EPA-R2-72-035
August, 1972
S02 FREE TWO-STAGE COAL COMBUSTION PROCESS
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
endorsement or recommendation 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
A. Bunas is the EPA project officer of the conti-
nuation of this work.
ii
-------
ABSTRACT
Under EPA sponsorship, (Contract No. CPA 70-146) Applied Technology
Corporation is developing a new SO -free Two-Stage Coal Combustion
Process. In this process, coal is dissolved by injection into molten
iron and simultaneously the dissolved coal carbon is oxidized with
injected air to carbon monoxide. Under the reducing conditions
existing during combustion, coal sulfur is not oxidized, but trans-
ferred from the iron to a slag consisting of coal ash and added lime-
stone. Thus, a sulfur-free offgas (carbon monoxide, hydrogen, and
nitrogen) is produced for use in power plant boilers.
Experimental combustor studies have shown that a boiler stack gas,
containing less than 50 ppm sulfur dioxide is produced. Results
indicate total coal solubility is obtained at injection depths
of less than three feet; and total oxygen utilization results at
air injection depths of 5 inches. Combustor design parameters were
established, and indicate that three 38-foot I.D. combustors are re-
quired for a 1000 MW plant. Laboratory work has shown the slags can
be successfully desulfurized with steam to produce elemental sulfur
and a H^S/SO« gas stream suitable for additional elemental sulfur
recovery in a Glaus Plant.
Process economics are favorable, indicating that a retrofitted instal-
lation would cost $23/KW. If the process is incorporated into a
grass roots, gas-fired power station, the capital cost increase would
be only $1.5/KW over a S0~-polluting coal-fired plant. Operating
costs for this new facility would be comparable to polluting plants
and considerably less if S0_ abatement equipment is included in the
polluting coal-fired facility.
This report was submitted in partial fulfillment of Contract No.
CPA 70-146 under the sponsorship of the Environmental Protection
Agency.
iii
-------
CONTENTS
Section
I Conclusions
II Recommendations
III Introduction
IV Process Description
V Experimental Program
VI Discussion of Results
VII Combustor Design Basis
VIII Economic Evaluation
IX References
Page
1
3
5
7
13
17
41
45
53
-------
FIGURES
PAGE
1 CONCEPTUAL DESIGN OF COMBUSTOR 8
2 TWO-STAGE COAL GASIFICATION PROCESS 10
3 CARBURIZATION RATE AS A FUNCTION OF IRON BATH CARBON CONTENT 18
4 EFFECT OF LANCE POSITION ON COAL SOLUBILITY 19
5 EFFECT OF COAL SURFACE AREA ON SOLUBILITY 20
6 EFFECT OF LANCE POSITION ON OXYGEN UTILIZATION 24
7 VARIATION OF METAL CARBON CONTENT WITH AIR INJECTION TIME 26
8 DECARBURIZATION RATE AS A FUNCTION OF OPERATION VARIABLES 27
9 EFFECT OF BATH CARBON ON S02 EMISSION 28
10 EFFECT OF LANCE POSITION ON S02 EMISSION 28
11 EFFECT OF IRON BATH CARBON CONTENT & LANCE POSITION ON
S02 EMISSION 29
12 EFFECT OF PROCESS PARAMETERS ON SLAG DESULFURIZATION 37
13 VARIATION OF OFFGAS COMPOSITION WITH TIME AT 2000°F 38
14 COMBUSTOR DIAMETER VERSUS OFFGAS VELOCITY 42
15 PROCESS FLOW DIAGRAM 46
vi
-------
TABLES
No. Page
1 Typical Combustor Slag/Iron Compositions 32
2 Stream Quantities for Two-Stage Coal Combustion Process 45
3 Estimated Fixed Capital Requirement Two-Stage Coal
Combustion Process 47
4 Estimated Capital Requirements for 1000 MW Power
Plant 50
5 Estimated Operating Cost 51
6 Effect of Coal Composition on Operating & Economic
Data 52
vii
-------
CONCLUSIONS
Analysis of the results obtained to date leads to the following con-
clusions about the Two-Stage Coal Combustion Process:
1. Boiler stack gas containing less than 50 ppm
S0« can be attained, regardless of coal
sulfur content.
