion Process Progress Report".
Appendix A is a discussion of the experimental combustor de-
sign, construction, and operation. Appendix B details the
laboratory work conducted for this study. In Appendix C is
presented the Two-Stage Coal Combustion Process simulation
and economic evaluation.
-------
CONTENTS
Page
Abstract i
Appendix A - Combustor Design, Construction, and Operation 1
Experimental Combustor System . 1
Injection System 3
Injection Lances 5
Process Instrumentation and Control System 5
Raw Materials and Supplies 7
Experimental Procedure 9
Gaseous Sampling and Analysis . 12
Analytical System Operation 16
S02, NO , NO , Analysis . . 16
Hydrogen Analyzer 17
CO, CO- Analyzer 17
Oxygen Analyzer . 18
Iron and Slag Sampling Analysis 18
Appendix B - Laboratory Studies 25
Summary of Laboratory Results 25
Slag Composition 25
Slag Viscosity 25
Equilibrium Partition Ratios 25
Heat Capacity 25
Crushing Energy Requirements 26
Slag Granulation 26
Slag Desulfurization 26
Slag Fluidity 26
Introduction 26
Experimental Procedure 27
Discussion of Results 28
Slag Viscosity 32
.Introduction 32
Experimental Procedure 32
Results and Discussion 36
Sulfur Partition Between Iron and Slag 42
Introduction 42
Slag Preparation 43
Effect of Time and Temperature on Approach to
Equilibrium . . 43
Effect of Slag Temperature and Basicity on Equilibrium
Partition Ratio 44
ii
-------
CONTENTS CONT'D
Page
Heat Capacity ' 46
Introduction 46
Procedure 46
Results 48
Total Specific Surface Area 52
Procedure 52
Discussion of Results 52
External Surface Area . 53
Introduction 53
Theory 53
Equipment and Experimental Procedure 54
Discussion of Results . ., . 54
Crushing Energy 55
Introduction 55
Experimental Equipment and Procedure 57
Discussion of Results 59
Granulation -Study 62
Introduction 62
Experimental Equipment 62
Experimental Procedure 62
Discussion 65
Conclusion 71
Slag Desulfurization 71
Introduction 71
Experimental Procedure 71
Discussion of Results 74
Appendix C - Process Simultaion and Economics 79
Equipment Cost 79
Coal Preparation Complex 79
Slag Preparation Complex 85
Flux Preparation Complex 85
Air Preparation Complex 87
Combustor Complex 88
Slag Desulfurization 89
Cost Controlling Variables 93
Total Purchased Equipment Cost 93
Estimated Fixed Capital Requirements.. 98
Power Plant Costs 98
Operating Costs 98
iii
-------
CONTENTS CONT'D
Page
Power Plant operating Costs and Process Parameters. . . . 101
Summary of Operating Ranges for Important Process
Parameters 110
References Ill
iv
-------
FIGURES
Figure Page
A-l Installation Plan View 2
A-2 Dense Phase Pneumatic Coal Injection System 4
A-3 Process Flow Control Diagram 6
A-4 Analytical Instrument Panel Schematic 13
A-5 Carbon Determination Equipment 20
A-6 Sulfur Determination Equipment 22
1-B Modified Herty Fluidity Test Apparatus 29
2-B Effect of Basicity and CaS Content on Slag Fluidity . . .31
3-B Viscosimeter and High Temperature Furnace 33
4-B Viscosimeter Bob and Cup 35
5-B Effect of Bob Location on Dampening Constant 37
6-B Viscosity of Slag V-2 (40% CaO, 40% Si02> 20% Al^). . .39
7-B Effect of Sulfur Content on Apparent Viscosity of
0.2 Basicity Slag . . . 41
8-B . Effect of Basicity on Apparent Viscosity of High
Sulfur Bearing Slags 41
9-B Effect of Heating Time on Partition Ratio 45
10-B Effect of Slag Basicity on Partition Ratio 47
11-B Correction Factor for Heat Losses During Sample
Transfer to the Calorimeter . „ 49
12-B The Effect of Temperature on the Relative Heat
Content of Various Slags 50
13-B The Variation of Slag Heat Capacity with Temperature. . .51
14-B Effect of Slag Basicity and Particle Size on External
Specific Surface 56
-------
FIGURES CONT'D
Figure • Page
15-B Crushing Energy Apparatus. . 58
16-B Calibration Curve for Aluminum Wire Used in
Crushed Energy Test 60
17-B Vacuum Tower Granulation Apparatus 63
18-B Drop Formation Apparatus 64
19-B Effect of Orifice Diameter on Surface Area of
Granulated Slag 69
20-B Effect of Temperature on Surface Area of Granulated
Slag 70
21-B Desulfurization Kinetics Experimental Apparatus 72
22-B Effect of Process Parameters on Slag Desulfurization . . 76
23-B Variation of Offgas Composition with Time at 2000°F. . . 78
1-C Process Flow Diagram 80
2-C Process Equipment Layout 82
3-C 48-Foot Diameter Combustor 90
4-C Operating Cost/Air Preheat Temperature versus
CombustorTemperature . 104
5-C Operating Cost/Air Preheat Temperature versus
% Moisture in Coal 105
6-C Operating Cost/Air Preheat Temeprature versus
Process Limestone in Flux 107
7-C Operating Cost/Air Preheat Temperature versus
Slag Basicity 108
8-C Operating Cost/Air Preheat Temperature versus
% Sulfur in Combustor Slag 109
vi
-------
TABLES
Table Page
I-A Typical Pig Iron Composition 7
II-A Coals Used in Experimentation 8
III-A Inducto-87A Refractory 9
IV-A Harmix CU Refractory 10
V-A Korundal XD Refractory 11
I-B Premelt Composition of Slags 27
II-B Effect of CaS Content on Fluidity of High Basicity
Slags 30
III-B Effect of Immersion Depth on Equipment Dampening
Constant (k) at Varying Bob Distance from
Crucible Bottom 36
IV-B Premelt Slag Composition 38
V-B Comparison of Measured Viscosity and Literature
Values (12,13) 38
VI-B Slag Composition for Partition.Ratio Studies 43
VII-B Total Surface Area of High Sulfur Bearing Slags ... 52
VIII-B Comparison of Glass Bead Surface Area Measured
by Micrometer and by Air Permeability Method 55
IX-B Variation of Crushing Energy with Basicity 59
X-B Comparison of Crushing Energy Requirements for
Silica and Slags 61
XI-B Physical Properties of Liquid Oil and Mercury .... 66
XII-B Vacuum Granulated Slag-Experimental Results 67
XIII-B Vacuum Granulated Slag—Particle Size Distribution. . 68
vii
-------
TABLES (CONT'D)
Table
I-C
II-C
III-C
IV-C
V-C
VI-C
VII-C
VIII-C
IX-C
X-C
XI-C
XI I-C
XIII-C
XI V-C
xv-c
XVI-C
Page
Process Stream Description—Two Stage Coal
Combustion Process 81
Coal Preparation—Equipment Costs 83
Slag Preparation—Equipment Costs 86'
Flux Preparation—Equipment Costs 85
Air Preparation—Equipment Costs 87
Combustor—Equipment Costs 91
Desulfurization-Equipment Costs ,. . . .92
Cost Controlling Variables . . 94
Equipment Cost Factors .'95
Coal Composition 96
Process Stream Rates—1000 MW Power Plant 96
Total Purchased Equipment Cost 97
Estimated Fixed Capital Requirement Two-Stage Coal
Combustion Process—1000 MW Power Plant 99
Estimated Capital Requirements for 1000 MW Power
Plant Systems
100
Estimated Operating Cost 1000 MW Power Plant 101
Effect of Coal Composition on Operating and
Economic Data 103
viii
-------
APPENDIX A
COMBUSTOR DESIGN, CONSTRUCTION, AND OPERATION
The following is a discussion of the design, construction, and operation
of the experimental combustor system in which an induction melting fur-
nace was used to simulate the combustor and an offgas processing system
was incorporated to prepare the combustor offgas for release to the
atmosphere. The coal injection system, process instrumentation, and
controls, sampling, and requisite auxiliary equipment will be discussed.
Experimental Combustor System
An installation plan view of the experimental combustor system is shown
in Figure A-l. The installation was constructed on three levels. The
induction melting furnace was positioned on the lowest level in a fire-
brick-lined pit and the furnace control panel, coal injection system,
and offgas processing system were located on the second or main operating
level. An emergency spill cavity was constructed in the pit to capture
molten metal if a furnace run-through or an emergency pour should occur.
A two-ton (capacity) jib crane was located on the main operating level
to move equipment and iron ladles. An induced draft fan was located on
the third or uppermost level to maintain negative pressure on the
experimental combustor process system. The entire operating area was
enclosed with a firewall and a roof-mounted ventilating fan was installed
for rapid removal of unwanted gases.
An induction melting furnace was selected to simulate the combustor
because it is a convenient means of preparing a molten iron bath and
maintaining it at a specified operating temperature. The induction
furnace was designed to prepare a three ton molten bath in four or five
hours. Whenever it is desired to remove the contents of the induction
furnace, the furnace is disconnected from the rest of the system, tilted,
and the molten metal poured into a ladle for casting into small pig
molds or large starter block molds. The iron pigs and starter blocks
are reused in the furnace.
The combustor offgas contains carbon monoxide and hydrogen in addition
to other constituents. Therefore, prior to atmospheric elimination,
the gas is burned to carbon dioxide and water, cooled, and scrubbed.
This is accomplished in the offgas processing system shown in Figure A-l.
The combustor (1) is connected to the offgas handling system using a gas-
tight transition section (2) mounted on top of the furnace. The
design of this piece of equipment is such that a gas-tight seal could be
made with the induction furnace while the total surface area was minimized
to limit radiant heat loss. The latter consideration necessitated
the redesign and fabrication of a second transition section when the
-1-
-------
i...
i
2
3
4-
€
t
7
s
1
10
\\
(2.
COMBUSTION &UKU
PB/MA2y SP/EflV CHAMBER.
'
£>CJC7~
r/>u
'SPILL PIT
; f'aevscw we.
I/IEW
6-2
A - i
-------
excessive heat loss resulting from the initial design led to slag
crusting problems which caused operational difficulties. The transition
section (which also provides support for coal and air injection and
sampling equipment) directs the combustor offgas into a flare section (3)
for complete combustion to carbon dioxide and water. Air is injected
into the flare section which contains a natural gas pilot burner (5)
as an igniter for the gas-air mixture. Complete combustion of. the offgas
occurs in the second flare section (4).
The offgas from the flare flows to a quench section (6) where it is
cooled by. contact with water sprays. The water-saturated gas stream
enters a wetting section (7) where it undergoes further water cooling
(provided by radially mounted sprays) to remove large-sized particulate
matter (if any). The cooling water is separated from the saturated gas
stream in the dewatering section (8). The scrubber section (9) located
downstream of the dewatering section (8), removes any particulate
matter which may be present in the offgas. The clean gas exits the
scrubber section into draft fan (10) and is eliminated from the system
through the stack (12). The draft fan maintains negative pressure
throughout the combustor system and discharges the cooled and scrubbed
gaseous products of combustion to the atmosphere.
The transition and flare sections were'lined with high alumina castable
refractory designed to withstand a 2800-3000°F combusted offgas temp-
erature and yield a 400°F outer steel shell temperature. The temperature
of gas leaving the quench section was controlled in the range 150-
200°F; hence, a refractory lining was not used for downstream equipment.
Injection System-
The pneumatic injection system is shown in Figure A-2. To minimize the
transport gas requirements, a dense-phase pneumatic system was selected.
In such a system, material is fed batchwise into an injection tank,
pressurized and fluidized using part of the transport gas. The material
is subsequently conveyed out of the tank via a hose to the injection lance.
Additional (secondary) transport gas is added at the tank exit to the
hose to insure adequate transport velocity and to balance pressure drops.
.The material injection blow tank was located on a weight scale with a
loss-of-weight recorder. During injection, an instantaneous weight can
be obtained from the scale dial and a permanent loss-of-weight record
for the entire run is obtained from the recorder.
The pneumatic injection system is used to inject several materials into
the molten bath. Graphite and several coals were injected using the
system to determine solubility and gasification rates. Slags are
synthesized by injecting the requisite ingredients beneath the metal bath.
The materials dissolve as they float to the surface where the fluid slag
collects to the desired height.
-3-
-------
COMPKESSSD lS-/?S-
0)1
I
COUTEOL l/dute _-
SygTEVv?
A-2.
-------
Injaction Lances
To a large extent, the successful operation of the experimental combustor
.depends on suitable injection lances. Since both solids and air are
injected beneath the molten iron, a submergible lance is required. Two
types of non-cooled lances were used—graphite fluxing tubes and
refractory or ceramic coated lances. The graphite fluxing tubes were
found to be better suited to submerged injection and were used almost
exclusively. Unlike the ceramic coated lances which were prone to thermal
shock failure, or refractory coated lances which were prone to slag
attack (chemical reaction) failure, the graphite fluxing tubes showed
adequate useful life for our purposes. This useful life was limited
only by the graphite dissolving in the molten iron bath or reacting
with the air flowing through the lance.
For extended experimental and commercial life, internally-cooled metal
lances should be used. In coal injection applications, the internal
cooling also offers additional protection against the possibility of
coking the coal within the lance—which would plug the lance. Water
cooled lances are currently used above molten iron in basic oxygen steel
refining. However, it is known that the introduction of water into
molten iron can present serious safety hazards. Even though safety
features can be incorporated within the lance design to minimize these
hazards, the need for a totally fool-proof system in power plant appli-
cations is paramount to maintain reliable and constant operations.
An experimental study is now being conducted to design and evaluate a
non-aqueous cooled submergible lance system for both experimental and
commercial combustor operations. Typical non-aqueous cooling fluids
such as oils, molten salts, and low-melting point alloys could be used
which, in general, eliminate explosive possibilities because they do
not generate large amounts of gases capable of instantaneous expansion.
Oils are intrinsically safe because the coking and cracking that occurs
at the high temperatures involved produce only small amounts of gases
that under typical expected combustor operation should not generate an
explosive condition. Molten salts may create problems. Nevertheless,
they are suitable as cooling agents insofar as they will prevent an
explosive condition. Probably the most desirable cooling agents are
low melting point metals or alloys. Alkaline metals, such as sodium,
vaporize when exposed to molten iron and will subsequently react with
atmospheric oxygen or water vapor. Much more suitable are the low-melting-
point, high-boiling-point alloys such as lead-bismuth eutectic which
will safely remain in the iron if leaked into the molten bath. These
materials, operating in a closed-loop cooling and heat exchange system
offer much promise for a lance system.
Process Instrumentation and Control System
The design of the experimental combustor process instrumentation and
control system is shown in Figure A-3. The system does not include the
coal injection system which is treated separately.
-5-
-------
-------
Process water, is taken directly from the local water supply main (1) and
used first to cool the induction furnace and its power supply. It
exits (2) at temperatures in the range of 90-100°F which are suitable
for cooling and scrubbing the offgas. The water exiting the furnace
power supply panel is used for the evaporative cooling of combusted
offgas in the quench section (3), in the wetting section (4), in cooling
of the internal baffle of the dewatering section (5) and for scrubbing
any particulates from the gas in the wet scrubber (6). Water flow
rates are measured with rotameters using an integrating meter as a
check. Process temperatures are measured at the points marked "T",
on four potentiometric recorders. All high-temperature gas streams are
monitored with platinum 13 percent rhodium thermocouples, and low
temperature gas and water streams are monitored with copper-constantan
thermocouples. Gas flow measurements of the combustion air injected
into the flare section and the total offgas stream in the reduced draft
fan are determined using orifice meters.
Raw Materials and Supplies
Raw materials used for experimentation are: iron, slag ingredients,
sulfur, air and nitrogen. Iron was purchased in pig form and had .the
typical composition shown in Table I-A. The slag materials were lime,
calcium sulfide, sulfur, silicon, and alumina. The oxides were
purchased from refractory suppliers to minimize contaminants without
.going to the cost of purchasing reagent grade materials.
In this study, several coals and graphite of the compositions shown on
Table II-A were used. The air and nitrogen were received via tank
trailer. The trailers were connected to an outside station from which
various streams were piped to the system where required.
Solids injection beneath the molten bath was effected through graphite
"fluxing tubes".
TABLE I-A
TYPICAL PIG IRON COMPOSITION
Carbon 4.37
Silicon 1.17
Sulfur .026
Phosphorous .085
Manganese .93
-7-
-------
TABLE II-A
COALS USED IN EXPERIMENTATION
Stillwater
Coal
Consol
Sea
Coal
Bureau of
Mines -1/4"
Coal
Bureau of
Mines -1/8"
Coal
Bureau of
Mines -1/16"
Coal
Proximate Analysis
Moisture
Volatile Matter.
Fixed Carbon
Ash
1.97.
38.8
54.1
5.2
100.0
1.9%
34.0
51.3
12.8
1.9%
34.0
51.3
12.8
1.9%
, 34.0
51.3
12.8
100.0
100.0
100.0
I
00
Ultimate Analysis
Hydrogen 5.1
Carbon 67.8
Nitrogen 1.0
Oxygen 8.2
Sulfur ' 3.5
5.5
79.0
1.6
7.1
1.6
4.9
72.2
1.5
7.1
1.5
4.9
72.2
1.5
7.1
1.5
4.9
72.2
1.5
7.1
; i.s
Screen Analysis
7o Retained
Tyler Mesh
3
10
20
35
48
60
100
200
PAN
0.0
1.7
10.9
26.3
17.8
0.0
0.0
12.8
29.6
20.0
14.1
12.0
11.5
100.0
0.0
6.8
31.3
25.2
13.1
.8.3
7.1
8.1
100.0
-------
The specific lances used were pitch impregnated graphite and were
composed of essentially 99 percent carbon and a typical 0.13 percent
ash (the pitch impregnation material being graphitized out during
either processing by the manufacturer or preheating in the experimental
combustor). The lances were obtained in either six or eight foot
lengths with a three inch O.D. and a 3/4 inch I.D. Refractory material
used to line the induction furnace during experimentation included
high alumina dry ram material (Inducto - 87A), high alumina castable
material (Harmix CU) and a high alumina brick (Korundal XD). The
specific properties and compositions of these refractories are shown
respectively in Tables III-A, IV-A, and V-A.
TABLE III-A
INDUCTO - 87A REFRACTORY
Technical Data:
Maximum Temperature
Density -
Chemical Nature -
Thermal Conductivity -
Mean Specific Heat
Modulus of Rupture
(After Firing)
Thermal Expansion
Chemical Analysis:
.Alumina (Al?0,)
Silica (Slop
Magnesia (MgO)
Zirconium Oxide
(Zr20)
Calcium Oxide (CaO) -
Iron Oxide (Fen00) -
2 3
3500°F
195 Ib/ft
Amphoeteric
26 btu/in (§2500°F
.22 btu/lb/ft
700 lbs/in2 @2500°F
.05% @2000°F
90.5%
6.6%
trace
trace
trace
trace
Experimental Procedure
The experimental program consisted of three types of specialized
operations: carburization tests, decarburization tests, and slag
injection. All these experimental operations began with the basic
combustor startup; the individual differences beginning at certain
points per the following discussions.
The combustor startup procedure began with a detailed clean-out of
the induction furnace interior to remove all foreign material. A
specified amount of iron was then charged into the furnace in starter
—9—
-------
TABLE IV-A
HARMIX CU REFRACTORY
Technical Data:
*Physical Properties: (Typical)
Weight Required for Ramming
If Shipped Dry
If Shipped Wet
173 pcf
180 pcf
Approximate Amount of Water Required
for. Ramming Dry Mix
Per 100 Ibs.
Per 45.36 Kg.
2 U.S.Qts,
Bulk Density
After Drying at 230°F. (110°C)
Ibs./cu.ft.
173
Modulus of Rupture
After Drying at 230°F.
After Heating at 2300°F. (1260°C)
Ibs./sq.in^
400 to 800
1500 to 2500
Cold Crushing Strength
After Drying at 230°E (
After Heating at 2300°F. (1260°C)
Permanent Linear Change (7»)
After Drying at 230°F. (110°C)
After Heating at 2300°F. (1260°C)
2500 to 4500
8000 to 10,000
Negligible
0.0 to +0.5
*Chemical Analysis:
(Approximate)
(Calcined Basis)
Silica
Alumina
Titania
Iron Oxide
Lime
Magnesia
(S102)
(A1203)
(Ti02)
(Fe203)
(CaO)
(MgO)
11.4%
84.6
2.8
1.1
Trace
Trace
Alkalies (Na20+K20+Li20) 0.1
-10-
-------
TABLE V-A
KORUNDAL XD REFRACTORY
Classification:
Physical Data:
(Typical)
High Alumina Brick
Bulk Density
Pounds/cu.ft.