2. Combustor design parameters have been
established: three 38-ft diameter com-
bustors can comfortably fuel a 1,000
megawatt power plant. For complete dis-
solution, the coal should be injected
3 feet under the molten iron surface; for
complete utilization, the air should be
injected 5 inches below the molten iron
surface.
3. Combustor slags can be desulfurized with
steam to produce elemental sulfur and a
mixture of l^S and S02 suitable for elemental
sulfur recovery in a Glaus plant.
4. The capital cost for the additional
equipment required for this process is
approximately $23 per kilowatt. In a new
installation, equipment savings result in
a capital costcf only $1.5 per kilowatt
above the cost of a S02 polluting conven-
tional coal fired plant. Operating costs
for such a new facility would be comparable
to those of polluting conventional plants.
-------
RECOMMENDATIONS
The work described in this report provides the basic carburization and
decarburization rate information required for combustor design. The
results of this work also show that sulfur dioxide emissions from the
power plant stack will be extremely low (less than 50 ppm).
The successful conclusion of the above basic studies has led to additional
experimental studies. A study of safe, non-water cooled, lances for
coal and air injection into the combustor is underway. Also planned
as part of the current program is a series of more detailed studies on
selected key areas. These include (1) simultaneous carburization and
decarburization of the iron bath to more closely approximate commercial
operation, (2) improved measurement of particulate emissions, (3) pro-
longed "endurance" runs to obtain more reliable refractory life estimates,
(4) rapid desulfurization of high temperature liquid slag, (5) deter-
mination of sulfur bearing slag weathering characteristics, (6) ex-
ploration of trace element accumulation in the iron bath, (7) review of
the data by design engineers to determine indicated additions or changes
in the experimental program, and (8) boilermaker-oriented study of com-
bustor offgas utilization.
Essentially, the work completed and planned for the experimental phase
will provide all the answers that can be obtained from the scale of
equipment (28 inch diameter combustor) that has been used to date.
Since the data are most encouraging, both from the technical and from
the economic feasibility point of view, a pilot-demonstration combustor
operation is indicated as the next step. This installation should be
operated in conjunction with a 50 megawatt to 150 megawatt boiler.
An estimate of the cost of a 30-month program, including 9 to 12 months
of continuous operation of the combustor providing all of the fuel for
a commercial boiler, has been prepared. The cost of the project will
depend on the size of the installation, and will vary from $5.5 million
for a 50 megawatt unit to $13 million for a 150 megawatt unit.
-------
INTRODUCTION
Because of the large quantities emitted into the atmosphere, sulfur
dioxide is regarded as the most serious gaseous air pollutant and the
need to eliminate it from power plant flue gases has been well documented.
To minimize sulfur dioxide emissions, some power plants may operate
on natural gas or low-sulfur fuel oil. However, the high cost and
limited availability of these fuels necessitate the power industry's
dependence on coal as its primary fuel.
A few power plants are located near western low sulfur coal deposits.
However, the high cost of transportation of low sulfur coal to the
industrialized east precludes the use of such coal on a large scale.
Consequently, considerable money and research efforts have been expended
to develop a practical method for high-sulfur coal combustion without
sulfur dioxide emissions from power plant stacks.
In general, the problem has been approached from two directions, i.e.,
the removal of sulfur prior to combustion and the removal of sulfur
dioxide from the flue gases after the sulfur bearing coal has been
burned. The precombustion sulfur removal processes are either very
expensive (de-ashed coal) or are ineffective (beneficiation and washing
techniques) for lowering coal sulfur content to desired levels. On the
other end of the scale, the treatment of flue gases to remove sulfur
dioxide has thus far proven to be rather expensive or inoperable. These
processes, such as scrubbing,. limestone and/or dolomite injection, or
sorption techniques inherently become expensive because of the dilute
sulfur dioxide concentration and the large volume of gases that must be
treated.
For these reasons, the Applied Technology Corporation (ATC) has
developed a process for removing sulfur (while it is still in concen-
trated form) during the actual combustion phase. Thus, unlike other
processes, the ATC coal gasifier prevents the sulfur from oxidizing
and removes it during the coal gasification phase.
Essentially, the ATC process consists of injecting coal particles into
a molten bath of iron. Because iron in the liquid state has a
tremendous affinity for sulfur and carbon, the coal solubilizes to re-
lease organic and inorganic sulfur constituents for reaction with the
active iron melt. Iron sulfides are formed which then migrate to a lime
containing slag floating on the molten iron bath (much like a steelmaking
practice) where they are removed from the combustion process. At the
same time, the carbon that is dissolved in the molten iron is reacted with
.air to produce an offgas which consists of nitrogen, carbon monoxide,
and hydrogen. This hot (above 2500°F) gaseous mixture, essentially
free of sulfur dioxide, is introduced into a steam boiler along with
secondary air to recover all of the heating value of the coal. Sulfur
is recovered in elemental form from the slag produced in the ATC process,
along with iron contained from the coal pyrites and a desulfurized slag
-------
road aggregate. All three by-products have commercial value.
The Office of Air Programs of the Environmental Protection Agency is
sponsoring (Contract No. CPA 70-146) a study in which ATC has been
obtaining experimental data from a coal gasifier of about 2 megawatt
capacity. The results of this work have been extremely encouraging
in that the use of molten iron to solubilize coal and then burn it
produces a hot offgas essentially sulfur-free that is suitable for
power plant boilers or gas turbines.
This report presents a description of the process, a discussion of the
state of the art, a program for additional work required to demonstrate
the process on a commercial scale, and the cost of the proposed demon-
stration work.
-------
PROCESS DESCRIPTION
Combustor Concept
What for years has been a problem in the steel industry can be a boon
for the energy industries. The strong affinity that exists between iron
and sulfur making it difficult to produce low-sulfur iron and steel
is the basis of the combustion-gasification concept. Instead of combusting
or gasifying coal by conventional techniques which convert the coal
sulfur into sulfur dioxide, in this new method, coal is injected into
a mass of molten iron. Both the fixed carbon and sulfur contained in
the coal are retained by the iron. That is, the fixed carbon content
of the coal sulfur are solubilized in the iron, while the coal volatiles
are cracked and exit in the offgas. This is termed carburization and is
represented as follows:
(Fixed Carbon) -- 3*~ (Carbon)
Coal Iron
dissolves
(Organic and Inorganic Sulfur) - ^(Sulfur)
Coal Iron
. .
(Volatiles) - »-(CO, 0*4, H2)
Coal Offgas
The dissolved carbon in the iron is then gasified by reaction with air
or oxygen to yield carbon monoxide. This is termed decarburization and
is represented as follows: •
(Carbon) + (1/2 02) - a— (CO)
Iron Air Offgas
By equalizing the carburization and decarburization rates, defined as
the weight percent change of carbon in iron per minute, a steady state
carbon level in the bath can be maintained. Thus, the combustor-
gasification principle involves the simultaneous carburization with
coal and decarburization with air of a molten iron bath to produce a
sulfur-free carbon monoxide rich offgas.