Grains/cc
Apparent Porosity, 7o
Cold Crushing Strength on Flat
Pounds/sq.in.
Kilograms/cm2
Modulus of Rupture
Pounds/sq.in.
Kilograms/cm
Reheat Test
Permanent Change @
At 3140°(1725°C)
At 3300°F (1816°C)
Load Test, 1% hr., 25 psi (1.76 kg/cm2)
% Linear Change
At 3000°F. (1650°C)
At 3200°F. (17600C)
o
Load Test, 24 hr., 25 psi (1.76 kg/cm )
At 3200°F. (1760°C)
Panel Spalling Test
7o Loss
Preheat - 3000°F. (1650°C)
181 to 185
2.90 to 2.96
14 to 18
9000 to 14000
634 to- 986
2500 to 3500
176 to 246
+0.5 to +1.5
+1.2 to +2.4
0.0 to +0.4
-0.2 to -0.4
-1 to -3
0.0
Chemical Analysis:
(Approximate)
Silica
Alumina
Titania
Iron Oxide
Lime
Magnesia
Alkalies
(Si02)
(A1203)
(Ti02)
(Fe203)
(CaO)
(MgO)
8.57.
90.8
0.1
0.2
0.07
0.07
0.15
-11-
-------
block and pig form. The transition base was then sealed, as air tight
as possible, to the induction furnace using a non-binding asbestos
refractory mud; the transition stack was then connected to the flare
section. The induction furnace control panel and all auxiliary
equipment were then inspected and started up proceeding to a molten
iron bath. In the interim between furnace startup and the appearance
of a molten iron bath, the sampling system analytical equipment was
calibrated. Once a molten bath was established, the entire experimental
combustor system was checked and started in sequence, beginning with
the water flows, through the safety flare burner, and ending with the
I.D. fan. Once the entire experimental combustor was operable, the
desired tests were performed.
Decarburization testing began by preheating a lance sized to effect air
injection at a specified depth. During the final minutes of the preheat
cycle, the flexible air transport line from the trailer was connected
to the lance and the air flow rate preset on a rotameter. Immediately
prior to decarburization, iron and slag samples were taken. The lance
was then lowered to the specified depth with air flow to prevent
lance pluggage. The experiment was then carried out as prescribed.
At the end of the test, the lance was withdrawn, weighed, and measured,
while iron and slag samples were taken.
Carburization testing began with the charging of coal into the pneumatic
injection system. During the lance preheat cycle mentioned above, the
injection tank was pressurized to the specified amount using either
nitrogen or air. Similar to decarburization testing, the flexible
transport line from the injection tank was connected to the graphite
lance during the final minutes of the preheat cycle. Experimentation
began by lowering the lance into the iron at the end of the preheat
cycle. As with decarburization testing above, iron and slag samples
were taken before and after experimentation.
Slag synthesis was performed in the same manner as for the carburization.
Prior to charging the slag ingredients into the injection tank, they
were individually weighed and thoroughly mixed. Following mixing, the
mixture was sampled and charged into the injection tank. As in the
carburization test procedure, the ingredient mixture was injected
beneath the molten bath. The materials then dissolved and floated to
the top.
Gaseous Sampling and Analysis
During experimentation, samples of gas streams throughout the system
were taken from gas sample taps and sampling ports located in the
transition section, flare, and stack sections. The combustor offgas
was continuously monitored by the various instruments located on the
panel shown schematically in Figure A-4. The instruments determined
the concentrations of oxygen, carbon monoxide, carbon dioxide, hydrogen,
sulfur dioxide, nitric oxide, and nitrogen dioxide in the gaseous
stream. Located on a separate panel were seven strip chart recorders
-12-
-------
FIGURE A-4
ANALYTICAL INSTRUMENT PANEL SCHEMATIC
-------
FIGURE A-4 (Cont'd)
ANALYTICAL INSTRUMENT PANEL DESCRIPTION
Item No.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
Filter
Drierite
Temperature Gauge
Vacuum Gauge
Main Pump Switch and
"ON" Light
Relief Valve
Pressure Gauge
Check Valve
Main Flo wine ter
Main Total Indicating Flow
Sample Bomb
Sample Bomb Flowmeter
Hydrogen Flowmeter
Hydrogen Analyzer
Carbon Monoxide and Carbon
Dioxide Flowmeter
Carbon Monoxide and Carbon
Dioxide Analyzer
Carbon Monoxide Trim Pot
Carbon Dioxide Trim Pot
Oxygen Analyzer
NO, NO and S02 Control
Module
Nitric Oxide Analyzer
Nitrogen Dioxide and
Sulfur Dioxide Analyzer
Main Sample Inlet
Sample Bomb Outlet
Main Sample Outlet
Control.Module Inlet
Control Module Outlet
0.3 microns
10-20 mesh
50-500°F
0-30 in. Hg
0-30 psi
0-50 SCFM
0-50 SCFM
20-280 cc/min
0-25%
20-280 cc/min
0-50%
0-5 SCFM
0-500 PPM
0-500 PPM
-14-
-------
integrated with the instruments which provided a continuous record of
these concentrations. The panel was also designed so that the sample
gas stream could be split. Part of the gas was routed to the con-
tinuous analyzers and another part of it was sent to a series of glass
sample bombs which could be opened and closed to take retainer gas
samples at various times during an experiment.
-15-
-------
Analytical System Operation
Flow meters are located on the panel so that flows can be independently
adjusted to the instruments. A stainless steel bellows pump is used
to pull the gas from the combustor and to deliver it to the instruments
and glass sample bombs. This type of pump was chosen since no lubricating
oils or foreign material contacts the sample gas except the AM-350
stainless steel bellows which are inert to reaction with combustor gases.
Also located on the panel upstream of the pump is a filter and a
drierite column to remove water vapor from the stream. Pressure gauges,
vacuum gauges, and thermometers are also.mounted on the panel to monitor
the sample gas characteristics.
Gas sampling ports are located at several positions in the area of the
transition section of the combustor. Sample taps at this section are
prone to plugging; therefore, a manifold arrangement is provided so
that switching from one sample port to another can be done quickly during
the course of an experiment if desired. The sample line itself is
1/4 inch stainless steel tubing and runs from the combustor sample
ports into an adjacent room in which the instrument panel is located.
The approximate hold up time in the line is 15-30 seconds. The response
time of the instruments ranges from less than one second to about ten
seconds. A brief description of each of the continuous gas analyzers
is given below.
Analyzers
The instruments for monitoring sulfur dioxide, nitric oxide, and nitrogen
dioxide in the combustor offgas were manufactured by Envirometrics, Inc.,
(Marina Del Key, California). The measurement of these gases is
accomplished by using patented plug-in sensors which Envirometrics
calls Faristors . According- to Envirometrics, the Faristor is a
liquid-state device containing a chemically-sensitive activating surface
layer upon which pollutant molecules are strongly adsorbed by non-
thermal catalytic action. This results in a change of oxidation state
producing a charged surface relative to a reference layer. The magnitude
of this charge is determined by the rate at which the gas molecules
reach the activating surface, this in turn being directly proportional
to the pollutant concentration at a constant sampling rate. Transfer
of this charge across the activating layer results in a current flow.
The net effect is somewhat analogous to a negative (nonohmic) resistance
in a solid-state device. Faristors are made individually selective by
virtue, of the specific free energies of reaction and varying chemical
kinetics associated with the catalysis of the individual pollutants*.
The equipment includes three separate instruments (shown as item
numbers 20, 21, 22, on Figure A-4). Item number 20 is a sampler control
module which contains a stainless steel bellows pump, water trap, and
*"Instruction Manual, Coupled SO./Nitrogen Oxide Analyzers", from
Envirometrics, Inc.
-16-
-------
and dust filter. This module conditions the gas sample before it .is
sent to the NS-280 analyzer. The NS-280 analyzer contains Faristors
to measure S0« and NO„ concentrations (2-10,000 ppm). The sample gas
then passes into the NS-200 analyzer which determines NO (2-10,000 ppm).
The NS-200 and NS-280 analyzers are equipped with analog output signals
(0-10 mv) which are connected to strip chart recorders.
Calibration gases are located behind the instrument panel and are
connected directly to the sampler control module. By adjusting the
three way ball valves on the module, the calibration gases can be
sent to the analyzers. The following calibration gases are required:
nitrogen, nitric oxide in nitrogen, nitrogen dioxide in nitrogen, and
sulfur dioxide in nitrogen. A complete calibration is performed weekly
but zero and span checks.are done approximately every hour (while the
instruments are in service).
Hydrogen Analyzer
A Bendix Thermal Conductivity Monitor is employed to determine the
amount of hydrogen present in the sample gas stream. The principle of
thermal conductivity is used in many continuous analyzers. The value
of the thermal conductivity of a gas is indicative of its ability to
remove heat from its surroundings. Hydrogen has a very high thermal
conductivity and can remove heat rapidly compared to gases such as
nitrogen and oxygen which have lower themal conductivities and remove
heat slowly. The thermal conductivity detector has a sealed reference
cell filled with air and a measuring cell through which the sample gas
passes. The detector elements are WX, 32 ohm filaments which are
part of a Wheatstone bridge, which in turn heats the filaments. As
hydrogen passes through the measuring cell, heat is conducted from the
filament, changing the resistance and changing the current flow in the
bridge. The change in current is proportional to the concentration of
hydrogen.
The instrument has a range of 0-25 percent hydrogen. Calibration is
done on a daily basis (when operating) and includes the use of a zero
gas (air or nitrogen) and a span gas (a gas mixture containing 18 percent
H-, 28 percent N-, 54 percent CO).
CO, CO- Analyzer
Analysis of the sample gas stream for carbon monoxide and carbon dioxide
is accomplished with a Peerless #206 two-gas analyzer manufactured by
the Peerless Instrument Company, (Elmhurst, New York.) The operation
of the instrument is based on the selective absorption of radiation of
certain wave lengths by CO and C0?. According to Peerless, "The
infrared (IR) radiant energy from the sources is formed into two beams:
the reference beam and the sample beam. The gas in the sample cell
absorbs IR in proportion to the gas concentrations while the reference
cell contains a neutral gas... The difference in energy in the two
beams coming out of the two cells (and ultimately reaching detectors)
-17-
-------
**
is proportional to the gas concentrations in the sample . The signal
beams striking the detectors are ultimately transformed into electrical
signals which are noted on two meters located on the front of the instrument.
The electrical signals are also sent to strip chart recorders for the two
gases.
The range of the instrument is 0-50 percent for both carbon monoxide and
carbon dioxide. Zero calibration is done with air or nitrogen while the
span calibration is done by either using a gas mixture containing CO and
CO- or by using the convenient "internal" calibration knob located on the
front of the instrument. Calibration is done on a daily basis while operating.
Oxygen Analyzer
The oxygen analyzer used was manufactured by Servomex Controls and was
distributed by Bendix. The operation of the instrument is based on the
fact that oxygen exhibits paramagnetism to a much greater extent than other
common gases. Thus, the instrument is made to respond to the magnetic
susceptibility of the sample being tested. This response is eventually
transformed into a meter reading denoting the percent oxygen in the sample
directly. This is accomplished according to the following description given
by Bendix: "Inside the instrument, a small dumb-bell shaped body is suspended
on a platinum ribbon in a non-uniform magnetic field. It experiences a torque
which is proportional to the volume of magnetic susceptibility of the gas
surrounding the dumb-bell. This torque is counteracted by the electro magnetic
effect of current which is made to flow through a single turn of platinum wire
wound around the dumb-bell. The current required to do this is proportional
to the original torque and is, therefore, a measure of the susceptibility of
the sample gas."
The range of the instrument is 0-25 percent oxygen. Calibration is done on
a daily basis using two calibration gases: (1) nitrogen to obtain a zero
check, and (2) air to obtain a span adjustment (21 percent oxygen).
Iron and Slag Sampling Analysis
Iron samples are obtained with "batester" immersion sampling devices. The
devices operate on a simple capture and freeze principle in that when the
end reaches molten metal, a protective cap melts and allows metal to enter
a sample chamber where it solidifies into a pin sample with an enlarged re-
tainer sample at one end.
Slag samples are obtained using a rod with a cupped end. Representative
samples are taken by immersing the rod to a specified depth at a given location
in the fluid slag. Both the iron and slag samples are taken through sample
ports located on the transition section and are analyzed primarily for carbon
and sulfur.
"Operation and Service Manual", Peerless Instrument Co., Inc.
-18-
-------
The procedure used to determine the amount of carbon that is contained in
the above iron samples taken from the molten bath is based on the combustion
method of carbon analysis. In this method, the carbon contained in the
sample is burned to carbon dioxide; the amount of carbon dioxide produced
is measured and the percent carbon in the sample is subsequently determined.
The experimental equipment is shown in Figure A-5. The oxygen used for
combustion is treated in a Leco purifying train (//516-000) prior to
entering the furnace in which the combustion occurs. The purpose of the
purifying train is to remove trace amounts of moisture, carbon dioxide,
and acid gases that may be present in the oxygen. These gases, if not
removed, would interfere in the present method. The furnace in which the
combustion takes place is a Burrell Model H-9 resistance furnace. The
furnace is heated by four symmetrically arranged glow bars which surround
and are parallel to the ceramic reaction tube. The temperature is con-
trolled by two voltage taps located on the front of the furnace. The
exit end of the combustion tube is packed with glass wool to trap dust
particles produced during combustion. The gas leaving the furnace then
passes through a MnO« trap to remove sulfur oxides and then to the Leco
gasometric carbon analyzer (#572-100). The Leco carbon analyzer traps
all of the gases that exit the furnace (except sulfur oxides) during the
entire combustion procedure. These gases consist primarily of oxygen
and carbon dioxide. The volume of this gas at standard conditions is
measured on a graduated burette located below the large bulb on the
analyzer. The gas is then contacted with a concentrated solution of
caustic (KOH) to remove the carbon dioxide. After removal of C0~, the gas
is sent back to the bulb outfitted with the graduated burette. The
difference in volume represents the amount of CO,, that was originally
present in the offgas. The amount of carbon present in the sample is,
of course, directly proportional to the volume of C0~ produced.
The experimental procedure consists of (1) sample preparation, (2) com-
bustion of the sample at 2400-2700°F with an oxygen flow rate of 1 liter/
minute, (3) measurement of the offgas produced, (4) scrubbing the C0«
from the offgas, and (5) measuring the volume of offgas less carbon
dioxide. Sample preparation consists of crushing the disc-pin iron sample
in a mortar and pestle into small fragments. Approximately 200 milligrams
of sample is placed in a ceramic combustion boat. The following com-
bustion accelerators are added: one scoop (1 gram) of copper coated tin
and two strips of copper. The combustion boat containing the sample is
then placed into the furnace for one to two minutes to preheat, followed
by introduction of an oxygen flow rate of one liter per minute. After a
ten minute burn time, the products of combustion are scrubbed with the
caustic solution and the volume is then remeasured to find the volume
of carbon dioxide produced in the combustion. The percent carbon in the
iron sample is then calculated from the amount of carbon dioxide produced.
-19-
-------
•Boat Containing Sample
Leco Purifying Train
O
I
Combustion Furnace
Burrell H-9
Leco Gasometric
Carbon Analyzer
FIGURE A-5
CARBON DETERMINATION EQUIPMENT
-------
The determination of the percent sulfur contained in iron and slag
samples is based on the combustion-iodometric method. In this method,
the sulfur in the sample is burned to sulfur dioxide which is subsequently
titrated with potassium iodate to determine the amount of S0? that
evolves, from the sample. The reactions proceed according to the following
equations:
KIO + SKI + 6HC1 = 6KC1 + 31 + 3^0 (1)
S02 + I2 + 2H20 = H2S04 + 2HI (2)
Thus, the S0« that is formed during combustion is bubbled into a
titration vessel containing a weak solution of HC1 to which has been
added potassium iodide and starch. The titrant is KIO_. Equation (1)
above produces free iodine which forms a deep blue complex with starch.
As sulfur dioxide enters the solution, it is rapidly oxidized to s.ulfuric
acid. This oxidation reaction also destroys the free iodine and, there-
fore, the blue color caused by the starch-iodine comples. Addition of
KIO titrant restores the blue color until additional SO. evolves from
the combustion of the sample. Persistance of a blue color in the
titration vessel indicates the absence of sulfur dioxide. The total
amount of sulfur dioxide that evolves during the combustion is determined
by the amount of KIO,. titrant added (three moles of SO- are equivalent
to one mole of KIO ).
The experimental equipment is essentially the same as shown in Figure A-5,
(determination of carbon in iron). In place of the Leco carbon analyzer
is a Leco sulfur titrator model (#517). This titrator operates on the
combustion-iodometric method described above.
Sample preparation for determination of sulfur in iron consists of
crushing the sample and placing 100-200 milligrams in a combustion boat
with 1 gram of copper coated tin and two strips of copper combustion
accelerators. The combustion boat is then placed in the furnace at 2200°F
and allowed to preheat for two minutes prior to initiating the flow of
oxygen at one liter per minute. The boat is placed inside a thimble
located in the hot zone of the furnace as shown in Figure A-6. The
purpose of the thimble is to lower the rate at which the S0» would enter
the titration solution. Without such an arrangement, S0_ would evolve
too rapidly and would escape before the titration procedure could be
conducted properly. This arrangement is used for determining sulfur
in both iron and slag samples and permits the entire titration to be
extended to ten minutes.
The determination of sulfur in slag is done similarly except that,
sample preparation consists of crushing the slag samples and removing
foreign iron particles by magnetic separation before the sample is
placed in a combustion boat with 1/2 gram of electrolytic iron accelerator,
1/2 gram of copper tin accelerator, and 2 strips of copper accelerator.
-21-
-------
i
K3
tsJ
XNimble With Small Hole
In End
z
^— r
Combustion Boat
Combustion Furnace
(Burrell Model H-9)
I
—To Sulfur Titrator-
FIGURE A-6
SULFUR DETERMINATION EQUIPMENT
-------
The chemical analyses for SiCL, Al-0.,, Fe^O , and CaO in combustor slag
samples utilizes both volumetric and gravimetric methods.
The first part of the analysis is a gravimetric analysis to determine
SiQj. The crushed slag sample is placed in a one to one solution of HC1
and H^O. The solution is filtered with coarse filter paper. The
material remaining on the filter paper consists chiefly of SiO_ and
other acid insolubles. The filter paper and its contents are placed in
a platinum crucible and fused with sodium carbonate to form soluble
carbonates from the acid insoluble constituents in the slag. The fused
.material is then redissolved in concentrated HC1. Fifty milliliters
of water are added and the solution is brought to a gentle boil. This
solution is then filtered again to remove insoluble SiO~. The filter
paper is placed in a platinum crucible and charred at 200°C in a muffle
furnace. After the paper has been charred off, the crucible is placed
in a muffle furnace at 1000°C for several hours until a constant weight
is achieved. The contents in the crucible are then reacted with 5-
10 cc of concentrated hydroflouric acid. Silica in the crucible forms
gaseous SiF, and escapes. After-the excess HF has been driven off, the
crucible is reweighed. The difference in weight represents the loss of
silica as SiF,. From this weight loss and the original sample weight,
the percent SiO~ in the sample can be determined.
The filtrate from the above filtration contains soluble forms of the
calcium oxide, iron oxides, and alumina originally present in the slag.
Ammonium hydroxide is added to the filtrate and the hydroxides or iron
and alumina are precipitated. The precipitate is then filtered from the
solution, washed, dried, and weighed. The weight recorded is proportional
to the total weight of iron and aluminum in the original sample existing
as Fe203 and A120~ (denoted as R2°3^ *
The filtrate from the above precipitation is retained to determine the
amount of calcium in solution, and therefore, the amount of CaO existing
in the original sample. Calcium is determined by performing a
complexiometric titration with EDTA. This is a well known titration and
is routinely performed. The filtrate is diluted to volume in a volumetric
flask arid an aliquot is extracted. The aliquot is buffered, calcine
indicator .is added, and the titration with EDTA is performed. The
end point is achieved when the green flourescence seen in UV light and
caused by a free calcium-calcine indicator complex dissappears.
The determination of the relative amounts of Al_0- and Fe_0 present in
the sample from the total amount of these components (R 0«) is accomplished
by analyzing for Fe~0~. A new slag sample is taken and digested in a
50 percent HC1 solution. The iron is then reduced to the ferrous state
with stannous chloride (SnCl2) and titrated with a standard solution of
potassium dichromate (K., Cr,,07) . This is known as the ferric oxide
referee method (ASTM #CI14).
-23-
-------
-24-
-------
APPENDIX B
LABORATORY STUDIES
Summary of Laboratory Results
To provide guidelines for conducting the experimental combustor program,
a laboratory program was completed to ascertain the physico-chemical
characterization of slags. In-addition to the operational guidelines,
engineering data were also generated for the process engineering studies.