The dissolved coal sulfur in the iron migrates to a lime-bearing slag
layer floating on the molten iron to form calcium sulfide as follows:
(Sulfur) + (CaO) - »~(CaS) + (1/2 02)
Iron Slag Slag Iron
i
i
Thus, sulfur is continuously removed from the iron into a slag. The
slag consists of lime, silica, alumina, and calcium sulfide. Lime-
stone is continually added to the system, whereas, silica and alumina
occur in the coal ash:
-------
Coal Lance
Air Lance
oo
FIGURE 1-CONCEPTUAL DESIGN OF COMBUSTOR
-------
(Ash) ». (Silica and Alumina)
Coal . Slag
The calcium sulfide bearing slag is continuously removed and desulfurized
with steam to yield a sulfur bearing offgas and desulfurized slag:
(CaS) + Steam s^-(CaO) + (S, H2S, SC>2)
Slag Desulfurized Offgas
Slag
A portion of the desulfurized slag is recycled to the iron to utilize
its lime content and the rest is removed from the system. The sulfur
from the desulfurization offgas is condensed out and the H2S - S0£ gas
mixture is sent to a Glaus plant for sulfur recovery.
The ATC coal gasification method produces a sulfur-free, carbon monoxide
rich offgas, marketable sulfur, and coal ash-bearing desulfurized slag.
The iron normally found in coal accumulates in the bath, which is period-
ically trapped for by-product iron credit. No pollution causing by-
products are formed and, because desulfurized lime bearing slag is re-
cycled, considerably less than the stoichiometric requirement of lime-
stone (for sulfur removal) is required. Also, a number of carbonaceous
materials, from petroleum coke to sub-bituminous coals, with a wide range
of sulfur and ash contents, can be gasified.
A conceptual drawing of a furnace structure for the submerged combustion
of coal which would be compatible with conventional boilers used in power
generating plants is shown in Figure 1. As shown, the combustor is a
cylindrical refractory-lined vessel which contains a molten iron bath
into which coal is continuously injected beneath the surface. Located
above the combustor is a transition section which connects the combustor
to a conventional boiler (not shown).
As indicated in Figure 1, the sulfur-bearing coal is injected into the
molten iron where it dissolves. The combustion of carbon from the coal
by addition of air near the iron bath surface produces a hot carbon-
monoxide rich offgas (2500-2700°F) which proceeds via a transition
section to the conventional boiler where it is further combusted (to
carbon dioxide) to yield additional heat.
The operation of the combustor requires limestone and/or lime-bearing
slag. Limestone is converted to lime in the combustor and forms a slag,
which has an affinity for sulfur and thus removes the sulfur from the
molten iron. The slag is removed from the combustor and sent to a
desulfurizer where sulfur is recovered. A portion of the slag is fed
back to the combustor to again remove sulfur from the iron.
The Overall Coal Gasification Process
Figure 2 shows a diagram of the entire process. Coal from storage is
crushed and pneumatically conveyed into the combustor through tubes
(lances). Recycled desulfurized slag and limestone are added to the
coal and proportioned to the coal lances. Air for combustion and pneumatic
-------
0
w
to
O
I
CO
O
M
O
o
o
i
•n
§
n
w
CO
CO
Coal
Slag Limestone Compressor
i
60
1
1
Sulfur
Condenser
V
Air
Compressor
Iron
Granulator
Slag
Desulfurizer
Steam
Granulated Iron
Desulfurized Slag
To Crusher
-------
conveying of coal is supplied by separate compressors, because the large
volume of combustion air is required at a lower pressure (5 psig). The
carbon-monoxide-rich offgas generated in the combustor by the combustion
of coal enters the boiler where it is oxidized to carbon dioxide by
secondary air to produce steam for subsequent power generation.
The sulfur bearing slag formed in the combustor is separated from the
molten iron and transported to the desulfurizer, via enclosed metal
conveyors. Prior to entering the slag desulfurizer, the slag is crushed.
In the slag desulfurizer, the sulfur-bearing slag is contacted with steam
to produce a sulfur-rich gas which proceeds to a sulfur condenser and
Glaus unit where the sulfur is removed from the gas and deposited into
pits. Part of the desulfurized slag is sent to the combustor for reuse
and the rest is sold as road aggregate.
Because coal contains iron (mainly as pyrite) some iron is produced in
the combustor. Iron is periodically removed from the combustor and sent
to a granulator where it is contacted with a water spray to produce
salable granulated iron.