The results of the laboratory work can be summarized as follows.
Slag Composition
Slags having good fluidity or flow characteristics and partition ratios in
excess of 20 can be produced with basicities of 0.2 or less. A 0.2
basicity slag composed of 20 percent CaS, 13.4 percent CaO, 44.4 percent
Si02, and 22..2 percent Al.O. is suitable for combustor operation.
All percentages are given on a weight basis. The use of higher basicity
slags will result in poorer flow characteristics. No attempt was made
to study additional effects of iron oxide on flow characteristics. How-
ever, it can be safely surmised that improved fluidity will be observed
under actual conditions due to the inclusion of iron oxides and alkaline
constituents into the slag.
Slag Viscosity
For slags having the composition specified above, the apparent viscosity
over the temperature range of 2700 - 2800°F will be 50 to 100 poise. These
data were obtained with an oscillating bob viscosimeter which tends to
report higher viscosity values if the slags contained suspended solids.
Inasmuch as the high sulfur bearing slags contain undissolved CaS, the
measured viscosity is believed to be substantially higher than anticipated
in the combustor.
Equilibrium Partition Ratios
For typical lime-silica-alumina slags covering a basicity range of 0.2 to 1.0,
equilibrium sulfur partition ratios in excess of 20 will be realized. These
values were for a slag containing 20 percent calcium sulfide (approximately
8.9 percent sulfur).
Heat Capacity
For slags containing 20 percent calcium sulfide, the heat capacity shows no
dependence on slag basicity. Consequently, for process engineering cal-
culations, the heat capacity of 20 percent CaS bearing slags is adequately
defined by
C = 0.119 + 3.66xlO~4T - 15.90xlO~8T2
P
which is valid over the temperature range of 200 to 1200°C. In the equation,
-25-
-------
specific heat has the units of calories per gram per degree C, and T is
defined as degrees K. .
Crushing Energy Requirements
For sizing the necessary crushing and grinding equipment, the slags can be
assumed to have characteristics comparable to silica. Comparison of crushing
energy requirements for slag and silica indicates that the same Rittinger's
Number characterizes both materials. The fact that the slags behave as silica
during crushing should facilitate the selection of commercial equipment.
Slag Granulation
Experimental results indicate that a vacuum shot tower granulation method
is an effective means of increasing the total surface area of typical
CaO-SiOp-Al-O slags from two to five times the value for crushed slag.
Based on these findings, it appears that slag granulation can serve as a
means of providing increased total surface area slags for desulfurization.
Slag Desulfurization
Experimental results indicate that slag desulfurization is mass transfer
controlled and no chemical reaction kinetics limitations are present.
Although complete design information is not yet available, it is believed
that reasonably sized reaction vessels can be used. Experimental results
have shown that 99 plus percent desulfurization occurs with an offgas con-
taining approximately 50 percent sulfur.
SLAG FLUIDITY
Introduction
The operating cost of the process is influenced to a large degree by slag
fluidity and the ability of the slag to contain a maximum amount of sulfur.
The slags must be fluid enough to facilitate removal from the combustor but _ ,
at the same time they must be viscous enough to minimize refractory corrosion ' .
Consequently, a compromise slag fluidity is required to maintain the most
desirable operating conditions. Ideally, the slag should satisfy not only
the fluidity requirement but it should also contain the stoichiometric
equivalent, of lime to convert all of the sulfur introduced to the molten iron
bath by the coal to calcium sulfide. However, from a practical viewpoint an
excess of lime will be required.
Because the ATC process requires a low excess lime content in the slag, very
low basicity (defined as the ratio of lime to the sum of the silica and alumina
contents) high sulfur slags will be required. A great deal of data are
available pertaining to the effect of slag composition on the viscosity of
slags containing less than 2 percent sulfur. However, combustor slags will
contain about 6-8 percent sulfur. The only available data for such high-
sulfur slags were obtained using high-basicity magnesia-bearing slags . In
power plant application of the ATC combustor, the use of magnesia-bearing
high-basicity slags would have an adverse economic effect by increasing the
quantities of flux added to the combustor. The choice of limestone and the
-26-
-------
composition of the coal ash restricts the slag composition to the ^
Al.,0., system. Most coals used for power generation contain an ash whose
composition is relatively fixed at a silica to alumina ratio of about two .
Since the flux to be added to the combustor contains only minor amounts of
silica and alumina, the slags produced will contain silica and alumina in
the same proportion as that found in coal. For this reason this ratio was
maintained in all experimental work. Since literature data on the effect
of high sulfur concentration on slag viscosity are scarce and because a
large number of slag compositions are possible, a quick screening test was
required to relatively rank the various types of slags with regard to fluidity.
A simple and rapid fluidity test has been developed by Herty ' and has been
used in steelmaking operations for some time. The test with some modification
has proven to be a valuable aid for the evaluation of slag fluidity characteristics.
The principle of the test is to cause molten slag to flow down an inclined
plane. The extent to which the slag travels down the plane prior to
solidification is an indication of its fluidity (and, indirectly, viscosity).
Obviously, a test as simple as this is influenced by a number of physical
constants as well as heat transfer. Nevertheless, it is used as a relative
guide for comparing differences in fluidity2 and was adopted for use in this
work.
Experimental Procedure
Ten different compositions of slag (Table IB) with varying amounts of CaS
and basicities ranging from 0.2 to 1.2 were used for this investigation.
Five pounds of each of these slags were prepared from reagent grades of CaO,
Al-O.,, SiO-, and CaS by melting at 2700°F with an argon atmosphere in a
1
2
3
4
5
6
7
8
9
10
TABLE I-B
Premelt Composition of Slags
Basicity
.80
.80
1.00
.50
1.00
1.20
.20
.80
1.00
1.00
CaO
40.4
35.6
45.5
26.0
39.9
49.6
13.4
38.0
51.5
42.5
Weight
A100
16.9
14.8
15.2
18.0
13.3
13.8
22.2
16.0
16.5
14.2
-27-
Percent
SiO
*•
33.7
29.6
30.4
36.0
26.6
27.6
44.4
31.0
32.0
28.3
CaS
9.0
20.0
9.0
20.0
20.0
9.0
20.0
15.0
0.0
15.0
-------
covered pyrolitic carbon crucible. It required three soaking periods of
eight hours each and periodic stirring to obtain slags of uniform composition.
Uniformity of the slags was judged by their appearance. The slags were cooled,
crushed and stored prior to conducting the Herty tests. Samples of approximately
300 grams each were remelted and used for the Herty tests. These tests were
conducted at temperatures of 2600°F, 2650°F, and 2700°F.
A schematic arrangement of the modified Herty apparatus is shown in Figure
IB. The apparatus contains two concentric pyrolytic graphite cups each
having a 0.5- inch diameter hole in its base. The cups rest in a refractory
brick cavity designed to accommodate the crucibles. This assembly rests on
a platform attached to the top of an inclined plane. The inclined plane,
(6 x 30 inches) is positioned at an angle of 14 degrees from the horizontal.
Although a 30-degree angle is recommended in the Herty5 procedure, the 14-
degree angle was adopted so that the slag would remain intact after solidification.
It was found that an inclination of 30 degrees was too steep and as a result,
the slag upon solidifying tended to break up and slide down the inclined plane.
In conducting the Herty tests, the two concentric cups (one placed inside the
other, bottom holes 180° apart) and the graphite thermowell are heated to the
respective slag temperatures. The holes in the cups are mis-aligned to prevent
slag flow until the start of the test. These crucibles along with the
refractory brick, (heated to 2000°F) are positioned on the platform at the
top of the inclined plane. Molten slag is then poured into the inner cup.
The holes of the cups are then aligned so that the molten slag will flow
from the crucible and down the incline plane.
The temperature of the slag, the length and weight of the stringer are re-
corded.
Discussion of Results
Work completed for the Water Quality Office has shown that the addition of
calcium sulfide to slags containing magnesium oxide lowered the slag fluidity.
The bulk of this work was conducted with slags having basicities of 0.8 or
more. To determine if a similar effect could be observed for the magnesium-
oxide free slags employed in this work, a series of fluidity tests were
completed on slags over the basicity range of 0.8 to 1.2. The results of
this work, in which the slag calcium sulfide content was varied from zero
to 20 weight percent, are presented in Table IIB. The data indicate that
slag fluidity rapidly decreases with increasing calcium sulfide content.
At a calcium sulfide content of 20 percent (which corresponds to the expected
sulfide level in the commercial combustor), the slag showed a no flow condition.
In these tests, a stringer length of zero inches indicates an apparent viscosity
in excess of several thousand poise. Visual observation of the slags during
the course of experimentation indicated that the slags were not solid but
rather were very viscous and froze before any appreciable stringer could be
formed. As the calcium sulfide content was lowered to 15 percent or less,
-28-
-------
I
NJ
Inclined Plane
Recorder
FIGURE IB-MODIFIED HERTY .FLUIDITY TEST APPARATUS
Thermocouple
.Thermowell
Concentric Cups
-Brick
-------
TABLE IIB
Effect of CaS Content on Fluidity of
High Basicity Slags
Slag
Number*
6
9
3
10
5
Basicity
1.2
1.0
1.0
1.0
1.0
Average
Stringer Length
Inches
0
31.0
24.3
0.0
0.0
3
8
2
0.8
0.8
0.8
9.0
15.0
20.0
11.0
•3.0
0.0
See Table IB for Compositions
acceptable flow characteristics were realized for the 0.8 basicity slags.
The data suggest that increasing calcium sulfide content has a stronger ef-
fect on lowering the fluidity of the higher basicity slags particularly as
the CaS content increases from 9 to 15 percent (compare 1.0 with O.S basicity
slags).
The flow characteristics of multicomponent systems such as these slags are
complicated functions of composition. Because such high basicities would
require an excessive amount of flux (over arid above that required to stoi-
chiometrically combine with the sulfur) for the combustor, no further at-
tempts were made to resolve the functional dependence of fluidity with varying
calcium sulfide content. Instead, work was directed towards obtaining the
fluidity or flow characteristics of economical slags containing 15 to 20
percent calcium sulfide. The results of this work are shown in Figure 2B.
The data show that for slags containing 20 percent CaS (8.8 percent sulfur)
decreasing the basicity increases the fluidity of the slag. As the basicity
changed from 0.8 to 0.2 stringer length increased from zero to two inches.
When the CaS content is decreased to 15 percent, the effect of decreasing
basicity becomes more pronounced and even better fluidity characteristics
are observed.
-30-
-------
oo
d
(U
CO
1.5
% Basicity
FIGURE 2B-EFFECT OF BASICITY AND CaS CONTENT ON SLAG FLUIDITY
-31-
-------
Inasmuch as the fluidity measurement is only a relative guide and the
data indicate that 0.2 basicity slags containing about 20 percent cal-
cium sulfide can be poured and readily removed from the combustor,
fluidity work was terminated in favor of more precise measurements of
slag viscosity. As will be shown later, the choice of favorable slag
compositions (about 0.2 basicity, 20 percent CaS) as determined by the
Herty test is confirmed by the viscosimeter measurements.
SLAG VISCOSITY
Introduction
Fluidity test data discussed in the preceeding section establish the fact
that the flow characteristics of the low basicity slags (0.2) are better
than those obtained at higher basicity levels. It is important that the
high-sulfur-bearing MgO-free slags have good flow characteristics so that
they can readily be removed from the combustor. However, the slag must
not be too fluid to prevent excessive erosion of combustor refractory.
By maintaining a relatively low viscosity, using an MgO-free slag and
producing a slag that contains about 10 percent sulfur, process economics
and operability become attractive. Inasmuch as viscosity data are not
available in the literature for the high-sulfur, MgO-free slags that will
be used in the combustor, it became necessary to characterize the slags
with respect to this property. For this reason, an oscillating bob
viscosimeter with a high-temperature furnace capable of achieving temperatures
of up to 3000°F was designed and constructed.
The theory and technique for determining viscosity with an oscillating bob
will not be discussed in this report, as detailed explanations can be found
in the literature 1^» H. Briefly, the viscosity measurement is obtained
by the logarithmic decrement of oscillation method. With this technique,
an initial torque is applied to a torsion wire from which a serrated graphite
bob is suspended and immersed in a molten slag contained in a graphite crucible.
Depending on the magnitude of the viscous forces resisting the rotation of
the bob, a characteristic dampened oscillation frequency occurs. The ratio
of two successive dampened oscillation amplitudes is a function of the
viscosity of the liquid.
Experimental Procedure .
The viscosimeter illustrated in Figure 3-B was constructed by laboratory
personnel and is similar to that employed by Machin et.alH. However, it
differs from the equipment used by Machin in that the angular displacement
of oscillation is measured by a gradual interruption of a beam of light
focused on a photocell while Machin used a mirror that reflected light to
measure the angular displacement. The output of the photocell is made
proportional to the radial deflection of the suspended torsion wire. The
photocell output is recorded with a Bausch & Lomb VOM 5 recorder. Except
for this modification, and some changes with regard to the length of the
torsion wires and a more convenient manner for setting the oscillating
-32-
-------
Initial
Solenoids
Torsion Wire
. Light Source
Photocell
Photocell
Recorder
.-"-:"• :••'-'• JM-Cerafelt
Location of Globar
Thermocouple to
Control Unit
FIGURE 3-BVISCOSIMETER AND HIGH TEMPERATURE FURNACE
-33-
-------
system in motion, the two methods are the same.
Graphite is used for both the oscillating bobs and the crucibles.
Because the slags did not wet the graphite surfaces, both the bobs
and the crucibles were grooved. The grooves were made parallel to the
longitudinal axis of the bobs and crucibles (Figure 4-B). When the
slag filled the grooves of both the bobs and the crucibles, a coating
of molten slag adherred to the surface of the bobs and to the inside
wall of crucible. This adherent coating is necessary for the measurement
of viscosity, since the measurement depends on shearing stress between
planes of different velocities. The rotation of the bob causes the slag
to shear upon itself. The dampening constant is governed by the shape
and dimension of the bob and crucible, the length and diameter of the
torsion wire, the inertia of the mass suspended from the torsion wire,
the immersion depth of the bob and, to a certain degree, the distance
between the bob and the bottom of the crucible.
For the operating range of the viscosimeter, various pertinent factors
and the calibration curves were determined with standard viscosity oils,
the viscosities of which ranged from 3.9 to 620 poise at a room temperature
of approximately 26°C. The equipment dampening constants were established
for three certified torsion wire sizes (24, 27, and 30 B&S gauge) and two
sizes of bobs.
The viscosimeter furnace is also shown in Figure 3-B. It has a cavity
9 inch deep by 8 inch wide by 8 inch high with an indentation located
in the center of the bottom hearth plate for the graphite crucibles.
The heating elements are located within this cavity. The cavity is ac-
cessible through a 5 inch diameter opening at the top. A gas inlet is
provided at the bottom of the right side panel. The furance is capable
of attaining the holding a temperature of 3000°F. The temperature is
controlled with a Burrell Model SS-200 solid state controller, utilizing
a Pt-13 percent Rh/Pt thermocouple, the tip of which is positioned flush
with the bottom hearth plate.
To reduce atmospheric oxidation of the graphite bobs and the graphite
crucibles at high temperature, the furnace was purged with argon for
at least 1/2 hour prior to introduction of the crucibles containing the
slag. Although the flow of argon was continued throughout most of the
heating period, the'entrance of air could not be prevented entirely.
To prevent gas movement around the oscillating bob, the argon flow was
discontinued during each viscosity measurement.
Attempts were made to check the validity of the equipment dampening
constants under actual high temperature conditions with a standard lead-
silica glass sample (NBS Sample No. 711). However, difficulties in sample
-34-
-------
(a)
(b)
(a) Cross-sectional view of.graphite bobs
large bob: x = 1/2" y = 1/16" z = 1/8"
small bob: x = 3/8" y = 1/16" z = 1/16"
length of bob = 12" Length of grooved section
(b) cross sectional view of graphite crucible
height of • crucible :=. 7" inside
outside ,7-3/4"
= 3"
FIGURE 4-B-VISCOSIMETER BOB AND CUP
-35-
-------
temperature measurement were encountered. The attempt was finally
abandoned when it became apparent that PbO in the glass was reduced to
elemental lead under test conditions. The validity of the equipment
dampening constant under actual high temperature conditions was determined
by using sulfur-free slags of known viscosities. These slags were prepared
in the viscosimeter furnace with an argon atmosphere. During these experi-
ments, the temperature of the slag at various depths was measured immediately
after completion of the viscosity measurements. A two foot long Ft 13
percent Rh/Pt thermocouple was used for this purpose. The thermocouple
was shielded by a- two foot long quartz tube which was in turn protected
from the slag by a thin graphite tip. It was noted that stray currents «
were interfering with the thermocouple and affecting the temperature
measurements. Therefore, the power was turned off for short periods during
each measurement in order to obtain reliable temperature readings.
Results and Discussion
Data obtained during instrument calibration with standard viscosity oil
(Figure 5-B) showed that the distance between the bob and the crucible
bottom has a small effect on the instrument dampening constant, which
diminishes with increasing distance. This is reasonable because the
presence of the crucible bottom is felt by the oscillating bob less and
less as the distance between the bob and the crucible increases.
TABLE III-B
Effect of Immersion Depth on Equipment Dampening Constant (k) at
Varying Bob Distance from Crucible Bottom
27 Gauge Torsion Wire and Small Bob Were Used
Bob Distance
from.Crucible Average k value at immersion depth of:
Bottom 3/4" 1" 1%"
1" 684 -t 61 609 -t 91 797 i 225
1%" 648 -t 47 387 -t 33 878 ± 11
1%" 573 -t 73 479 -k 14 693 -k 17
1 3/4" • 408 * 35 ' 404 -k 21 479 t 43
Table III-B shows that at a bob immersion of 1 inch, the instrument
dampening constant is least influenced by the bob distance from the
crucible bottom for distances greater than 1 inch. Based on these
calibration data, it was decided that an immersion depth of 1 inch
and a distance of at least 1 1/4 inches from the crucible bottom should
be used for all measurements except for slags of exceptionally high
viscosities (1000 poise or more).
-36-
-------
160G
, 1200-
-P
cr
(0
w
c
o
o
to
c
I
Q
4J
§
g
D1
W
800-
400 f
5-B (a)
27 gauge torsion wire,
one inch immersion
0 1 2
Bob distance from crucible bottom, inch
1200-.
c
to
I 8004
o
u
•H
!•
o 400T
OJ
e
•H
P
O1
u
5-B (b)
27 gauge torsion wire,
3/4 inch immersion
1 2
Bob distance from crucible bottom, inch
FIGURE ^-B-EFFECT OF BOB LOCATION ON DAMPENING CONSTANT
-37-
-------
The k value (instrument dampening constant) that would be used in a
particular measurement would depend upon the actual parameter governing
that measurement, e.g. bob size, depth of immersion and the bob distance
from the crucible bottom. To establish the validity of the instrument
dampening constant obtained at room temperature with standard oils,
a series of tests were completed at high temperatures using slags of known
viscosities. Two slags for which literature values for viscosities were
known were selected for these tests. The slag compositions are shown in
Table IV-B for slags V-l and V-2. The results of these measurements are „ 1
shown in Table V-B which presents a comparison of the measure and literature '
values. In these experiments, the room temperature dampening constant
was used to calculate the viscosity of the molten slag. As is evident,
the observed viscosity agreed extremely well with the literature. For
slag V-l the reported reference 12 viscosity was 50 poise at a temperature
of 2700 - 2732°F. The measured viscosity at 2680°F was 46 poise. However,
since the literature only reported one point value, a second slag was tested
for which the viscosity was known over a wide temperature range (Figure 6-B).
TABLE IVB
Premelt Slag Composition
Slag No. % CaO %Si02 %AI2-3 %CaS B/A Ratio
V-l 26 49.5 24.5 0 0.35
V-2 • 40 40 20 0 0.67
TABLE VB
Comparison of Measured Viscosity and Literature Values Q.2, 13)
Slag No. Temperature °F Observed, Poise Lit., Poise Reference
V-l 2680 - 10 46+8 50 at 2732°F 12
V-2 2440 - 10 58.4 - 3.7 59 at 2440°F 13
For this slag (V-2) the observed viscosity at 2440°F was 58.4 plus
or minus 3.7 poise which agrees extremely well with the literature
value of 59 poise. Based on these results, it was concluded that the
effect of high temperature on the instrument dampening constant was
small and that the room temperature calibration was equally valid at
high temperatures.
-38-
-------
w
•H
O
•H
W
O
0
W
•rH
100 1
50 I
10 1
2300
2400
2500 2600
Temperature °F
2700
Data Points Obtained from Reference No. 13
FIGURE 63-VISCOSny OF SLAG V-2 (40% CaO, 40% Si02, 20% A1203)
-39-
-------
Viscosity measurements were then made to determine the effects of tern- .
perature, slag sulfur content and slag basicity on slag viscosity. In
Figure 7B is shown the effect of slag sulfur content on slag viscosity
for 0.2 basicity slags. As seen, the slag viscosity increases with
sulfur content. It is anticipated that slags with viscosities in the
order of 50 poise or less will be sufficiently fluid for combustor
operation; consequently, slag sulfur levels in the range of 6-8 percent
are usable.
The effect of temperature and slag basicity on slag viscosity is shown
in Figure 8B for slags containing 7.8 and 8.8 percent sulfur. As ex-
pected, slag viscosity decreases as the temperature increases. The
viscosity also is seen to decrease as slag basicity decreases. For
example, the slag viscosity is about 10 and 30 poise for basicities
of 0.1 and 0.2 respectively at a temperature of 2750°F. Consequently,
slags to be used in the combustor should have basicities in the range
of 0.1 to 0.2.
It should be pointed out that the considerable scatter of the viscosity
data obtained from the high sulfur bearing slag is attributed to the
fact that at these sulfur contents, not all of the sulfur goes into the
solution. The slags contain solid CaS particles that may be non-
uniformly distributed within the slag mixture contained in the viscosi-
meter cup. Experimental result variability was artificially introduced
into the measurement by swirling or mixing the slag prior to making a
viscosity measurement. By swirling the slag, the solid CaS particles
are forced to the outer perimeter of the cup and away from the oscil-
lating bob. Consequently, the oscillating bob then measures the vis-
cosity of slags containing little or no suspended solid particles.
Because solid particles tend to yield higher apparent viscosities when
measured with an oscillating bob viscosimeter, the stirred slag yielded
lower viscosity values. Accordingly, it is expected that the apparent
viscosities of high sulfur bearing slags measured in this equipment are
probably substantially higher than the actual viscosities themselves.
Visual observation of the molten slags as they are removed and poured
from the viscosimeter indicate that the true viscosity will be considerably
lower than the experimental values. These judgments are based on a
relative comparison of the slag flow characteristics with those of the
standard viscosity oils.
The above viscosity results were obtained for laboratory prepared synthetic
slags. These slags did not contain FeO which will be present in actual
combustor slags. The presence of even a small quantity of FeO has the
effect of substnatially reducing slag viscosity-1-^. Consequently, it is
expected that in actual operation the slag viscosities will be much lower
than the experimental results indicated.
-40-
-------
o
o
o
o
o
I
-p-
CO
•<-(
o
>» o
4J O
1-1 f-t
to
O
o
o>
1-t
> o
60
03
CO
d
e> •
M
BJ O
O. i-l
O.
Slag Temperature
2835°F
2710 to 2780°F
2650°F
4 8 12
Slag Sulfur Content, Percent
FIGURE 7B-EFFECT OF SULFUR CONTENT ON
APPARENT VISCOSITY OF 0.2 BASICITY SLAG
O
O
O
O
O
O O
P^ O
CQ
O O
O u"»
CO
Q
f-<
CO
Slag Temperature
2730 to 2750°F
2850°F
O 2900°F
Slag Sulfur Content: 7.8 to 8.87.
0.2 0.4 0.6 0.8
Slag Basicity, Weight Ratio CaO/SiO, +
FIGURE SB-EFFECT CF BASICITY ON APPARENT VISCOSITY
OF HIGH SULFUR BEARING SLAGS
-------
Based on the viscosity results of this section it can be concluded that
low basicity (0.2 or less), high sulfur slags (6-8%) will be sufficiently
fluid to readily pour from the combustor and at the same time exhibit
high enough viscosities to minimize refractory wear. For this reason
the bulk of experimental combustor work was conducted using slags at
0.2 basicity. The addition of coal to the combustor would add sufficient
ash to the yield slags in the range of 0.1 to 0.2 basicity.
SULFUR PARTITION BETWEEN IRON AND SLAG
Introduction
In the two-stage coal combustion process, sulfur is introduced into the
iron bath of the combustor that is injected beneath the molten surface.
From an economic point of view, it is desirable that all of the coal sulfur
be recovered from the iron bath as calcium sulfide in the molten slag.
The distribution of sulfur between the liquid slag and the molten iron
at steady state is of primary importance. This distribution (termed the
partition ratio) is defined as the ratio of the percentage of sulfur found
in the slag to the percentage of sulfur in the iron.
For optimum operating conditions, it is desirable that the partition
ratio be as high as possible. In this manner, the bulk of the sulfur
will be present in the slag rather than in the iron phase. If a high
partition ratio (10 or more) can be obtained with relatively low excess
lime, the lime consumption will be low and operating costs will be reduced.
Furthermore, as the partition ratio increases, the sulfur content of the
iron decreases which results in a higher value for the iron produced in
the process.
Based on the literature available for partition ratios of blast furnace
and steelmaking slags, high partition ratios are associated with increased
temperature, a reducing atmosphere and high slag basicity. However, these
data are generally confined to slags containing less than 2 percent sulfur.
Unlike steel manufacturing processes, the combustor will operate with slags
containing large amounts of sulfur (about 8 percent). The only partition
ratio data available for such high sulfur slags were developed for high
basicity magnesia-bearing slags .
Accordingly, a laboratory investigation was initiated to determine the
partition ratios and sulfur content for both the iron and the slag when
a low basicity high sulfur bearing slag (calcium oxide-silica-alumina
system) was in contact with molten iron. Although data are not available
to determine the effect of free calcium sulfide on partition ratio, no
attempt was made to measure the maximum solubility of calcium sulfide in
these slags. Rather, a study was made to evaluate the overall effect
of high-sulfur slags (regardless of solution state) on equilibrium sulfur
partition ratios.
-42-
-------
In this series of experiments, the equilibrium partition ratio was
obtained by transferring sulfur from the slag to iron, rather than
from the iron to the slag as expected to occur in the process. This
was done to save cost and time, since this approach was proved valid
by the data obtained from a previously completed study^ which showed
that equilibrium was reached by either method.
Experiments were divided into two groups. The first group involved the
determination of the minimum time required to reach partition equilibrium.
The second group was concerned mainly with the effect of temperature and
slag basicity on sulfur partition ratios.
Slag Preparation
In this investigation, synthetic slag samples were prepared by mixing
master batches of CaO, Si02, Al-0 , and CaS to obtain slags of varying
basicity as shown in Table VIB.
TABLE VIB
Slag Composition for Partition Ratio Studies
Basicity
0.2
0.5
1.0
These mixtures were then placed in graphite crucibles and heated to
2600°F - 2760°F for at least eight hours. Each crucible was removed
from the furnace, stirred and returned to the furnace. The stirring was
repeated at least twice to provide a uniform slag composition. The
slag was then removed from the furnace, cooled and crushed to minus 10
mesh. U.S. These synthetic slags were used in all subsequent experiments.
Effect of Time and Temperature on Approach to Equilibrium
A Burrell Model H-2-9 high-temperature electric furnace was used for
this work. The furnace, which contains two heating chambers (1-1/2 inch
ID combustion tubes), was placed on its side so that the chambers were in
a vertical position. The bottom openings of both heating chambers were
sealed except for a gas inlet.
A silicon carbide rod was placed in an upright position in the bottom of
each chamber and served as a base for small graphite crucibles containing
crushed iron covered with a granulated slag. At the start of the experiment,
the heating chambers were heated to test temperature and flushed with
CaO
13.4
26.0
40.0
blU0
Z
44.4
36.0
26.6
A100
2.- -3
22.2
18.0
13.3
CaS
20.0
20.0
20.0
-43-
-------
nitrogen through the gas inlet at the bottom of the chamber for at
least 10 minutes. Two small crucibles, one on top of the other, were
placed in one of the heating chambers. The crucibles (3/4 in. ID x 2
in. high) contained 6 grams of iron and 1.5 grams of slag. A larger
crucible (3/4 in. I.D. x 3 in. high) containing 12 grams of iron and 3
grams of slag was placed in the other chamber. Aluminum foil was instal-
led inside each heating chamber 1-2 inches down from the top to serve as
a radiant heat reflector. The nitrogen flow was stopped when the rubber
stoppers of both heating chambers were tightened. The heating intervals
were measured from the time the furnace temperature reached 2600°F until
the samples were removed. The heating times varied from 30 minutes to
24 hours.
The effect of time at temperature on the approach to equilibrium partition
ratios is presented in Figure 9B. In these experiments, furnace
temperature was maintained constant at 2600°F (sample temperature
estimated at 2450°F). Initially, when all of the sulfur is in the slag,
the partition ratio is infinite. However, as the slag and molten metal
are held in contact, the partition ratio rapidly decreases and slowly
approaches an equilibrium value after 10 hours. The data scatter is
typical of what one may expect from measurements of this sort. However,
it is interesting to note that the data are independent of the type of
crucible used as well as the position in the furnace. Based on these
results, it was decided that in subsequent work for determining the
effect of basicity ratio.and sulfur content on expected partition ratios,
the contact time should be maintained at 10 hours.
Effect of Slag Temperature and Basicity on Equilibrium Partition Ratio
To expedite the work, the second group of experiments were completed in
the laboratory-constructed viscosimeter furnace. This furnace has a
cavity of 8 inches wide by 9 inches long by 8 inches high. A refractory
crucible holder with 5/8 inch diameter 1 inch deep holes to accommodate
six graphite crucibles and a thermocouple was placed in the furnace and
was heated to a predetermined temperature with an argon purge for at least
1/2 hour prior to installing test samples. The test samples, consisting
of approximately 0.9 grams of slag and 3.6 grams of iron granules (containing
4 percent carbon),. were contained in graphite crucibles approximately 1/2
inch OD and 3/8 inch ID by 2 inches long. The slags in this work were the
same as those listed in Table VIB.
After the crucibles were placed in the crucible holder, charcoal or
graphite powder was sprinkled over the entire assembly and a stream
of argon was maintained in the furnace to prevent disintegration of
the graphite crucibles by air infiltration. The samples were maintained
at a specified temperature for ten hours to insure equilibration. The
temperature was measured at the center of the crucible holder. The entire
contents in each crucible were crushed and separated into magnetic (iron)
-44-
-------
o
•1-1
AJ
(-1
CO
401
30 1
:.-,"*4-
':•.•'. Legend:
'•... O.- ':..-Large Crucible ' ;'/ '•
. .- @ ' ; Small Crucible,' -Top • •.' '..
.. Q .';• Small Crucible,' '.-Bottom'
0.3 • 0.5 :. . .1.0 ; . ••_,,;...; s'.o'.-.;;.'10.0.
. ' '" . . '. Heating Time, Hours .. '•';'• ':'. . ;-.
FIGURE' 9B-EFFECT OF HEATING TIME ON PARTITION RATIO
,' .30.0
-45-
-------
and non-magnetic (slag) portions and the sulfur content of each portion
was determined.
Figure 10B presents the effect of slag temperature and slag basicity on
the equilibrium partition ratio that is achieved. The data show that
the partition ratio for a given slag basicity rapidly increases with
increasing temperature. It- should be mentioned that, unlike the temperatures
in Figure 9B which were furnace temperatures, the data of Figure 10B are
based on'actual slag temperatures. Because the combustor economics are
vastly improved when using slags of 0.2 basicity or less, the partition
ratio data are encouraging in that high values can be realized even though
low basicity slags are employed. The data show that, for a 0.2 basicity
slag, partition ratios of the order of 25 to 170 may be realized over the
temperature range of 25-2700°F. The latter temperatures span the operating
temperature that will be achieved in the combustor operation.
HEAT CAPACITY
Introduction
Heat capacity data for the various slags at high temperatures are required
to adequately define the heat balances for the process. This section
presents the results of work done on three different types of slags having
basicities of 0.2, 0.5 and 1.0 (Table VIB). These slags were selected as
typical of those that may be used in the operation of the combustor.
Procedure
Standard calorimetry equipment and procedures were used to determine the
specific heat o.f solids. The calorimeter was a one-liter capacity Dewar
vacuum flask. To preclude the possibility of slag reaction with water,
ethylene glycol was used as the heat absorption liquid. Specific heat
data on the laboratory C.P. grade of ethylene glycol used were obtained
from the literature-^. A Beckman differential thermometer graduated to
0.01°C and capable of interpolation to 0.005°C was used to measure the
temperature rise of glycol. A standard laboratory thermometer accurate
to within 0.5°C was used to measure the absolute end point temperature
of the glycol and cooled solids.
Before beginning experiments, the heat absorption values for the Dewar
flask calorimeter, Beckman thermometer and stirrer were determined. A
cork stopper was used to cover the Dewar. ' The Beckman thermometer and
stirrer were introduced into the flask through holes punched in the cork.
The calibration of the Dewar flash for heat absorption and losses was
done by in-situ neutralization of sulfuric acid with sodium hydroxide.
The heat input or theoretical heat of reaction due to neutralization was
obtained from the literature . The difference in literature and experi-
mental values was the heat absorbed by the calorimeter.
-46-
-------
'AGO
o
•H .•
g .300
•H. ' • '
4-> ' ••.
•r-l . ' • • '
2001
• . 1001
••:•• o
.. Sample Temperature'
O 2490°F ...-...• .-. ;••' ':• .
O 25.70°?; .'•.;*..".;-.:'': i"•'..'•••'.;'•'•:.
A" 2680°'?'.'••:':;.-::.'.'•-••'.:i.'
., 0.2 ::.
-•-•'• :.-'••.•••:.' .-'.'•'i.o
,• "• :.. '•:'- • 0.5' ' .;•' -.'•';;..:••_ ......
• • '.'.. ';'• •.. ... ..'•• ;' '.•• Slag Basicity . '. • ' .'': . ;''.'."•'"•
FIGURE 10B- EFFECT OF SLAG BASICITY ON PARTITION' RATIO ' '
-47-
-------
Once the calorimeter constant was evaluated, the method was standarized
by determining the specific heat of aluminum and nickel over the-temperature
range 300 to 1200°C. A comparison of the experimental and literature
values for the specific heats of aluminum and nickel permitted an estimate
to be made of the heat losses experienced during the transfer of the hot
aluminum or nickel sample from the furnace to the calorimeter. The dif-
ference in the heat content measured and that calculated from the liter-
ature values for the specific heat of both materials was attributed to
radiant heat losses as well as glycol evaporization. Correction factors
were developed which related the difference in measured and literature
heat content values through the emissivity and surface area of the sample.
These corrections are shown in Figure 11B as a function of initial sample
temperature.
For determining the specific heats of the various slags, the following
procedure was used. Slag samples were briquetted into a cylindrical
disc of 1 inch diameter and 1/8 inch thick. The slag disc was then in-
serted into a ceramic combustion thimble and both were heated to temp-
earature in a muffle furnace. The weight of the thimble varied between
60 and 70 grams and the ratio of thimble to slag sample weight varied
between 4 and 8. Relatively massive thimbles were used to minimize
heat losses from the slag disc during transfer from the heatup furnace
to the calorimeter.
The calorimeter containing a known weight of glycol (usually 500 grams),
was allowed to equilibrate with the surroundings. The Beckman differential
thermometer was adjusted to read near the bottom of the scale so that
a maximum temperature increase (5.2°C) in the glycol could be measured.
In the event that the temperature rise of the glycol exceeded the range
of the thermometer, the test was discarded. Once the differential
thermometer had been standardized, the combustion thimble containing the
slag briquette was removed from the heatup furnace. The temperature of
the combustion thimble and slag sample were continuously monitored by
means of a thermocouple inserted within the thimble. Upon removing the
thimble and briquette from the furnace, the briquette was immediately
transferred to the calorimeter by simply dumping the contents of the
thimble. During the dumping of the slag briquette, glycol fluid agitation
was maintained by means of a glass stirrer. The temperature of the glycol
was measured by means of the differential thermometer as a function of
time. Measurements were continued until a constant rate of cooling of a
glycol was observed. In this manner, heat capacity data could be corrected
to account for heat losses from the system. It should be mentioned that
these corrections were relatively small. The effect of calorimeter heat
losses on the measured specific heat was less than 0.1 percent.
Results
The effects of slag basicity and temperature on the heat content (above
28°C) of various slags are presented in Figure 12B. As can be seen,
slag basicity has little or no effect on the enthalpy of the solids.
-48-
-------
. 15. _
D
y,
w
V io.
a-
e
CO
H .. .
00
>,
0-)
• r-l
W . '
•r-l
£
W
.:•'© .'Aluminum
Q Nickel;V:
%-#&£iE^}Q
• •'= •-..-'• .2DO
•r; :;,;.,.-.4t)0 .. ;' 6'00 • •/V •; • 800 "v
• '. ' ;•'.:.. ".•'. ••'••••. ."•'•'.' * .Temperature .°C .'..-•'. ' " .
FIGURE.ilB-CORRECTION FACTOR FOR HEAT LOSSES DURING SAMPLE
• ••..-••••• /:•: ;., /'•/•V.r'- ';.:.':-TO THE CALORIMETER"'--' •'.'•' '"-'.•-•• •-
1000
.TRANSFER-
-49-
-------
Within the limits of experimental data, all of the slags exhibit similar
characteristics. The data of Figure 12B were correlated by means of
a least square polynomial fit. For comparison purposes, the correlating
curve is shown along with a heat content curve for open hearth and blast
furnace slags . As is evident, the high sulfur bearing slags used in
this work exhibit characteristics not too far removed from those of
commercial blast furnace and open hearth slags.
To obtain the heat capacity of the various slags, the polynomial ex-
pression correlating the enthalpy with temperature was differentiated
with respect to temperature. The derivative, which is defined as the
heat capacity, is presented as a function of temperature in Figure 13B.
A comparison of the experimental values for the combustor slags with
those of open hearths and blast furnace slags indicates that the m'aterials
are somewhat comparable in nature. The high sulfur slags tend to exhibit
a somewhat higher heat capacity.
The Effect of Temperature on the Relative Heat
Content of Various Slags .
300
6
200
O 1.0 Basicity Slag
D 0.5 Basicity Slag
O 0.2 Basicity Slag
Polynomial
Fit. • •
100
W
Literature
Reference 17
I
1000
• , •:' ; 200 400 6(!)0 . ;•-:;• 800,.../.
. . •• "•..••'".'".'. ' Temperature °G'.
• - '.•'•.'*'-""
. Figure 12B . ..
The equation correlating the heat capacity as a function to temperature
is as follows:
Cp =0.119 + 3.66xlO~4T - 15.90xlO~8T2
where Cp is in calories/gram/degree C. It is valid for temperatures
between 200°C and 1200°C and slag basicities between 0.2 and 1.0.
-50-
-------
The Variation .of Slag Heat. Capacity with Temperature
o
o
0.3.
u
M
U
o
w
K
Literature Reference. \i
200
400 600 800
'Temperature °C
Figure 13B";.
1000
-51-
-------
Total Specific Surface Area
One of the measurements of the physical properties of high-sulfur bearing
slags required as a possible correlating variable for the desulfurization
and granulation studies on such slags is their total surface area.
Tests were conducted on three slags of 0.2, 0.5, and 1.0 basicity ratio
(basicity defined as percent of calcium oxide divided by percent silica
plus percent alumina). The composition of these slags is shown in
Table VI-B. A Perkin-Slmer 212D Model Sorptometer was employed for
these measurements.
Procedure
The principle of the surface area measurement is based on the amount
of gas absorbed by a solid sample. In this method, a known mixture of
nitrogen and helium is passed over the sample (1.0-2.0 grams) in
a sample tube and the effluent is monitored by a thermal conductivity
detector. While the gas is flowing through the tube containing the
sample, the tube is cooled by immersion in a bath of liquid nitrogen.
As a result, the cooled sample adsorbs a certain amount of nitrogen
from the gas stream, which is indicated on a recorder chart as a peak.
The area of this peak is proportional to the volume of nitrogen adsorbed
by the sample. _After equilibrium is established, the sample tube is
removed from the liquid nitrogen bath. As the sample warms the adsorbed
gas is released and enriches the effluent: gas passing through the sample
tube. A desorption peak is then obtained, which is in the reverse
direction of the adsorption peak. When desorption is complete, a known
volume of nitrogen is added to the nitrogen-helium stream to produce a
calibration peak. By comparing the areas under the desorption and
calibration peaks, the volume of the nitrogen adsorbed by the sample
can be calculated.
Discussion of Results
Prior to measuring the total surface area of the slags, the sorptometer
was standardized by measuring the surface area of a calibration sample.
A standard sample of Titanium dioxide (TiO«) supplied by the American
Instrument Company was used for this work. The known total surface
area was 10.3 + 0.2 square meters per gram. Excellent agreement was
found between the standard sample surface area and that determined by
means of- a sorptometer. The experimental value was 10.67 square meters
per gram.
The results of surface area determination on the three different
basicity slags are shown in Table VII-B.