11
-------
EXPERIMENTAL PROGRAM
Purpose
Although most parts of the Two-Stage Coal Combustion Process are
commercially used in steelmaking practice, the combination of this
technology applied to the sulfur dioxide elimination problem is
unique. Consequently, an experimental program was undertaken under
Office of Air Programs sponsorship to study the application of a
molten iron bath as a coal gasifier for conventional power plant
boilers (Contract No. CPA 70-146).
To define the technical and economic feasibility of the process,
it was necessary to provide answers to the following questions:
1. What combustor depth and cross-sectional area is
required for a given power plant size?
In particular, it is necessary to determine how
deep into the metal bath each type and size of
coal must be injected to insure that it dissolves.
It should be noted that any part of the coal that
does not dissolve will be carried with the offgas
and probably burn to produce sulfur-bearing gases.
It is, therefore, important to completely dissolve
the coal in the iron bath. Obviously, some coal
losses can be sustained without upsetting air
pollution code requirements. With regard to the
oxidation of the dissolved carbon, it is necessary
to determine the optimum point of 'release of the
air that oxidizes the carbon. Optimization is a
matter of trade-off between the compression cost
associated with injection of air deep into the
metal bath and the loss of air utilization ef-
ficiency when the air is injected without in-
timate contact with the iron. Further, because
of the large amount of gas generated by carbon
oxidation, it is necessary to establish experi-
mentally the allowable metal and slag turbulence
(a function of gas velocity) in order to deter-
mine the combustor cross-sectional area require-
ment .
2. How clean is the gas produced in the combustor?
It is necessary to determine the operating tech-
niques that will eliminate, or at least minimize,
the sulfur dioxide and particulate matter content
of the gas produced. It is also desirable to
13
-------
determine to what extent two stage coal combustion
minimizes the formation of nitrogen oxides.
3. Are there any combustor materials or design pro-
blems?
If the combustor is to operate like other power
plant equipment over prolonged periods, approximating
one year between maintenance shutdowns, it is nec-
essary to establish experimentally that refractory
wear will not present a serious problem. It is also
necessary to show that reliable, and safe means of
injecting coal into the iron bath can be successfully
designed.
4. What slag composition should be used in the
combustor?
Since limestone must be added to the combustor
to flux the coal ash and react with dissolved
iron sulfur, process economics are favored by
minimizing the quantity of slag and the lime
content of the slag. Experimental data are
needed to determine slag quantity and composition
that are consistent with (1) a high sulfur re-
tention capacity, (2) slag fluidity that is suf-
ficient to permit continuous removal of slag
from the combustor, (3) a minimum slag reactivity
with commercially available refractory, and (4)
amenability to slag desulfurization.
5. How can sulfur be best recovered from the slag?
Previous research projects (EPA Contract No.
14-12-529 and No. 14-12-929) have shown that
sulfur is readily recoverable from slag by re-
acting the slag with steam at elevated tempera-
tures. However, the rate of reaction depends
on the physical properties of the slag. Since
the physical properties of the slag depend on
composition, kinetic data are required on the
desulfurization of slags of the composition
contemplated for the Two-Stage Coal Combustion
Process.
Method
The experimental program consisted of coal gasification-combustion
studies in a combustor containing up to three tons of molten iron,
bench-scale studies of slag properties and sulfur recovery kinetics
and a study of the offgas composition as a function of coal sulfur
14
-------
concent. An engineering evaluation of the technical and economic
aspects of the process was also made. No bench scale studies were
made of the coal solubility and gasification reactions which occur
in the combustor. It was felt that the 28-inch diameter, three-ton
electric furnace used was the smallest vessel in which these reactions
could be studied without excessive freezing of the iron and slag,
which would obscure the effect of the parameters under study. In
retrospect, it appears that even the large electric furnace used was
somewhat small for the purpose. The details for the experimental
work are described in the Appendix (presented in a separate volume).
While the combustor was being designed and installed, bench scale
laboratory studies of the physical-chemical characteristics of slags
which may be encountered in the combustor were carried out to pro-
vide guidance for the combustor experiments. The most important
aspects of slag characterization were concerned with the search for
slag compositions that would be generated by adding limestone in
the combustor to flux the silica and alumina from coal ash and to
react with the sulfur to produce calcium sulfide. Compositions of
such slags were explored to determine what lime content would be
required to produce slags that would be fluid at combustor operating
temperatures and would contain sufficient free lime to provide a
driving force for sulfur transfer from the iron to the slag. In
addition, studies were made of such slag properties as heat capacity
and crushing energy requirements. These studies were followed by
studies on slag granulation and slag desulfurization, which were
continued to parallel the combustor operations.
Experimental data have been obtained on a combustor capable of
gasifying coal at a rate equivalent to supplying the energy requirement
for a 2 megawatt power plant. An induction melting furnace was selected
as the combustor vessel because it is a convenient means of preparing
a molten iron bath and maintaining it at a specified operating temperature.
Experimentation was divided into alternate carburization and de-
carburization runs. That is, coal was injected into a molten iron bath
(2600-2800°F) to bring the carbon level to about 3 to 4 percent and
then the bath was decarburized back to a carbon level of about 2 to 3
percent. During experimentation, a continuous offgas sampling system
was used to monitor the SO-, NO , NO, H , CO, 0-, and CO content of
the offgas. Bomb gas samples were also 'taken and analyzed in a gas
chromatograph to confirm the continuous analysis and to detect nitrogen
in the gas. A detailed description.of experimental equipment and pro-
cedures is presented in the Appendix.
The primary experimental objectives were to determine, under various
experimental conditions:
1. Carburization and decarburization rates.
15
-------
2. Carburization efficiency or the fraction of the coal
fixed carbon which actually increases the carbon level
of the iron bath.
3. Decarburization efficiency, or the extent to which all
the oxygen in the air is reacted with the carbon dis-
solved in the molten iron.
4. SO- content of the offgas during carburization-decarburization.
5. The sulfur partition ratio, defined as the ratio of the
percent sulfur in the slag to the percent sulfur in the
iron.
16
-------
DISCUSSION OF RESULTS
This section presents a summary of the results of the experimental study
on the Two-Stage Coal Combustion Process. The results are discussed
by subject of study, such as slag characterization, carburization, and
decarburization of molten iron. Stress is placed on the significance of
these results on the design and economics of the Two-Stage Coal Combustion
Process.
Slag Characterization
The search for suitable combustor slag compositions (slags that would
be fluid in the combustor and would be capable of removing sulfur from
the iron) is described in detail in the Appendix. The results of this
search show that slags with good fluidity and high affinity for sulfur
can be produced with very little excess lime--in the order of 10 to 15
percent of the slag weight. A slag composition that was used exten-
sively in the combustor consisted of:
22 percent alumina
45 percent silica
13 percent lime, and
20 percent calcium sulfide.
This slag was used on the basis of bench-scale results that show that
it has adequate fluidity (viscosity of approximately 50 poise at 2700°F).
This slag also showed strong affinity for sulfur in bench-scale studies,
with an equilibrium sulfur partition ratio of 20 (ratio of percent sulfur
in the slag.to percent sulfur in iron in contact .with the slag).
Carburization
Early in the experimental work, it was established that the solution
rate of carbon in the iron bath is a function of the carbon content
of the bath. Figure 3 shows how the carbon solubility decreases as the
carbon content of the iron increases and approaches the 5 percent sat-
uration level. The linear relationship between the rate of solution
and the carbon content of the iron suggests a carbon diffusion situation
represented by the equation:
dCt ka(Cs-Ct-)
dt v
where Ct = carbon concentration in iron at time t
Cs = carbon concentration at saturation
k = mass transfer coefficient
a = particle surface area
v = volume of metal bath
The data of Figure 3 were gathered by injection of a carbonaceous
material (graphite) under various conditions. The differences in the
slope of the lines in Figure 3 are caused by variations in the operating
parameters, such as graphite feed rate, pneumatic transport gas flow rate,
17
-------
1.0
3
•rj
C
o
CO
CJ
c
0)
o
M
-------
100-.