TABLE VII-B
Total Surface Area of High Sulfur Bearing Slags
Basicity Surface Area SQ. M/gram
0.2 , 0.3
0.5 less than 0.1
1.0 less than 0.1
-52-
-------
As is evident from this table, only the 0.2 basicity slag exhibited a
sufficiently high surface area for meaningful measurements. The outer
two slags had surface areas too low for measurement by the sorptometer,
which can measure surface areas down to 0.1 square meters per gram.
The surface area shown for the 0.2 basicity slag represents an average
value for five different measurements. The average specific surface
was 0.31 square meters per gram and the range of the experimental data
was. from a low of 0.20 to 0.43 square meters per gram. Variability in
the measured surface areas was attributed to variations in slag com-
position. A number of tests were conducted on the other two slags;
however, surface areas were outside the range of the instruments.
Because the slags are glass-like in nature, it can be expected that the
total surface areas will be minimal. In particular, because the higher
basicity slags tend to be more glassy, total surface areas decrease with
increasing basicity.
EXTERNAL SURFACE AREA
Introduction
As part of the detailed slag characterization, the external specific
surface area of slags over the basicity range of 0.2 to 1.02 was
determined. The external specific surface is a useful correlating
variable for subsequent work, such as slag desulfurization reaction
kinetics and slag surface expansion by granulation and is necessary to
determine the grinding energy requirements for the various slags. This
section presents the results obtained from the experimental investigation
of the external surface areas of high sulfur bearing slags.
Theory
The relationship of the specific surface area of the solids to the
pressure drop obtained in laminar flow of ..fluids through packed beds
of granular materials had been formulated . The theoretical expression
relates the volumetric flow of fluids to the pressure drop established
per unit length of bed, the specific surface of the solids, bed
porosity, the viscosity of the fluid, and the cross-sectional area of
the packed column. Since all variables with the exception of the
specific surface are readily measured experimentally, the specific
surface can be calculated from the following equation:
(Pv (1 - E)')
where d = the apparent density of the solids, grams/cubic centimeter
E = porosity, dimensionless
v = kinematic viscosity of fluid, square centimeters/second
S = specific surface, square centimeters/gram .
P = permeability of packed bed, centimeters/ (second)
-53-
-------
If laminar flow of fluid is maintained throughout the packed bed, the
above equation adequately describes the relationship between the measured
variables and the specific surface area. Data reproducibility can. be
maintained within plus or minus 10 percent.
Equipment and Experimental Procedure
The apparatus used to measure the specific surface consisted of a flow
meter, an inclined manometer, and a volumetrically graduated sample
holder of known cross-sectional area. To determine specific surface
of the slags, it is-required that the apparent density of the slag
particles be known. An apparent density measurement, rather than a
true solid density determination, is required because the latter value
is independent of the porosity, or pore volume, of the particle. Since
fluid flow does not occur within the particle, it is necessary that the
apparent density be used to evaluate the open porosity of the bed.
Apparent densities were determined by total immersion of the slag
particles, and the density and the volumetric displacement of the carbon
tetrachloride by the slag particles, the apparent density was readily
calculated.
The experimental procedure for determining the specific surface was as
follows. A weight of solids of known apparent density was poured into
a graduated glass sample tube holder and vibrated to maximize the packing
of the solids. Knowing the length of the packed bed and the cross-
Sectional area of the sample tube holder, the sample volume was then
calculated. Bed porosity was determined from the weight, density, and
the volume of the packed bed. Once the bed parameters were measured,
a known volume of air was caused to flow downward through the bed. The
pressure drop across the packed bed was measured when steady flow
conditions had been achieved. In general, pressure drop measurements
were obtained for at least five different air flow rates.
For the equipment used in this work, an air flow rate of five cc per
second or less insured a laminar flow condition. For laminar flow, a
plot of pressure drop versus the volumetric flow rate yields a straight
line passing through the origin. The slope of the straight line represents
the permeability of the bed. Using this permeability value along with
the viscosity of the air, the porosity and apparent density of the
solids, the specific surface was calculated by the equation presented
earlier.
Discussion of Results
Earlier work established the validity of the specific surface area
measurement. In that work, glass beads having nominal diameters of 3,
A, and 6 milimeters were employed. The specific surface of the glass
beads was determined by micrometer measurements of the diameter and by
obtaining a particle count per unit weight for the different size glass
-54--
-------
beads. In this manner, the specific surface in terms of square centimeters
surface area per gram of sample was determined. A comparison of the
surface areas obtained by the micrometer measurement and air permeability
methods is presented in Table VIII-B. Agreement between the two methods
was excellent.
TABLE VIII-B
Comparison of Glass Bead Surface Area Measured by Micrometer
and By Air Permeability Method
Glass Beads Specific Surface, sq cm/gram
Diameter, mm Air Permeability Micrometer
3.15 7.4, 7.7, 7.9 7.6
4.06 5.5, 5.8, 6.1 5.9
5.95 3.9, 4.1, 4.4 4.1
The effect of basicity and particle size on the external specific surface
area of slags shown in Table VI-B is presented in Figure 14-B. The
data show that the specific surface rapidly increases with decreasing
particle size. Scatter in the data is attributed to variation in the
sampling and composition uniformity of the slag. Because calcium
sulfide has a limited solubility in lime-silica-alumina slags, it is
probable that the fracture characteristics of the particles are highly
dependent upon localized concentration of segregated slag components.
No significant differences could be observed between slags of 0.2, 0.5,
and 1.0 basicity. For comparison purposes, the smoothed curve obtained
for magnesia-lime-silica-alumina slags containing about 18 percent
calcium sulfide is also shown in Figure 14-B. These data were obtained
for completed work for the Water Quality Office . Evidently, since the
slags are glass-like in nature, the fracture characteristics tend to
generate similar external surface areas which are relatively independent
of compositional changes. The crushing energy requirements also tend
to substantiate this conclusion.
CRUSHING ENERGY
Introduction
In the Two-Stage Coal Combustion Process, high sulfur bearing slags are
periodically withdrawn from the combustor and processed to recover
elemental sulfur. The molten slag may be granulated by means of water
quenching and/or vacuum treatment to produce a particulate material.
Alternatively, depending upon process designs parameters, the slag may
be cooled in molds and subsequently ground to produce a size distribution
that would optimize process economics. Slag particle size influences
the desulfurization equipment size and the commercial utility of the
coarse road aggregate produced. Regardless of the process employed to
-55-
-------
'••': looo
E
CO '
M
60.
e •
O'
•
Cf
w
O
C3
o
-rl
O
' G
Q.
CD
C
X
w
-100
Symbol .-:.•/ •. • Slag Basicity
S'^^-'^^^^
'••• 'T\.•-'•:.'.'•-.'•:;'.v'v>''. ' .'.'•.•.'.•'..*.':-'>
•"..'o ••:::.••
::-.10 L
•".-.'• . Reference '2
;.•:.. 0.04 •....:•; o.i. ; , _ -. .. .' ;. i.o .-••••>/:.•/., ...
/'. •. . '•-. ...'• ••'"' .' '•'''.Particle Diameter, millimeters; .•• ' .'•" •'.•' • .. •'.'•'.
•FIGURE-IAHB-EFFECT OF SLAG BASICITY AND PARTICLE. SIZE ON. EXTERNAL'SPECIFIC
;•,.••,••'.•;.•'••..;•'..,•.••.••.•••..';;•.:/;•. SURFACE..'.. •'•••.-. \-,.-.-.. ::-: •.•'•'..••:•.'••'•''. .•
-56-
-------
produce a particulate slag material, the possibility exists that crushing
and grinding equipment will be required. Because both the capital and
operating costs for crushing equipment depend on the crushing energy
required to produce the particle size of the slag, a study of crushing
energy requirements v?as undertaken.
Experimental Equipment and Procedure 1q
A modified version of the drop-weight machine used by Gross was used
in this work. A schematic diagram of the equipment is presented in
Figure 15-B. A stainless steel cylindrical split die and stainless
steel plunger comprised the assembly that housed the slag crushing
chamber. The crushing chamber (6.1 cm diameter) consisted of a cavity
formed by the volume of the slag contained between the bottom of the
plunger and the upper surface of the die base. A 2.62 kilogram drop-
weight was used to generate the energy required to crush the slag. The
walls between the plunger and die were lubricated with graphite to
minimize plunger friction.
A known weight (about 15 grams) of closely sized (minus 10, plus 20
U.S. mesh) slag was introduced into the die. The external surface
area of a given slag sample was previously measured. The plunger is
carefully inserted in the die and slowly lowered to rest on top of the
bed of solids. The assembled die containing the slag sample rests on a
circular piece of aluminum wire (0.064 diameter) and is centered directly
beneath the drop weight. The aluminum wire had been previously calibrated
to determine energy absorption as a function of wire diameter after
deformation. The drop weight is raised by means of an overhead pulley
to a known height above the plunger resting on the test solids. Once
the elevated drop weight stops oscillating, the string is cut to permit
the hemispherical drop-weight to impact on the center of the plunger.
After impact, the aluminum wire is removed from beneath the die and
measured in seven locations to determine wire deformation. If the
deformation was not uniform along the entire length of wire, the test
was discarded. Uneven deformation is the result of an off-center hit
.with the falling drop-weight which gives rise to binding friction between
the die and the plunger. Consequently, the energy absorbed by the
crushed solids is not known. When a uniform deformation of the wire is
observed, the energy absorbed by the impacted solids is readily calculated
as the difference of the energy input of the drop-weight minus the energy
absorbed by the wire..
The crushed solids are then removed from the die and weighed to determine
weight loss, (usually less than 0.5 percent). The external specific
surface area of the crushed solids is then measured to determine the
increase in surface area after crushing. Knowing the new surface area
generated and the energy absorbed by the solids, the crushing energy,
or Rittinger's Number, is calculated.
-57-
-------
Drop Weight- .
Plunger
Base
Die
.Slag Sample
-Aluminum Wire
' 'FIGURE 15-B- CRUSHING ENERGY APPARATUS
-58-
-------
Previously developed calibration curves which related the deformed
aluminum wire diameter to the energy (kilogram centimeters) absorption
were used in this work. The validity of the calibration curves were
re-established at several different points. In these calibration
experiments, no solids were introduced into the die. To obtain deformation
data, the drop-weight was elevated to varying heights above the plunger
and released. Prior experience has shown that alignment difficulties
between the plunger and the drop-weight occur at high energy inputs and
data reproducibility becomes poor. Accordingly, in these calibration
tests and all subsequent experimentation, the drop-weight was not raised
to an elevation greater than 15 centimeters. The results of the
calibration tests are presented in Figure 16-B. As is evident, the
data confirm the validity of the calibration curves.
Discussion of Results
The effect of slag basicity (over the range of 0.2 to 1.0) on the
crushing energy requirements for slag containing 20 percent calcium
sulfide are presented in Table IX-B. Chemical composition of the slags
used in this work is shown in Table V-B. The data suggest that an
optimum Rittinger's number is achieved at a slag basicity of 0.5.
However, the extreme variability in the data limit the practical
significance for the commercial application of the results. The
difference between the results of the various basicity slags are not
great enough to warrant a selection other than one common type of
crushing and grinding equipment.
TABLE IX-B
Variation of Crushing Energy with Basicity
o
Basicity cm /Kg-Cm Crushing Energy
0.2
0.2
avg
4.56
2.12
1.25
1.72
2.73
2.47
0.5 4.23
2.96
4.40
avg 3.86
1.0 1.97
1.49
0.92
3.39
avg 1.94
-59-
-------
'•:.: '.>;" '"-: .0640
•.:--. 0630
.§;..:. 0620
CD • -;/ '. • .
°:;:V.0610
0) •••••.'
40 ;• 50 .. :'\.,;.'',- .'
•.• ".'•"'• •'•';'..' Energy Input Kilograms-Centimeter ..; .' ;. .•'•;'
.FIGURE. 16B-CALIBRATION CURVE FOR ALUMINUM WIRE USED IN
::- ';:•;> .., ;•;•:';•.,".. CRUSHING ENERGY, TEST •;.;.;. ''..-;.' " .".',•' ./;
-60-.
-------
The variability in the crushing energy results is attributed to the
inclusion of undissolved crystalized calcium sulfide that is interspersed
within the slag matrix. Calcium sulfide has a limited solubility
in lime-silica-alumina slag system . Depending upon the composition
of the slag system, sulfur (in the form of calcium sulfide) solubility
varies from 2 to 6 percent. In this work, the sulfur content of the
various slags was about 9 percent. The excess of CaS in the slags
creates a suspension of undissolved particles bound by a slag matrix.
Since the cooling rate of the slag was uncontrolled during preparation,
a variety of crystal sizes can be generated which give rise to a large
variability in the surface areas generated during crushing.
Rittinger's Number was determined for silica sand to obtain a reference
point for comparison of slag on a relative grindability scale to a
common material for which commercial crushing equipment is available.
A comparison of the slag to silica sand is presented in Table X-B and
shows that the crushing energy requirement for these slags compared
favorably with that of silica sand. Consequently, the selection of
commercial equipment can be based on silica sand characteristics.
TABLE X-B
Comparison of Crushing Energy Requirements for Silica and Slags
Material Average' Crushing Energy sq cm/Kg-cm
Silica 3.33
0.2 Basicity Slag 2.48
0.5 Basicity Slag 3.86
1.0 Basicity Slag 1.94
19
It is of interest to point out that Gross indicates a Rittinger's
Number of 17.56 for silica as compared to 3.33 reported in Table X-B.
The discrepancy is the result of the method by which surface area was
determined. Gross used a rate-of-solution method rather than an air
permeability technique to determine surface area. The rate-of-solution
technique yields higher surface areas because of solvent penetration
into the interstices of the particle. Air permeability measurements
afford little opportunity for any significant penetration within the
particle. Consequently, surface areas determined by the latter technique
would tend to be low by comparison.
21
Rittinger's Numbers determined from measurements on commercial units
agree reasonably well with those obtained in this work. Literature
values for the grinding of quartz in various sizes of ball mills range
from 2.6 to 6.8 sq cm/Kg-cm as compared to the 3.33 value.obtained.
Based on the results of the crushing energy studies, the high sulfur
bearing slags that will be produced in the combustor can be assumed
to have grinding characteristics comparable to silica. This criterion
should facilitate the selection of commercial equipment for 'use in
the process.
-61-
-------
GRANULATION STUDY
Introduction
As an alternative to slag crushing prior to desulfurization, a vacuum
tower slag granulation technique was investigated. This granulation
technique was studied because it is desirable to avoid the use of water
in granulation because hydrogen sulfide formation is prevented and
desulfurization heat requirements are lowered if the molten slag can be
granulated without being cooled by contact with water
The proposed experimental technique involves the introduction of slag
through a crucible into a vacuum shot tower which causes dissolved gases
to leave and expand the slag; thereby, increasing the porosity and total
surface area of the slag granules.
The purpose of .this study was to determine if expanded slag particles
could be produced by this vacuum tower technique and if so, to relate
the orifice size in the crucible, the vacuum in the tower, and the slag
temperature to the size, shape, and surface area of the slag particles
formed.
Experimental Equipment
The concept behind increasing the surface area of slag granulated by the
vacuum tower technique is simple. The objective is to form a small drop
of molten slag at the bottom of an orifice and to expose this drop to a
vacuum for as high a residence time as practical to enhance the opportunity
for dissolved gases to escape from the slag droplet and create pores
in the solidifying slag particle. Orifice size, slag flow rate, and
temperature will affect the residence time requirement of the slag
droplet at the orifice.
The experimental equipment is shown in Figure 17-B and consists of a
3 inch I.D. 15-foot steel pipe located on a sample collection box. At
the column top is a flange on which concentric graphite crucibles were
mounted. The crucibles had a hole (s) of known diameter and could be
aligned to permit molten slag to flow through. Fittings on the collection
box permitted the attachment of two vacuum pumps; a vacuum gauge was
connected to the column. The slag was melted in the crucible using a car-
bon rod attached to an electric arc welding machine. Provision was
made for an argon purge into the crucible to prevent the oxidation slag
sulfur.
Prior to actual experimentation, a simulation study of drop formation
was made using room temperature liquids which -simulate molten slag.
The experimental equipment is shown in Figure 18-B and consists of a
vacuum flask connected to a vacuum pump. A fine orifice was fitted into
a rubber plug and connected to a buret with a stopcock. In this manner,
drop, bead, or stringer formation from an orifice of known diameter into
a vacuum could be studied at room temperature.
Experimental Procedure
The procedure to study the formation of droplets in the flask involved
setting the vacuum at the desired level (50 mm Hg) and introducing liquid
-62-
-------
to Electric
Arc Welding]
Machine
Carbon Rod
._- Argon Purge Line
A
Vacuum
Tower ~
Inner Crucible
— Outer Crucible
< Vacuum Gauge
Sample Collection Box
FIGURE 17-B -VACUUM TOWER GRANULATION APPARATUS
-63-
-------
-Burette
•Stopcock
to Manometer
to Vacuum Pump
Capillary Tip
Vacuum Flask
FIGURE 18-B - DROP FORMATION APPARATUS
-64-
-------
from the buret through the stopcock. The formation of droplets or stringers
could then be observed.
The procedure to operate the vacuum shot tower equipment was as follows.
The crucibles, with holes misaligned, were cemented to the top of-the
tower. A sample collection pan, either dry or containing water, was
Placed in the sample collection box and the vacuum pumps started. When
a .vacuum of 28 inches of mercury was obtained, the inner crucible was
charged with approximately 200 grams of slag and the carbon rod inserted
in the slag. The electric arc was then started and the slag melted for
30 to 45 minutes with argon being purged into the inner crucible. When
the slag was fluid, the holes in the crucibles were aligned and the slag
was drawn into the tower by the vacuum. The granulated slag was collected
in the sample collection pan. Several times during a run the temperature
of the molten slag in the crucible was determined.
The total surface area of the minus 20 mesh fraction of the granulated
slag was determined using a Perkin-Elmer Sorptometer; the external
specific surface of the minus 10 plug 20 mesh fraction was determined
using an air permeability technique . The density of the minus 10 plus
20 fraction was measured by a liquid displacement technique. Particle
size distribution for the granulated slag particles were determined.
Discussion
A qualitative comparison of the various shapes of granulated slag particles
and their associated densities and surface areas is helpful in the
evaluation of the granulation study.
(1) A solid sphere or bead should have the same apparent density as
ungranulated slag. If it is assumed to be solid and smooth, its total
surface area should be relatively low compared to crushed slag, which
possess a certain amount of porosity and fissures from the crushing
process.
(2) Long stringers would have the same characteristics as solid
spheres. Both of these configurations are undesirable as compared to
granulated slag particles because of their low surface areas.
(3) If conditions of residence time at the orifice, slag temperature
and vacuum level in the tower are such that regular droplets (spheres)
are not formed but irregularly shaped particles with rough surfaces,
these latter particles should show an increase in specific surface area
but no change in apparent density. If pore formation and honeycombing
are negligible, the total surface area will be low.
(4) Spheres with open channels will have a different apparent density
from ungranulated slag. Specific and total surface areas will probably
be increased over ungranulated slag particles.
(5) Hollow spheres will have a decreased apparent density compared to
ungranulated slag. The specific surface area should not change, while
the total surface area may or may not change.
-65-
-------
In the study of droplet formation in the vacuum flask, molten slag was
simulated by a high viscosity oil and by liquid mercury. The physical
properties of these two liquids are shown in Table XIB.
TABLE XIB
.Physical Properties of Liquid. Oil and Mercury
Density Viscosity Surface Tension
(S/cc,I Poise dynes /cm
Oil (25°C) 0.87 50 26
Mercury (25°C) 13.5 0.015 480
Molten Slag 2.5 50 480
(0.2 basicity,
8% sulfur,
2800°F)
The test was run under identical conditions (0.034 in. diameter orifice,
50 mm Hg pressure) for the liquids. The mercury would not form droplets
but jetted through the orifice, while the high viscosity oil formed
stringers of approximately 1 cm in length at a rate of approximately 1
per second. In the vacuum tower study, however, droplets could be formed
from molten slag having the same surface tension as mercury but viscosity
of approximately 100-150 poise resulting from operation at a temperature
of approximately 2600 F. This drop formation was, therefore, accomplished
with molten slag having a viscosity approximately 10,000 times greater
than that of mercury through an orifice only three times the diameter of
that used in the vacuum flask study. Because of the great differences
in the behavior of the mercury and the slag having similar surface tensions
but greatly different viscosities, it was concluded that viscosity was
the controlling variable in droplet formation.
The results of the vacuum granulation study are presented in Tables XIIB
and XJIIB and in Figures 19-B and 20-B. Figure 19-B shows the effect
of orifice diameter on the total and external specific surface areas
of the granulated slag produced in a 20 in. Hg vacuum from the molten
slag at approximately 2600 F. As seen, both surface areas increase with
increasing orifice size. Below one-sixteenth inch orifice diameter, it
was not possible to obtain flow of the slag through the orifice. The
increase of total surface area with increasing orifice diameter may be
explained by an increase of the slag droplet residence time at the
orifice which permits a greater opportunity for hot slag expansion to
occur.
Although run reproducibility is poor, Table XIIB tends to indicate that
the total surface area of the granulated slag is not a strong function
of vacuum at an orifice diameter of 3/32 in. and a molten slag temperature
of approximately 2600 F. On the other hand, the external surface area
decreases with increasing vacuum at these same conditions. Introduction
of the molten slag into the tower at atmospheric pressure produced slags
-66-
-------
Table XIIB
VACUUM GRANULATED SLAG
Run No.
7-2-1
7-2-2
7-12-1
7-12-2
8-2-2
8-3-3
8-11-1 '
8-11-2
, 8-12-1
7" 8-12-2
8-23-1
8-30-1
8-30-2
(Original
urigranul.
slag)
\
Vacuum
(in. Hg)
28
28
17-21
15-24
7-16
0
16-25
8-12
17-22
26-28
27
9-18
16-22
—
Orifice
Diameter
(in.)
3/32
3/32
3/32
1/16
1/8
1/4
1/8
3/32
3/32
3/32
3/32
3/32
3/32
-
Temperature
Range (°F)
2440-2600
2500-2950
2290-3240
2340-2940
2200-2490
2520-3240
2250-2850
2210-2980
2660-2910
2340-2720
2300-3280
2290-2930
2280-2800
Density
(g/cc)
2.48
2.30
2.39
2.47
2.30
1.85
2.22
2.41
2.37
2.21
2.49
2.35
2.33
2.55
- EXPERIMENTAL RESULTS
Specific
Surface
(cm2/g)
18.91
41.6
57.42
43.23
53.39
72.71
61.24
44.45
36.02
38.62
31.78
48.80
49.10
Total
Surface
(m2/g)
0.33 Str
0.97 Bea
0 .41 San
1.40 San
9.13 Str
<0.1 Sma
5.28 Str
0.40 Str
1.91 Var
1.26 Str
2.54 Str
0.42 Sma
0.26 Sma
0.38
Sample
Description
Stringers, beads, s
Beads, hollow spher
Sandy, porous mater
Sandy, porous mater
Stringers, beads
Small, porous, irre
Stringers, beads, h
Stringers, various
Various size beads
Stringers, beads, r.
Stringers, beads, >.
Small beads, porous
Small beads, porous
:heres
:s , filaments
rial
particles
3How spheres
size beads
~llow spheres
2How spheres
material
material
-------
Table XIIIB
00
Run No.
7-2-1
7-2-2 (beads)
7-2-2 (non-sp!
7-12-1
7-12-2
8-2-2
8-11-1
8-11-2
8-12-1
8-12-2
8-30-1
8-30-2
VACUUM GRANULATED SLAG - PARTICLE SIZE
DISTRIBUTION
Weight Per-'Cents
1 i
+— -— +6
Stringers 4 4
20.9 31.8
22.1 33.8
.erical) - 14.8 66.0
30.7
3.0
8.6 — ; 4.2
11.4 — 5.0
1.4 2.7 5.1
0.5 — 4.4
2.5 ,6.7 17.6
1.7 2.5
9.9
-6 +10
26.4
24.5
17.3
11.7
45.0
22.5
22.3
30.8
31.5
29.9
19.0
17.3
-10 4-20
17.9
14.5
1«7
43.5
51.0
48.9
40.0
44.2
47.6
29.8
57.8
51.6
-20
2.8
0.1
0.2
14.1
1.0
15.9
21.3
15.8
16.0
13.6
19.0'
21.2
-------
O Total Surface
D External Surface
7.0
6.0
5.0
4.0
CD
0)
J-l
-------
toO
Csl
CO
0)
CD
o
N
00
-70-
-------
2
of total surface area less than 0.1 m /g, that is, less than that of
slags that have been crushed and ground (see Table XIIB). A possible
explanation for this decrease may be surface sintering that sealed off
the pores leading to the interior of the particle.
Figure 20-B shows the relationship between molten slag temperature
as measured by an optical pyrometer and the total and external surface
areas of slag produced through a 3/32 in. diameter orifice in a 28 in.
Hg vacuum. As seen, the total and external surface areas both increase
with increasing temperature. It is believed that as the temperature
increases, the viscosity decreases and the slag droplet becomes more
fluid. Thus, dissolved gases more readily escape the fluid and cause
the slag to expand; thereby, increasing the surface area.
Conclusion
Experimental results indicate that the vacuum shot tower technique is
an effective means of increasing the total surface area of typical CaO-
SiCL-Al-O,. slags from two to five times the value for crushed slag.
Based on these findings, it appears that slag granulation can serve as
a means of providing increased total surface area slags for desulfurization.
SLAG DESULFURIZATION
Introduction
The slag produced in the combustor will contain between six and eight
percent sulfur. By desulfurizing the slag, the problem of disposing of
a high sulfur bearing material, which could be a potential source of
pollution, is eliminated. Instead, part of the desulfurized slag is
recycled to the combustor to take advantage of its lime content
and the rest becomes a salable by-product for use in road construction.
The elemental sulfur recovered from the slag is either a useful by-
product or a harmlessly disposable waste (depending on sulfur market
conditions). Because of the importance of desulfurization it was
decided to conduct a study to determine the effect of various process
parameters on reaction kinetics. Slag desulfurization was accomplished
by reacting the slag with steam at elevated teirparatures to produce an
off gas containing elemental sulfur, H~S and SO,,.
Experimental Procedure
The experimental program was conducted to determine the effect of slag
properties and reaction variables on the kinetics of desulfurization.
In particular, the effects of reaction temp.erature, slag external sur-
face area, sulfur content of the slag, and concentration of water in
the reaction gas on the time to achieve desulfurization and on the
gaseous products of reaction were investigated.
The experimental setup is shown in Figure 21B. The reaction chamber
was a 1-1/8 I.D. ceramic tube thirty inches long. Steam was carried
into one end with an inert gas (N_), contacted the slag, and exited
into a series of condensers and traps designed to collect each individual
species in the offgas. Steam for reaction was produced in a flask heated
by a Variac-controlled mantle. At the beginning of each run the flask
was charged with water and weighed. By varying the nitrogen flow rate
into the flask and by adjusting the Variac both the mole fraction of
steam in the gas and the throughput of steam to the reactor could be
controlled. After each run the flask was stoppered, cooled, and
weighed to determine the steam used.
-71-
-------
Rotameter
1-0
Variac
Electric Furnace
Heating Mantle
Leco Sulfur
Dioxide
Titrator
Condensers
Offgas Rotameter
CuSO, Solution
FIGURE 21B-DESULFURIZATION KINETICS EXPERIMENTAL APPARATUS
-------
The ceramic reaction tube was charged with 40 to 60 grams of slag
before it was placed in the electric resistance furnace. After a
nitrogen purge of 15 minutes to remove air the ends of the tube were
sealed, and the tube was placed in the furnace. After the tube reached
the run temperature, its exit end was connected to the series of con-
densers and traps as shown in Figure 21B. The steam flask was connected
to the tube entrance and with the introduction of steam and nitrogen
the run was started.
As the gaseous reaction products exited the tube they were cooled by
two water-cooled condensers outfitted with Erlenmeyer flasks at the
bottom to trap condensing material. Elemental sulfur and unused steam
condensed in these flasks. The remaining offgas (N?, H^S, and S0_) passed
out of the flasks, through a rotameter, and to a section of the offgas
system where a bomb sample could be taken (Figure 21B).
The offgas then proceeded to a flask containing a 0.1 Molar solution
of copper sulfate (CuSO,). The offgas was bubbled into approximately
500 milliliters of this solution which precipitated hydrogen sulfide
as copper sulfide according to the following equation:
H0S + CuSO. j=~ CuS I + H-SO.
24 tt 2 4
Additionally, sulfur dioxide was absorbed in the solution due to
the high solubility of SO- in water. Sulfur dioxide that escaped
this solution passed to a Leco titrator where it was absorbed in
an acid solution and was titrated with potassium iodate.
At the end of a run, the two final absorption flasks were disconnected
from the system preceeding it and a flow of N was introduced into
the CuSO, to push some of the dissolved S0? into the Leco sulfur dioxide
titrator. The remaining SO- that would not leave the CuSO, solution
was determined by a direct titration of the solution with KIO . The
reactions and theory explaining this titration for SO- are identical
to those pertaining to the Leco sulfur dioxide titrator (explained in
Appendix A under Combustor Operation). The sum of the above three titrations
indicated the cumulative amount of SO- produced. Hydrogen sulfide was
determined by titrating an aliquot of the CuSO, solution (after it has
been filtered and diluted to volume) with ethylene-diaminetetraacetic acid
(EDTA) to determine the concentration of cupric ion (Cu ) remaining in
solution. A 50-ml aliquot of the diluted CuSO, solution was withdrawn
into a beaker, to which was added 10 ml of a pH 5.0 to 6.0 buffer solution,
5 drops of PAN indicator and 8 drops of xylenol orange indicator. The
sample was initially pink. Upon titration with 0.020 molar EDTA solution,
it turned blue and then to green. The blue to green color change
was the end point of the titration. The difference between the cupric
ion concentration initially in solution (i.e., the initial CuSO concentration)
and the amount at the end of the experiment indicated the amount of CuS
precipitated and therefore the amount of hydrogen sulfide that passed
through the solution.
-73-
-------
To complete the experimentation, the ceramic reaction tube containing
the desulfurized slag was sealed, removed from the furnace, and cooled.
The slag was dumped from the tube and the total sulfur remaining in the
slag was determined by the combustion-iodometric method (described in
Appendix A) .
Discussion of Results
Slag desulfurization is a heterogeneous reaction between water vapor
in the gas phase and solid sulfur, presumably existing as calcium
sulfide in the slag. Consequently, a number of rate-determining
steps may be postulated. The desulfurization reaction rate may be
controlled by the transfer of water vapor from the bulk gas phase
to the slag particle, diffusion of reactant and products through the
pores of the slag particle, or by chemical reaction kinetics. The
latter is generally a strong function of temperature whereas the mass
transfer steps show a much smaller dependence.
Although the reaction steps involved in the desulfurization of slag
with water vapor are not known, it can be assumed that a simple
stoichiometric relationship exists between the reactant and product
species involved. Assuming that the reaction is controlled by the
mass transfer of water vapor from the bulk gas phase to the slag, a
simple material balance around a differential reactor can be written
as :
_ _
P dt 8 (1)
where w = weight of the slag
p = density of the slag
C = sulfur content of the slag
t = time
c = stoichiometric conversion constant
k = mass transfer coefficient
a - slag specific surface
y = mole fraction of water in the gas phase
Equation (1) assumes that chemical kinetics are very rapid, and that
the rate of sulfur removal from the slag particle is directly proportional
to the mass transfer of water vapor from the gas phase to the slag
particle. The mass transfer coefficient is a function of the flow
characteristics in the reaction system and depends as well upon the
physical properties of the gaseous constituents. In general*, k is
a function of the Reynolds Number and the Schmidt Number. The Schmidt
Number, a dimensionless group of variables characterizing the physical
properties of the gas, remains relatively constant for small temperature
*Perry, J. H., Chemical Engineers Handbook, McGraw Hill, New York
1950, p. 547
-74-
-------
variations. Consequently, if it is assumed that the reaction proceeds
over small temperature ranges, the Schmidt number will present no
significant effect on the mass transfer coefficient. With this assumption,
the relationship between the mass transfer coefficient and system flow
parameters given in the reference* can be expressed by the proportionality:
kg « D-°-41G°-59 -
where D = diameter of the slag particle
C = flow rate of the fluid
Substituting Equation (2) into one, rearranging and integrating for a
constant flow system with slag particles having initial sulfur con-
centration C . , the relationship:
where C~ = constant defines the sulfur level in the slag as a function
of reaction time.
The results of the experimental work show that the slag desulfurization
reaction is essentially controlled by the mass transfer of water vapor
from the bulk gas phase to the slag particle. A correlation relating the
percent sulfur removed from the slag with slag physical properties and
system flow characteristics is presented in Figure 22B. As can be
seen, the direct proportionality between slag desulfurization and system
properties is little influenced by temperature over the range of 1800-
2100°F. The lack of a strong temperature effect implies that chemical
reaction kinetics are fast relative to the mass transfer rate.
Perhaps the most significant variable is time at temperature. For small
particle sizes, and particularly at long times (20 minutes or more),
the percent desulfurization data deviate considerably to the right
from the linear mass transfer relationship. This deviation exists
because of the sintering effect experienced with fine slag particles at
temperatures of 2000°F or more. It was observed that at slag residence
times of 20 minutes or more (depending upon slag particle size) , the
discrete slag particles tended to fuse together and render the packed
bed reactor impervious to gas flow. This was evidenced, not only
by considerably increase in back pressure during experimentation, but
also by difficulty in removing the packed bed from the system.
Inasmuch as the correlation of Figure 22B is based on the slag surface
area characteristics prior to insertion into the reactor, the true
specific surface as a function of time and temperature is unknown.
Additionally, the essentially constant level of desulfurization at long
residence times for the slags implies that a change in mechanism also
*Perry, J. H. , Chemical Engineers Handbook, McGraw-Hill New York
1950, p. 547 ' ' '
-75-
-------
100
A
oe
«
o
o
(U
VJ
3
to
C
01
u
75
o
en
50
Synthetic Slags.
d 2100°F
O 2000°F
1900°F
C 1800°F
O 2100°F
O 2100°F
-10 +20 Mesh
-10 +20
-10 +20
-10 +20
-40 +60
-60 +100
D 2100°F -140 +200
6
Combustion
2100°F -40 +60
-10
20
30 .40 . 50
[aG-59yt/WC..D-41l xlO"2
*fr
100
500
FIGURE 22B-EFFECT OF PROCESS PARAMETERS ON. SLAG DESULFURIZATION
-------
occurred. It is probably true that the sintering effect brought about
a collapse or decrease in the pore size of the slag particles making
them relatively impervious to the transport of reaction water to the
slag interior. Speculating, it would appear that Knudsen diffusion
may be the primary mechanism for bringing about desulfurization at this
time.
With the exception of two points, all of the data presented in Figure 22B
were obtained with synthetic slags produced in the laboratory. These
slags did not contain iron oxides. On the other hand, the slags produced
in the combustor contained about one percent iron oxides. Since the
latter impurities could have an effect on slag desulfurization, two
experiments have been completed on the desulfurization of combustor
slags. These data are presented as triangles on the correlation. The
results show that for a combustor slag initially containing 8.5 percent
sulfur, desulfurization in excess of 99 percent can be achieved in less
than two hours at 2000°F. It should be mentioned that the combustor
slags did not exhibit the tendency to sinter that was exhibited by the
synthetic slags.
Although the analysis of the desulfurization data (particularly with
regard to correlating the offgas analysis with regard to system
variables) is still continuing, typical offgas compositions as a function
of desulfurization time are presented in Figure 23B. The data shown
are for a coarse minus 10 plus 20 mesh slag, desulfuriz'ed at 2000°F.
The offgas is a rather complicated function of water flow rate, particle
size, temperature, and partial pressure of water vapor in the gas phase.
The relationship between these variables is not yet understood. Never-
the less, Figure 23B is used to illustrate the general way in which
hydrogen sulfide, sulfur dioxide and elemental sulfur content vary with
time. In all of the experimentation conducted thus far, H S to S0~ ratios
have varied from a low of about 1 to a high of about 15. Elemental sulfur
recovered depending upon operating conditions, varied from about 10 to
55 percent. In general, higher temperatures favor the formation of
elemental sulfur and lower H~S and SO,., ratios. Low water input rates
tend to produce high H?S and S0« ratios.
The desulfurized (99 plus percent) experimental combustor slags pro-
duced an offgas whose cumulative elemental sulfur content was 53 per-
cent and the cumulative H-S to S0_ ratio was 1.6 at the end of two
hours.
-77-
-------
o
ex
o
o
"O
4)
•a
C
C
0)
o
0)
§
3
(0
(0
00
U-l
U-l
o
01
N
3
CO
0)
Q
* TC
-------
APPENDIX C
PROCESS SIMULATION AND ECONOMICS
Equipment Cost
To determine the capital investment for the Two-Stage Coal Combustion
Process, equipment was grouped into six main equipment complexes desig-
nated as series 1000-coal preparation, 2000-slag preparation, 3000-
flux preparation, 4000-air preparation, 5000-combustor, and 6000-slag
desulfurization. The cost of each item of equipment in a complex
was estimated based on a size (cost) controlling process variable(s).
Those items of equipment having a common size controlling variable
were then combined to yield expressions of the form:
C = C SA (1)
o
where C = estimated cost (1972)
C = constant
S = magnitude of size controlling variable, and
A = constant
which permit combined equipment costs to be estimated for any value
of the size controlling variable.
The energy and material balance computer program is run to determine the
capacity and/or temperature of the process streams which control the
size and, consequently, the cost of the equipment. Individual equipment
costs are determined using equation (1) and combined to yield the cost
of an equipment complex. Process equipment cost is the sum of the costs
of the individual equipment complexes.
To aid in the following discussion, the process flow diagram is presented
as Figure 1C with Table 1C designating the process streams shown in
Figure 1C. Also a schematic layout of the process equipment designated
by number appears as Figure 2C.
Coal Preparation Complex
A list of the coal preparation equipment, the design basis, and
estimated cost are presented in Table IIC.
Lump coal from storage is belt conveyed '(equipment number 1010-1011)
into coal bucket elevator (1020) and then into a surge bin (1030)
located above a coal crusher (1040-1043). The crushed coal is
screened (1050-1052) and the plus 1/8 inch fraction recycled to the
crusher using belt conveyor (1070-1071). Crushed coal is dried in
coal dryer (1055) and belt conveyed (1060-1063) to a coal bunker (1080).
A belt conveyor (1090-1092) is used to transport coal to surge bin
(1100) prior to entering the pneumatic coal injection system (1110).
-79-
-------
FIGURE - 1C PROCESS FLOW DIAGRAM
f\lK
I
00
o
I
-------
TABLE 1C
PROCESS STREAM DESCRIPTION - TWO STAGE COAL COMBUSTION PROCESS
Stream Stream Description
1 Coal
2 Coal transport air
3 Coal-air mixture to combustor
4 Desulfurized slag into combustion air stream
5 " . Preheated combustion air
6 Preheated air-flux mixture to combustor
7 Total air to combustor
8 Molten iron from combustor
9 CaS bearing slag to desulfurization complex
10 Lime/limestone into combustion air stream
11 Total combustor offgas
12 Air to steam generation unit
13 Stack gas from power plant
18 CaS slag to crusher
19 Silica/alumina into combustion air stream
20 Recycled desulfurized slag to cool stream 9
21 Crushed slag to desulfurization reactor
22 Desulfurized slag not recycled to stream 9
24 Desulfurization offgas
25 Desulfurized slag from desulfurization reactor
26 Desulfurized slag to storage and sale
27 Steam to desulfurization reactor
29 Granulated iron from combustor
30 Water to granulator
31 Steam from granulator
32 Combustor offgas sent to air preheater
33 Compressed combustion air
34 Air to air preheater
35 Flue gas from air preheater
36 Combustor offgas to power plant
41 Sulfur product
42 Offgas from sulfur condenser
-------
FIGURE -2C PROCESS EQUIPMENT LAYOUT
i
00
-------
TABLE IIC
COAL PREPARATION - EQUIPMENT COSTS
Number Equipment
1010 Conveyor to Bucket Elevator
1011 Motor and Drive
Design Basis
400 foot length, 36 inch wide belt with
10 foot rise, 460 TPH coal, 15 hp motor
Cost, $M
37.4
10.3
1020
Bucket Elevator
45 foot high, 250 TPH coal
52.0
1030
Surge Bin above Crusher
hour residence time, 4160 ft-% carbon steel
15.6
i
co
1040 Coal Crusher
1041 Coal Splitter
1042 Motor & Drive
1043 Speed Reducer
4 inch nominal feed, Gundlach Gage-Paktor
model 50-2C 4R, 45 hp, 300 TPH coal, upper motor
300 hp, lower motor 200 hp
19.7
1.5
10.3
27.6
1050 Screens
1051 Motor and Drive
1052 Splitter
6 foot x 12 foot - triple deck screens, 2 % hp motor, 11.6
125 TPH coal 0.9
5.5
1055 Coal Dryer
1060 Conveyor to Coal Bunker
1061 Motor and Drive
1062 Tripper
1063 Cover
No 135 Heyl Patterson Fluid Bed Dryer System
400 TPH Coal, 8% Water to 1% Water
400 foot length, 36 inch wide belt with 50 foot
elevation, 500 TPH coal, 35 hp motor
335.0
37.4
16.0
22.8
9.5
1070 +1/8 inch recycle conveyor
1071 Motor and Drive
100 foot length, 16 inch wide belt, 50 TPH
coal recycle, 5 hp motor
7.3
8.8
-------
CONTINUATION OF TABLE IIC
Number
1080
Equipment
Bunker
Design Basis
12 hour storage, 350,000 ft3 capacity
\ inch carbon steel
Cost, $M
363.0
1090 Conveyor- to Injection System
1091 Motor and Drive
1092 Cover
200 foot length, 36 inch wide belt with 60 foot
elevation, 460 TPH coal, 40 hp motor
49.6
20.3
11.4
i
CO
•P-
1100 Surge Bin-Coal Injector
1110 Injection System
% hour storage, 1500 ft3 capacity, 150 TPH coal
3 injection cycles per hour, 1350 ft capacity
100 TPH coal
5.7
-------
The pneumatic injection system transports the coal into the combustor
lances. The size and cost of all the equipment of a coal preparation
complex is a function of the process coal rate (stream 1).