90-.
4J
g 80.
M
O)
70--
60..
o
dJ
•H
M-t
W
c 50.
o
0)
40..
•d
0)
30--
20-.
10
Coarse (1/4 inch)
Average' Carbon Content
of Iron Bath
A. 2.2 to 2.8
• 2.9 to 3.0
X 3.1 to 3.3
3.4 to 3.5
8
10
12
14
16
18
20
22
24
26
Lance Position - Inches Below Surface
FIGURE 4-EFFECT OF LANCE POSITION ON COAL SOLUBILITY
-------
c
0)
o
M
0)
P-i
u
c
0)
w
c
o
.0
0)
•H
(K
110
105
100
95
90
85
80
f
100 200 300 400 500 600 700 800
2
External Surface Area of Coal - cm /gin
FIGURE 5-EFFECT OF COAL SURFACE AREA ON SOLUBILITY
20
-------
and depth of graphite injection. All these variables affect the total
particle surface area that is present in the metal bath at any given
time ("a" in the above equation).
Efforts are continuing to quantify the effect of these variables in
order to normalize the data in a single plot of solution rate vs. con-
centration. When this is accomplished, the strong effect of carbon
content will be accounted for, and other variables can be studied more
precisely. Until that time, data will be treated by grouping them
according to iron bath carbon content.
This approach was taken in a study to determine how deep under the iron
surface the coal must be injected to insure complete solution. Figure 4
shows the effect of lance position on fixed carbon efficiency. The
latter quantity is the ratio of the weight of carbon dissolved in the
iron to the weight of fixed carbon added with the coal. Only the
fixed carbon is considered, because most of the volatile matter carbon
is cracked to lighter constituents, carbon and hydrogen, and bubbles
through the iron too rapidly to dissolve. However, some of the volatile
matter cracked to carbon does dissolve and yields efficiencies (Figure 4)
in excess of 100 percent. Figure 4 also is in essential agreement with
the previously discussed effect of iron bath carbon content on coal
solubility rate. As seen, the carbon efficiency tends to be higher at a
lower iron bath carbon content, when the lance position is constant.
One data point on Figure 4 marked "coarse", corresponds to a lower
efficiency because the coarser coal had a lower external surface area
through which the carbon could diffuse into the iron. The effect is
shown clearer in Figure 5 in which the fixed carbon efficiency is plotted
as a function of the external surface area of three different size
fractions of the same coal. All data points were obtained by injection
of coal at the same depth (21 inches) and at approximately the same
iron bath carbon content (3.05 to 3.37 percent). It should be noted
that the surface area of the coal changes drastically after injection
into the molten iron bath. Although it is not possible to measure
these changes, it is reasonable to assume that the rapid evolution of
volatile matter causes the coal particles to expand and assume a structure
similar to popcorn, as happens during the processing of coal char.
Figure 5 indicates that the surface areas of the coals tested maintained
their relative ranking after injection into the molten iron.
Decarburization
In pneumatic oxygen steelmaking processes such as the basic oxygen
furnace, Kaldo furnace, or the Rotor furnace , decarburization of pig
iron to produce steel is directly proportional to the rate at which
oxygen is injected into the molten metal bath. In particular, for
those processes (Kaldo, Rotor) where mechanical as well as gas agitation
2 1
-------
a-re employed, the rate of decarburization is a linear function of time.
For example, in the Kaldo process, constant decarburization rates of
about 0.2 percent carbon per minute are realized over a metal bath
carbon composition of 3.6 to 1.6 percent carbon. In the basic oxygen
process which relies solely on pneumatic stirring (and, consequently,
does not have the agitation associated with the other processes) es-
sentially constant decarburization rates are achieved over a narrower
range of iron carbon content. Thus, commercial steelmaking process
data show that the decarburization rate is controlled by the gas phase
mass transfer of oxygen to the liquid metal bath. Oxygen utilization
efficiencies for such processes are in the range of 90 to 95 percent.