Slag Preparation Complex
A list of the slag preparation equipment, the design basis, and
estimated cost are presented in Table IIIC.
Desulfurized slag from the desulfurization reactor is belt conveyed
(2010) into a slag bunker (2020). Slag is belt conveyed (2030) from
the bunker into a surge bin (2040) prior to entering the desulfurized
slag pneumatic injection system (2050). The pneumatic injection system
transports slag into the combustor lances. An alternate method of
putting slag into the combustor might be to dump the slag via a star
valve located on the combustor. This, however, was considered in the
combustor evaluation. All equipment was sized to handle 100 TPH of
desulfurized slag; consequently, the slag preparation complex cost is
a function of the recycled desulfurized slag rate to the combustor
(stream 4).
Flux Preparation Complex
Flux preparation includes equipment for taking flux materials, (in
most cases limestone, however, silica and alumina may also be required
under certain process conditions) from a storage bunker and pneumatically
injecting these fluxing agents into the combustor. A list of the flux
preparation equipment, the design basis and the estimated cost are
presented in Table IV-C.
Fluxing agents fr0m a storage bunker (3010) are belt conveyed (3020)
to a surge bin (3030) prior to entering the flux pneumatic injection
system (3040). The flux pneumatic injection system transports the
fluxing materials to the combustor lances. All equipment was sized to
handle 25 TPH of fluxing materials; consequently, the flux preparation
complex cost is a function of the sum of the limestone and/or lime
added (stream 10), and the silica and/or alumina added (stream 19).
TABLE 1VC
FLUX PREPARATION - EQUIPMENT COST
Number Equipment Design Basis „ Cost, $M
3010 Fluxing Material Bunker 6 hour storage, 4800 ft 13.0
capacity, 1/2 inch carbon steel
3020
3021
3030
3040
Conveyor to Surge Bin
Motor & Drive
360 foot length, 20 inch wide
belt, 10 hp motor
Surge Bin-Flux Injector 1/4 hour residence time,
180 ft capacity
Flux Injection System
3 cycles per hour, 220 ft"
capacity
17.1
8.9
5.3
58.8
-85-
-------
Number Equipment
2010
2020
Conveyor from Desulfurization
to Bunker
Bunker
TABLE IIIC
SLAG PREPARATION - EQUIPMENT COSTS
Design Basis Cost, $M
200 foot length, 20 inch wide belt, 10 hp motor 40.0
4 hour residence time, 12480 ft^ capacity, % inch 37.7
carbon steel
i
00
2030 Conveyor to Surge Bin
2031 Motor and Drive
2040 Surge Bin-Slag Injector
100 foot length, 20 inch wide belt, 7% hp motor
hour residence time, 1840 ft^ capacity
9.0
8.9
28.1
2050
Desulfurized Slag Injection
3 cycles per hour, 984 ft capacity
192.0
-------
Air Preparation Complex
The air preparation complex consists of the coal transport air compressor
(4010) which compresses air (stream 2) for coal injection, the combustion
air compressor (4020) which compresses air (stream 33) prior to entering
the air preheater, and the air preheater (4030) which heats combustion
air (stream 33) prior to entering the combustor. A list of these items
is presented in Table VC where the design basis and estimated cost is
given. The injection air compressor(s) is sized to yield sufficient
10 psig transport air to pneumatically inject coal into molten iron at
an injection ratio of ten pounds of coal per pound of air. The estimated
cost shown in Table VC for the injection air compressor is based on a
270 brake horsepower requirement. The brake horsepower requirement can
be established provided the desired outlet pressure air volumetric flow
rate and compressor efficiency are specified; consequently, the injection
air compressor cost for any process case can be estimated by scaling the
data of Table VC once the brake horsepower requirement is established.
However, if the compressor efficiency and air outlet pressure are
specified, then the cost is only a function of the injection air rate
(stream 2). Regardless of injection air rate a minimum of two compressors
are costed.
TABLE VC
AIR PREPARATION - EQUIPMENT COSTS
Number Equipment Design Basis
4010 Injection Air Compressor Centrifugul steam driven
turbine, 270 bhp
4020
4030
Combustion Air Compres-
sor
Centrifugul steam driven
turbine, 0500 bhp
Cost, $M
73.3
800.0
Combustion Air Preheater Heat exchanger with direct
firing having 20000 ft of
heat transfer surface with
the following air temperature
ranges and materials
77-1000F, carbon steel
1000-1200F, type 501 stain-
less steel
1200-1400F, type 405/410
stainless steel
1400-1500F, type 430 stain-
less steel
1500-1600F, type 304 stain-
less steel
1600-1800F, type 310 stain-
less steel
1800-2100F, inconel
62.0
155.0
186.0
223.0
273.0
310.0
372.0
-87-
-------
The combustion air compressor (4020) is sized based on total air re-
quirement of the process, less the air required for coal injection.
Combustion air is pressurized to 5 psig using centrifugal steam driven
turbine compressors operating at 70 percent efficiency. As seen in
Table VC, the estimated cost for a 6500 bhp compressor is $800,000.
A minimum of two combustion air compressors are used; however, when
the brake horsepower requirement exceeds 6500 bhp, a number of equally
sized smaller compressors are costed. In the same manner, as for the
.injection air compressor, the estimated cost of the combustion air com-
pressor can be made a function of only the combustion air (stream 33).
The air preheater cost will depend upon the required heat transfer
surface and the materials of construction, both of which depend upon
the exit temperature of the combustion air. In Table VC are shown the
estimated costs for air preheaters having a 20000 square foot heat
transfer surface at various exit air temperatures and the design basis.
The air preheaters are assumed to be a parallel arrangement of heat ex-
changers each with provision for direct firing of the combustor offgas.
The air preheating equipment cost is determined using an overall heat
transfer coefficient of 5 BTU/ft -hr-°F to establish the required heat
transfer surface to raise combustion air (stream 33) from 77°F to the
desired exit temperature. Since the exit air temperature is known, the
cost for a 20000 ft air preheater can be determined from Table VC. The
air preheaters are limited to a maximum size of 20000 ft ; consequently,
several equally sized units are costed depending on the total heat
transfer surface requirement.
Combustor Complex
The combustor complex consists of a combustor with lances, enclosed
conveyors which remove liquid combustor slag, enclosed conveyors which
combine this hot slag with cooler slag for belt protection, and an iron
granulation system. The depth of molten iron in the combustor is
established by the required immersion lance depth to fully dissolve the
coal.• Experimentation has shown that a 24 inch immersion depth will
suffice; consequently, the combustor is conservatively assumed to contain
a 3.5 foot molten iron bath. The area of the combustor is controlled by
either of two design criterias. There must be a sufficient weight of iron
in the combustor to accommodate the dissolving coal such that a solution
rate of .25 weight percent carbon/minute is not exceeded. That is, the
combustor area (or the weight of iron since the.height is fixed) must
be sufficiently large, to dissolve the coal.' In addition, there must be
sufficient combustor area to prevent the offgas velocity from the bath
surface from being excessive with subsequent carryover of dust. Based
on steelmaking basic oxygen furnace technology, the superficial offgas
velocity was limited to 30 ft/sec. In general, when air was used to
burn from the molten iron, the superficial offgas velocity criteria (30
ft/sec) controls the combustor area (weight of metal in the combustor).
When oxygen is used to burn metal carbon, the solubility rate will
determine the bath weight. The volumetric flowrate of the offgas will
depend on a number of factors such as the combustor pressure and temperature,
and coal composition and the quantity of air needed.
-88-
-------
In order to facilitate estimating the cost of various sized combustors,
the 48 foot I.D. combustor as depicted in Figure.3C was used. As seen,
the combustor is a steel shelled refractory lined vessel with a 2 1/2
foot refractory lining at the base and sides. In Table VIC is shown
costs for the combustor shell (5010), refractory lining (5011), lances
(5012), and lance cooling pumps (5013). Twenty five lances and pumps
were assumed to be required. The combustor shown in Figure 3C will
contain 1200 tons of iron. The required weight of iron as determined
by either the coal solubility rate or offgas superficial velocity criteria
was used to scale the combustor costs of Table VIC. The number of
combustors to use in a given power plant application is assumed to be
three equally sized units—any two of which can supply the full energy
requirements of the plant.
The iron granulation system includes the equipment designated as 5040
to 5048 in Table VIC. The cost of this- system is based on the quantity
of molten iron contained in a single combustor and not by the iron
production rate from the iron contained in the coal. That is, the iron
granulation system is sized to remove all the iron from a combustor in
a specified time. To establish the cost of the iron granulation system,
the total combustor iron is divided by the specified removal time to
obtain an iron removal rate and the costs of Table VIC scaled accordingly.
The combustor complex includes enclosed belt conveyors for removal of
slag from the combustor and belt conveyors used to place a cool layer
of desulfurized slag on the above conveyors. These are designated as
slag pan conveyor (5020) and slag recycle apron conveyor (5030). The
estimated costs shown are based on 120 TPH slag removal rate from the
combustor and 20 TPH desulfurized slag recycle rate to the pan conveyor.
The cost of the pan conveyor is scaled based on the CaS bearing slag
leaving the combustor (stream 9). The cost of the recycle apron con-
veyor is based on recycled desulfurized slag (stream 20).
Slag Desulfurization
The slag desulfurization complexes receives CaS bearing slag from the
slag pan conveyor (5020-combustor complex) and crushes it in a primary
roll type crusher and a secondary crusher (6010-6021). A bucket elevator
(6030) is used to convey the crushed slag into a refractory lined shaft
reactor (6040-6041). The sulfur bearing offgas from the shaft reactor
proceeds to sulfur condenser (6060) where the sulfur is condensed out
and pumped (6080) to a sulfur collection tank (6070). The remaining
H~S-SO offgas is sent to a Glaus sulfur recovery plant (6090) where
sulfur is produced and sent to the sulfur collection tank (6070). The
desulfurized slag leaving the shaft reactor is split into three streams,
respectively, to the slag bunker via a conveyor (2050-slag preparation),
recycled back to the slag pan conveyor via the slag recycle apron
conveyor (5030-combustor) and sent to pile storage using conveyor (6050).
In Table VIIC, the equipment comprising the desulfurization complex, the
design basis, and estimated cost are presented. For cost estimating
-89-
-------
Air Lance
Coal Lance
O
I
Refractory
2 - 3 —
x, \T N \ \ \ \-\\ x
FIGURE 3C-48 FOOT DIAMETER COMBUSTOR
-------
TABLE VIC
COMBUSTOR - EQUIPMENT COSTS
Number Equipment
Combustor
5010 shell
5011 refractory
5012 lances
5013 lance cooling" pumps
Design Basis Cost, $M
48 foot I.D. combustor as shown in Figure 3C 308.0
40 psig, carbon steel, containing 1200 tons 168.0
iron high density alumina brick, 25 lances and 80.0
pumps 7.0
5020 Slag Pan Conveyor
5030 Slag Recycle Apron Conveyor
Link belt conveyor, 70 foot length, 120 TPH
15 hp motor
100 foot length, 50 TPH, 7% hp motor
60.0
20.0
i
VO
Iron Granulation System
5040 tundish
5041 . runner
5042 water sprays
5043 tank
5044 hopper guide
5045 discharge bell
5046 cooling and settling tank
5047 granulator pump
5048 granulated iron conveyor
10 feet L x 6 feet W x 6 feet D, 1.5 feet thick 4.3
brick
4-15 feet long 2.4
20 inch spray plate with % to \ inch holes .1
54 foot x 22 foot x 20 foot deep tank 68.7
20 foot x 12 foot guide 1.5
110 foot long, 36 inch wide rubber belt with 27.7
variable drive, 15 hp
28 foot x 22 foot x 20 foot deep tank 25.6
10000 GPM, 15 psig, 150 hp 14.2
110 foot long, 36 inch wide rubber belt with 27.7
variable drive, 15 hp
-------
TABLE VIIC
DESULFURIZATION - EQUIPMENT COSTS
Number Equipment
6010 Primary Crusher
Design Basis
Gunlach model 45 DACC, heavy duty Two-Stage
Four Roll Crusher, 75 hp, 220 TPH capacity
Purchase Cost
42.0
6020 Secondary Crusher
6021 Speed Reducer
Gunlach model 50-2C4R Cage Packer, 220 TPH
capacity
30.6
25.0
6030
Bucket Elevator
50 foot height, 100 hp motor, 220 TPH capacity 28.0
I
vo
N>
6040 Desulfurizing Shaft Reactor
6041 Refractory
6050 Conveyor for Slag to Storage
6060 Sulfur Condenser
6070 Sulfur Collection Tank
% hour residence, 607» loading, 5000 ft3 50.0
alumina refractory 12.0
100 foot length, 14 inch wide belt 20.0
Heat load 6.3 MM BTU/hr, 6000 ft2 heat transfer 50.0
surface, stainless steel, 14 TPH sulfur
24 hour storage, 336 tons sulfur, 11200 ft 33.6
6080
Pump
3500 GPM, 45 hp motor, 15 psig
11.2
6090
BS&B Sulfur Recovery Plant
Black, Sivalls & Bryson installed plant to
produce 6 TPH of Sulfur
270.0
-------
purposes, the equipment designated as 6010-6041 is a function of the
CaS bearing slag entering the shaft reactor (stream 18). The
sulfur condensing, pumping, and collection system is a function of' the
total sulfur condensed from the offgas which, based on laboratory
experimentation, was set at one half of the total sulfur produced in
the process (stream 41). The cost of the Glaus sulfur recovery plant
is based on recovery of the remaining sulfur. The cost for the Glaus
sulfur recovery plant is from internal sources.
Cost Controlling Variables
In Table VIIIC is presented a summary of the process equipment grouped
according to the cost controlling variables which are used to scale
the costs shown in Tables IIC - VIIC. As seen, most of the cost con-
trolling variables are the rates (TPH) for the various process streams.
In the case of the injection and combustion air compressors, as previously
explained, additional simplification in cost estimating is possible if
the outlet pressures are specified. With the outlet pressures specified
at 10 psig for the injection air and 5 psig for the combustion air,
the cost of these items are proportional to streams 2 and 33, respectively,
If the external heat transfer coefficient and the outlet air temperature
is specified for the air preheator then the heat transfer area, or in
essence, the air preheater cost is proportional to the quantity of air to
be preheated (stream 33). The quantity of iron in the combustor(s)
which determines cost is calculated using as a design criteria, the
offgas velocity. If the combustor offgas rate (stream 11), composition
pressure and temperature are known then the volumetric offgas flow rate
can be established. Using 30 ft/sec as the desired offgas velocity,
the required surface area of the combustor can be determined. This area
is then divided by two to establish the area of each combustor for a
two-combustor operation. Since a 3.5 foot high bath of molten iron is
required for total coal solubility, the known volume of weight contained
in each combustor is established. The cost of a combustor is, therefore,
related to its iron content. To insure totally reliability, three instead
of two combusors are costed—each capable of supplying 50 percent of the
offgas requirement for a power plant boiler. The iron granulator is
sized to empty the contents of a single combustor in eight hours.
In Table IXC are presented the factors C and A to be used in equation
(1) to determine the cost of the various items of equipment. Also shown
is the cost controlling variable S. These factors (C and A) were
determined using the cost data of Tables IIC - VIIC, by combining the
equipment associated-with each cost controlling parameter and determining
the combined cost over a range of values for the cost controlling variable.
Plotting the resulting cost versus the cost controlling variable enabled
values for C and A to be determined.
o
Total Purchased Equipment Cost
To establish the total purchased equipment cost, the energy and material
balance computer program was run to ascertain process stream rates. For
a 1000 MW power plant using coal of the composition shown in Table XC,
the process stream rates of Table XIC were determined.
-93-
-------
TABLE VIIIC
COST CONTROLLING VARIABLES
Equipment
Complex Number(s)
Coal Preparation 1010 - 1110
Slag Preparation 2010 - 2050
Flux Preparation
Air Preparation
Air Preparation
Air Preparation
Combustor
Combustor
Combustor
Combustor
Desulfurization
Desulfurization
Desulfurization
Desulfurization
3010 - 3040
4010
4020
4030
5010 - 5013
5020
5030
5040 - 5048
6010 - 6041
6050
6060 - 6080
6090
Cost Controlling Variables
Coal rate to combustor, stream 1
Recycled desulfurized slag rate to
combustor, stream 4
Flux rate to combustor, stream 10 plus 19
Brake horsepower requirements to compress
injection air, stream 2
Brake horsepower requirement to compress
combustion air, stream 5
Heat transfer surface required to heat
combustion air to temperature, stream 5
Combustor iron requirement based on offgas
velocity criteria, stream 11
CaS bearing slag to shaft reactor, . '
stream 18 .
Desulfurized slag recycled to conveyor,
stream 20
TPH iron removal rate which is function
of the weight of iron in combustor
CaS bearing slag to shaft reactor,
stream 18
Desulfurized slag to storage,
stream 26
Fraction of total sulfur produced in
process stream 41 which is condensed out
prior to sulfur recovery plant
Fraction of total sulfur produced in
process stream 41 which is formed
in sulfur recovery plant
-94-
-------
TABLE IXC
EQUIPMENT COST FACTORS
Equipment
Coal Preparation 1010-1110
Slag Preparation 2010-2-5-
Flux Preparation 3010-3040
Air Preparation
4010
4020
4030
Combustor
5010-5013
5020
5030
5040-5048
Desulfurization 6010-6041
6050
6060-6080
6090
17,080 .809 Stream 1
19,348 .606 Stream 4
14,958 .603 Stream 10 plus 19
6,220
2,060
302
817
956
1,110
1,320
1,454
1,812
.883
.941
.989
.979
.983
.989
.991
.996
.992
Stream 2,
Stream 5~
Stream 5~
Stream 5~
Stream 5?
Stream 5~
Stream 5«
Stream 5~
Stream 5
(0-10000F)
(1000F-1200F)
(1200F-1400F)
(1400F-1500F)
(1500F-1600F)
(1600F-1800F)
(1800F-2100F)
8.834 .600 Tons iron in each of
three combustors
2,655 .652 Stream 18
1,605 .647 Stream 20
5,294 .612 Iron Granulation Rate
3,549 .735 Stream 18
1,584 .647 Stream 26
18,983 .608 50% of Stream 41
91,900 .600 50% of Stream 41
1 Assumes injection air will be compressed to 10 psig and combustion air
to 5 psig.
2
2 Assumes an overall heat transfer coefficient of 5 BTU/hr-ft F and
air leaving at maximum temperature in range specified.
Using the process stream rates of Table XIC, the values of C and A
from Table IXC and equation (1) enable the costs for each individual
equipment complex to be determined. For the process case under
consideration, the total purchased equipment cost is shown in Table XIIC,
where the cost is seen to be $7.04 MM.
-95-
-------
TABLE XC
COAL COMPOSITION
Ultimate Analysis
Total Carbon
Hydrogen
Oxygen
Nitrogen
Sulfur
Ash
Higher Heating Value, 17, Moisture,
BTU/lb
Med Ash
68.17o
5.0
7.3
1.5
3.6
14.
12500
TABLE XIC
PROCESS STREAM RATES - 1000 MW POWER PLANT
Stream
1
2
3
4
5
6
7
8
9
10
11
12
13
18
20
21
22
24
25
26
27
29
30
31
32
33
34
35
37
41
42
Ton/Hr
361.
36.
397.
103.
1186.
1312.
1222.
6.
166.
24.
1536.
2177.
3595.
333.
167.
327.
153.
107.
320.
50.
72.
10.
6.
3.
117.
1186.
188.
306.
1418.
13.
94.
Temp-F
108.
100.
108.
1501.
1182.
1189.
77.
2500.
2500.
77.
2500.
77.
304.
2000.
1501.
1800.
1501.
1501.
1501.
77.
77.
151.
77.
212.
2500.
77.
77.
399.
2500.
832.
832.
-96-
-------
TABLE XIIC
TOTAL PURCHASED EQUIPMENT COST
4
Equipment . S (TPH) Cost, $MM
1010 - 1110 1.830
1. Coal Preparation 361 1.83
2010 - 2050 103 0.320
2. Slag Preparation .32
3010 - 3040 24 0.100
3. Flux Preparation .10
4010 36 0.030
4020 1186 1.600
4030 1186 0.850
4. Air Preparation 2.48
5010 - 5013 8502 1.500
5020 333 0.118
5030 167 0.004
5040 - 5048 106 0.091
5. Combustor 1.71
6010 - 6041 167 .260
6050 50 .020
6060 - 6080 6.5 .059
6090 6.5 .261
6. Desulfurization .60
TOT A L PURCHASED EQUIPMENT COST 7.04
1. For an air preheater in 1000-1200°F cost range.
2. 850 tons of iron in each of three 38 foot diameter combustors.
Calculation presented in discussion section of report.
3. Iron granulation rate = 850 tons/8 hour = 106 TPH.
4.. Magnitude of size controlling variable, equation (1)
-97-
-------
Estimated Fixed Capital Requirements
Once the total purchased equipment cost is known, the total fixed capital
requirement for the Two-Stage Coal Combustion Process can be calculated
as shown in Table XIIIC. The application of factors for installation,
etc., engineering, and construction are commonly used in engineering
estimates. A contingency of ten percent, contractor's fee of six per-
cent, escalation to 1980 and interest during construction are included
to yield the total fixed capital requirement. As seen, the total fixed
capital is $23.33 MM and represents the retrofitting cost for installing
the process into an existing facility.
Power Plant Costs
In order to estimate the fixed capital requirement for a grass roots
1000 MW Two-Stage Coal Combustion Power station, fixed power plant
capital costs were obtained . These costs are shown in Table XIVC
for coal-fired and gas-fired boiler operations producing 3500 psig -
1000°F steam. Costs are shown using the standard FPC costing procedure
and include escalation to 1980 and interest during construction. As
seen, the total fixed capital requirement: for the coal and gas-fired
plants are $163.2 MM and $139.5 MM, respectively, and differ because
of boiler plant equipment costs and an oil burning standby facility
included in the gas-fired plant. The Two-Stage Coal Combustion Process
produces an offgas for use in a power plant boiler; consequently, in any
grass roots installation, it is probable that power plant costs can
be approximated using the costs of a gas-fired facility. Therefore,
to establish the cost of a Two-Stage Coal Combustion Power Station, the
capital cost of Table XIIIC ($23.33 MM) was combined with the gas fired
fixed capital requirement to yield a total cost of $162.8 MM. This is
shown in Table XIVC. The results of Table XIVC permit a comparison to
be made between a conventional coal burning facility and a Two-Stage
Coal Combustion Process—gas fired facility; as seen, the fixed capital
requirements are $163.2 MM and $162.8 MM respectively.
The coal fired plant has a net output of 1012 MW whereas the ATC gas
fired facility produces 1000 MW. The capital cost per kilowatt are
then 161.3 $/KW for the coal fired facility and 162.8 $/KW for the
Two-Stage Coal Combustion power station. Thus, an additional capital
requirement of $1.5/KW yields a non-SO_ polluting power station.
Operating Costs
Operating costs for the coal fired power plant and the Two-Stage Coal
Combustion power station are compared in Table XVC for a 70 percent
"Technical and Economic Feasibility of Advanced Power Cycle and Methods
of Producing Non-Polluting Fuels for Utility Power Stations", NAPCA,
Contract CPA 22-69-114, 1970.
-98-
-------
TABLE XIIIC
ESTIMATED FIXED CAPITAL REQUIREMENT
TWO STAGE COAL COMBUSTION PROCESS
1000 MW POWER PLANT
EQUIPMENT COMPLEX $MM
Coal Preparation / 1.83
Slag Preparation 0.32
Flux Preparation . 0.10
Air Preparation 2.59
Combustor 1.71
Desulfurization Q.60
1. Total Purchased Equipment Cost 7.15
Installation, Piping, Electrical
Instrumentation, Utilities
(707, of 1) 5.00
2. Physical Plant Costs 12.15
Engineering and Construction
(307» of 2) 3.63
3. Direct Plant Cost 15.78
Contingency (107o of 3) 1.58
Contractor's Fee (57o of 3) .79
4. T 0 T A L 18.15
Escalation to 1980 (14.757. of 4) 2.67
5. TOTAL • 20.82
Interest During Construction
(12.07. of 5) 2.51
6. TOTAL FIXED CAPITAL (1980) 23.33
-99-
-------
TABLE XIVC
ESTIMATED CAPITAL REQUIREMENTS
For 1000 MW Power Plant Systems
All Figures in $ Thousand
Conventional*
Coal Burning
Gas-Fired Boiler*
With Combustor
Land and Rights
Structure and Improvements
Boiler Plant Equipment
Turbine-Generators
Electrical Equipment
Misc. Power Plant Equipment
Station Equipment
1. TOTAL
Other Expense (12.5% of 1)
2. TOTAL
Eng.-Design-Const.-Super.,
Contingency (12% of 2)
3. TOTAL
Escalation (1980) .
(14.75% of 3)
4. TOTAL
Interest During Construction'
(12.6% of 4)
5. TOTAL
Standby Oil Facility
6. TOTAL
Two-Stage Coal Combustion
Plant Cost
7. TOTAL
Net MW Output
Capital Cost $/KW
30
9,107
55,492
34,612
10,138
463
1,572
111,414
1,393
112,807
13,537
126,344
18,636
144,980
18,267
163,247
0
163,247
0
163,247
1,012
161.3
30
7,582
39,200
34,612
10,138
463
1,572
93,597
1,250
94,847
12,436
107,283
15,824
123,107
15,511
138,618
1,725
139,485
23,334
162,819
1,000
162.8
*A11 cost and factors except coal combustion process cost from "Technical and
Economic Feasibility of Advanced Power Cycle and Methods of Producing Non-Pol-
luting Fuels for Utility Power Stations", NAPCA, Contract CPA 22-69-114, 1970.
-100-
-------
load factor operation. A capital interest charge of 14 percent, operation,
supplies, and maintenance costs at 2 percent of the fixed capital require-
ment, and coal at $0.30/MMBTU were assumed. As shown, the operating cost
is 6.57 mills per kilowatt hour for the coal fired facility and 6.82
mills per kilowatt-hour for the Two-Stage Coal Combustion power station.
Thus, the Two-Stage Coal.Combustion power station has a slightly higher
operating cost (.25 mills per kilowatt-hour) than the coal burning - SCL
polluting power station. When credits for iron ($20/ton), slag ($0.5/ton),
and sulfur ($20/ton) are taken, the Two-Stage Coal Combustion power
station operating cost is reduced to 6.40 mills per kilowatt-hour.
Power Plant Operating Costs and Process Parameters
Incorporated into the energy and material balance computer program are
sub-programs which allow the fixed capital requirement and operating
cost to be estimated as a function of various process parameters. That
is, the economic effect of process variables such as coal composition,
moisture in the coal, combustor temperature, flux composition, combustor
slag basicity and the percent sulfur in the combustor slag can be
established. In this manner, optimum economic ranges for these important
process parameters can be determined.
TABLE XVC
ESTIMATED OPERATING COST
1000 MW Power Plants
All Figures in Mills per Kilowatt-Hour
Natural-Gas-
Conventional Coal Combustion
Coal Burning Process
Capital Charge (14% Rate) 3.683 3.717
Operating, Supplies & Maintenance 0.323 0.326
Coal ($0.3/MM BTU) 2.567 2.710
Limestone ($3/Ton) 0 0.068
Iron Credit ($20/Ton) 0 (0.129)
Slag Credit ($0.5/Ton) 0 (0.025)
Sulfur Credit ($20/Ton) 0 (0.263)
TOTAL POWER COST
Mills per Kilowatt Hour 6.572 . 6.403
TOTAL POWER COST .
Without By-Product Credits 6.572 6.820
-101-
-------
Table XVIC shows the effect of coal composition on a number of process
variables and on operating cost. Three coals were used and classified
as high, medium, and low ash. The important conclusions from Table XVIC
are:
1. The required air preheat temperature is in the
narrow range of 1000-1200°F for the wide range of coal
used.
2. The capital requirement for the Two-Stage Coal
Combustion Process plant is essentially independent of
coal composition and is about $23/KW.
3. The net capital cost increase of a Two-Stage
Coal Combustion Process - power station over a con-
ventional coal burning facility is essentially
independent of coal composition and equal to
approximately $1.5/KW.
4. The operating cost ranges from a high of 6.73
mills/KW hr for the high ash coal to a low of 6.40
mills/KW hr for the medium ash coal.
In Figure 4C is shown the air preheat temperature and operating cost
(mills/KW-hr with by-product credit) as a function of combustor operating
temperature for a process using the medium ash coal of Table XVIC. As
seen, both the air preheat temperature and operating cost increase as
the combustor temperature increases. Consequently, the combustor
should be operated at as low a temperature as possible consistent with
maintaining a fluid slag. The suggested combustor operating temperature
range is indicated and the operating cost for conventional coal burning
plant is shown for comparison in Figure 4C. In the expected combustor
operating temperature range the air preheat temperature will be between
1000 to 1200°F.
The moisture content of the coal is an important economic consideration
because the decomposition of water to hydrogen and oxygen in the com-
bustor is highly endothermic. In Figure 5C is shown the effect of
the moisture content 'of coal on both the air preheat temperature and
the operating cost. As seen, both the air preheat temperature and
operating cost increase as the coal moisture increases. This is
expected since increased quantities of coal are required to overcome
the endothermic decomposition of water. Based on this result, the
input coal moisture should be reduced to as low a value as
practical. The suggested coal moisture range for coal entering the
process is shown in the figure and should be about 1-1 1/2 percent.
-102-
-------
TABLE XVIC
*
EFFECT OF COAL COMPOSITION ON OPERATING & ECONOMIC DATA
(1000 MW Power Plant)
Coal (Ultimate.Analysis)
Total Carbon
Hydrogen .
Oxygen
Nitrogen
Sulfur
Ash
Higher Heating Value, 1% Moisture, BTU/lb**
Air Preheat Temperature, F
Coal, TPH
Limestone, TPH
Lime, TPH
Sulfur Bearing Slag to Desulfurization, TPH
Combustor Offgas to Steam Generation, TPH
Combustor Air, TPH
Sulfur Produced, TPH
Iron Produced, TPH
Slag Produced, TPH
Coal Combustion Process Plant Cost, $MM
Power Plant Cost, Gas-Fired Boiler Plus
Combustor $/KW
Power Plant Cost, Coal Fired Boiler, $/KW
Net Increase in Power Plant Cost, $/KW
Cost of Coal at $0.30/MMBTU, $/Ton
Operating Cost, Conventional Coal Burping
Boiler, Mills/KW-HR
Operating Cost - Mills/KW-HR ($0.30/MMBTU
Coal)
With By-Product Credit
Without By-Product Credit
High Ash
62.6%
5.0
7.3
1.5
3.6
20.0
11700
1200
390
9
16
178
1395
1202
14
7
100
23.49
162.98
161.31
1.67
7.02
6.57
6.73
7.19
Med Ash
68.1%
5.0
7.3
1.5
3.6
14.5
12500
1182
361
24
0
166
1418
1222
13
6
60
23.33
162.82
161.31
1.51
7.50
6.57
6.40
6.82
Low Ash
74.6%
5.0
7.3
1.5
3.6
8.0
13500
1015
350
12
0
159
1443
1240
13
6
30
23.16
162.65
161.31
1.34
8.11
6.57
6.48
6.87
**
2500°F Combustor Temperature, 0.1 Slag Basicity, 8% Sulfur Slag
Dulong's Equation
-103-
-------
FIGURE 4C
OPERATING COST/AIR PREHEAT TEMPERATURE
VS COMBUSTOR TEMPERATURE*
o
o
VO
o
o
01
M
D
CU O
n, o
B CM
(II
.c
f* j ^3
o
w o
•i-l r-l
o
o
CO
Operating
Range
ConventionaJL Plant
Operating Cost
vO
lA OJ
O
vO U
CO
l-i
-------
o
o
vO
O
O
oj O
a. o
B CNJ
0) i—I
H
to
0)
0)
n o
p I ^3
o
J-l r-l
o
o
oo
FIGURE 5C
OPERATING COST/AIR PREHEAT TEMPERATURE
VS % MOISTURE IN COAL*
Operating
Range
Conventional Plant
vO
r-l
-,-1
a
(0
o
•vf t>0
• c
VO -r4
4J
ni
(!)
Ot
o
ro
2 3
7» Moisture in Coal
* 14.5% Ash Coal, 2500 F Combustor Temperature,
.1 Basicity Slag, 87. Sulfur Slag, Limestone
-105-
-------
A choice exists whether to use lime, limestone, or combination of the
two as coal ash fluxing materials. In some instances, lime may be pre-
ferred (despite its high cost) because it yields a combustor operation
with a lower air preheat temperature requirement. This results because
the endothermic calcination of limestone to lime in the conbustor is
eliminated if lime is used. In Figure 6C is shown the effect of the
fraction of limestone in the flux on operating cost and air preheat
temperatures. As seen, the limestone fraction has a significant
effect on operating cost. For example, below about 50 percent limestone
in the flux, the process has a higher operating cost than the con-
ventional plant, whereas above 50 percent limestone, the operating cost
is lower. Consequently, a minimum of lime (preferably none) should
be used in the process consistent with maintaining a reasonable air
preheat temperature. For extremely high ash or low grade coals,
lime may be required to reduce the air preheat temperature (or in
essence, air pireheater capital costs) and can also serve as a combustor
temperature controlling variable when process coal composition varies
widely during operation.
Slag basicity is an important economic consideration because basicity
is related to limestone consumption in the process. Figure 7C shows
the effect of slag basicity on both the air preheat temperature and
operating cost. As seen, the air preheat temperature increases from
about 1200°F at a basicity of 0.1 to about 1500°F at a basicity of 0.8.
Also the operating cost increases quite substantially in the 0.1 to 0.8
basicity range. Operating costs are affected by basicity in two important
ways. The process limestone requirement or the limestone cost per
kilowatt hour increases proportionately with basicity. Arid desulfurization
capital increases because more slag handling and larger sized equipment
is required. To yield a minimum operating cost as low a basicity as
possible, consistent with a fluid slag should be used. Experimental
studies indicate that operable slags of 0.1 - 0.2 basicity can be used.
Perhaps the most important process parameter is the percent sulfur
contained in the combustor slag. The air preheat temperature and
operating cost for the process case under consideration are shown in
Figure 8C, as a function of the percent sulfur in the slag. As seen,
the process operating cost increases substantially when the percent
sulfur in the slag is decreased below about 4 percent.
At about 4 percent sulfur in the slag, the operating cost of the Two-
Stage Coal Combustion Power Station becomes equal to the conventional
coal burning plant. At less than 4 percent sulfur, operating cost
increases substantially. Fortunately, bench-scale laboratory experimen-
tation has shown that slags containing 6 to 8 percent sulfur are
sufficiently fluid for use in the combustor.
The main reason for the predominant economic effect of slag sulfur
content lies in the fact that increased quantities of desulfurized slag
-106-
-------
FIGURE 6C
OPERATING COST/AIR PREHEAT TEMPERATURE VS
PERCENT LIMESTONE IN.FLUX*
o
o
4) O
>-i O
3 O
±J r-l
to
t-i
O)
Q,
01
H
4J
TO O
OJ O
-------
. . FIGURE 7C
OPERATING COST/AIR PREHEAT TEMPERATURE
VS SLAG BASICITY*
o
o
\o
O
o
0)
M
3
0) •
0,
E o
tn
O
CJ
60
d
•H
0)
0.
O
0 .2 .4 .6 .8
Slag Basicity, Lime/(Silica & Alumina)
* 14.57» Ash - 1% Moisture Coal, 2500°F Combustor Temperature
87» Sulfur in Slag, Limestone
1.0
-108-
-------
FIGURE 8 C
OPERATING COST/AIR PREHEAT TEMPERATURE VS
% SULFUR IN COMBUSTOR SLAG*
O
O
O
CM
O
O
CO
M
3
a.
S
at
H
4J
nj
0)
.C
0)
M
PH
M
•H
o
o
o
o
o
o
CM
O
O
o
Conventional Plant
vO
I
Operating
Range
4 6
Sulfur in Combustor Slag
vO
Di
EC
(0
r-l
r-{
•r-l
to
O
60
C
•r-l
U
TO
M
-------
must be recycled to the combustor and that both slag handling and slag
desulfurization capital equipment costs are increased. Also more coal
is required in the process to make up for the increased addition of
desulfurized slag to the combustor which is at a temperature considerably
lower than the combustor. In addition, more combustor offgas is
consumed in the air preheating operation, since higher air temperatures
are required (Figure 8C). This results in higher coal consumption to
yield the fixed energy requirement of a 1000 MW power station.
Summary of Operating Ranges for Important Process Parameters
The process simulation - economic results determined above can be
summarized as follows:
1. Low as well as high ash coals yield approximately
the same capital and operating costs for the process.
The air preheat temperature falls into the narrow range of
1000-1200°F for these coals. The process is, therefore,
adaptable to a varying input coal composition as expected
under normal power plant operation.
2. Coal moisture content should be as low as
practical (1 to 2 percent).
3. The combustor operating temperature should be
as low as possible (2500°F) consistent with a fluid
slag operation.
4. Limestone rather than lime should be used to
flux the coal ash—provided that the air preheat
temperature is reasonable.
5. The basicity of the slag should be as low as pos-
sible to minimize the limestone requirement. A slag
basicity of 0.1-0,2 should be used and based on
experimental studies yields operable slags.
6. The sulfur content of the slag should be
as high as possible. Slag sulfur contents of
about 6 to 8 percent are preferred.
-110-
-------
REFERENCES
1. Robson, F. L. , et.al., "Technological and Economic Feasibility of
Advanced Power Cycles and Methods", UARL Report J-970855-13,
National Air Pollution Control Administration, U. S. Department
of Health, Education, and Welfare, Durham, North Carolina 27701
Contract: CPA 22-69-114
2. Evaluation of Acid Mine Drainage Treatment Process, Environmental
Protection Agency, Water Quality Office, Contract No. 14-12-529
3. K. Endell & Co Wens, Industrial Heating 4, (2), 143, 1937
4. Fehling, H. R., "Erosion of Refractories by Coal Slag", Institute
of Fuel, J. Vol II No. 59, pp 451-458, June (1938)
5. Herty, C. H., et.al., Min. Met. Invest. Coop. Bull., No. 46 1,
(1930)
6. Humphreys, K. K., Lawrence, William F., Technical Report No. 53,
Coal Research Bureau, West Virginia University
7. Herty, C. H., Blast Furnace & Steel Plant, 25, 1000 (1937)
8. Herty, C. H. , et.al., Min. Met. Invest. Coop. BulL, No. 49, 1,
1930
9. Herty, C. H., Jr. Trans. Amer. Inst. Min & Met Eng. Iron & Steel
Div., 1929 ,pp 284-299
10. Rait, J. R. , Trans. Brit. Ceram. Soc. , ^0_, 157-204, 231-269 (1941)
11. Machin, J. S., Lee, T. B., Hanna, D. L., J. Am. Ceram. Soc., 35,
322-325 (1952)
12. The Making, Shaping, and Treating of Steel, p. 314, U. S. Steel
8th Edition (1964)
13. Panov, A. S., Kulikov, I. S. and Tayler, L. M., 1ZV, AKAD,
Nauk, SSSR, OED,'Tekhn, Nauk, Met Toplive, 1962 No. 3, p. 30
14. The Making, Shaping, and Treating of Steel, page 316, U. S.
Steel 8th Edition (1964)
15. Perry, J. H. "Chemical Engineers Handbook," McGraw Hill Book
Company, Inc. p. 226 (1950)
-111-
-------
16. Hougen, 0. A., Watson, K. M. , Ragatz, A. R. , "Chemical Process
Principles", Part I, pp 314, John Wiley & Sons, Inc. (1954).
17. Ibid, p. 482
18. Dallavalle, J. M., "Mic'romeritics,"'Fitment Publishing Corp.,
New York, pp 272, 334 (1948)
19. Gross, J., "Crushing and Grinding", U. S. Bur. Mines Bur. 402,
(1938)
20. Cameron, J., Gibbons, T. B., and Taylor, J., Journal of the Iron
and Steel Institute, pp 223-28, December (1966)
21. Brown, G. G., "Unit Operations," John Wiley and Sons, Inc., New
York, pp 42-45 (1950)
-112-
-------
|