Assuming that the chemical reaction between dissolved carbon and
incoming oxygen is rapid, a simple material balance relating the change
in percent carbon per unit time in the metal bath is given by
W^ = -kE(P)V
where W = weight of iron
• c = percent carbon content of the molten iron
t = time
k = mass transfer coefficient
V = air injection rate
E(P) = efficiency of oxygen utilization as a function of injection
lance depth and other parameters
This equation simply states that the depletion of carbon in the metal
bath is solely a function of the injection air rate. The efficiency
E(P) is a function of a number of variables including the gas velocity
leaving the lance, the immersion depth of the lance into the metal, the
ratio of lance diameter to molten bath diameter and the ratio of lance
immersion depth to the total molten iron bath height. Theoretical cor-
relations are not available to estimate the effect of these variables
on the mass transfer coefficient and the interfacial area generated by
the kinetic energy of the gases impinging on the molten metal. However,
it is possible to determine the functional dependence experimentally.
As will be shown later, for the experimentation conducted in this work,
the efficiency of oxygen utilization for the combustion of carbon is
related primarily to the immersion depth of the lance.
Because of the rapidity of the combustion reaction for converting dis-
solved carbon to carbon monoxide, the mathematical interpretation pre-
sents the simple conclusion that the decarburization rate is equal to
the rate at which oxygen is made available to the molten bath. Obviously,
for a constant oxygen transfer, the decarburization rate will be constant
.and the carbon content of the metal bath will decrease linearly with
time.
A series of experiments were completed to determine the relationship
between decarburization rate and oxygen utilization efficiency and the
22
-------
following variables: iron bath weight, lance immersion, air injection
rate, ratio of lance immersion depth to metal bath height, and velocity
of injection gas leaving the lance. The weight of the iron bath was
varied between 2,000 and 5,000 Ibs., air flow rate from 70 to 250 SCFM,
and lance position from a high of about 4 inches above the metal surface
to an immersion depth of 19 inches. Substantial scatter in the data
exists and obscured the effect of minor variables. However, the oxygen
utilization efficiency did correlate well with the lance immersion
depth. The results of this work are presented in Figure 6.
As can be seen in Figure 6, the lance position relative to the slag
metal interface profoundly influences oxygen utilization efficiency.
When the position of the air injection lance tip is about 4 inches
above the surface of the molten iron bath, oxygen utilization is about
25 percent. When the lance tip is at the metal bath surface, the
oxygen utilization is attributed solely to the creation of interfacial
area between the gas and the molten iron generated by the air leaving
the lance. Obviously as the tip of the lance approaches closer to the
iron bath surface, the gas more efficiently creates interfacial area
between the air and the molten iron to bring about a higher mass transfer
rate of oxygen. When the lance is immersed into the molten iron at
a depth of about 4 to 5 inches, oxygen utilization efficiency approaches
100 percent. Any further increase in immersion depth has no further
effect upon efficiency and only results in a higher pressure drop
requirement to overcome the ferrostatic head. Since the objective
of the experimentation was to define or evaluate expected operation
for a commercial vessel, the bulk of the experimental work was confined
to lance submersion depth of less than 6 inches. In this way, compression
cost for the injection of combustion air would be kept minimial.
As is evident from Figure 6, considerable data scatter exists for oxygen
utilization efficiency. For example at an immersion depth of about 4
inches the measured efficiency varied from a low of 93 percent to a
high of 108 percent. The variation in this result is attributed to
the splashing that occurs within the vessel. It has been estimated
that splashing can deposit about 100 to 200 pounds of iron on the walls
of the vessel. For a three percent carbon melt containing 3,000 pounds
of iron, this can introduce an error of about 3 to 6 percent in the
efficiency calculation. Inasmuch as the oxygen utilization efficiency
is calculated from the bath weight and the difference in carbon content
of the molten iron at the beginning and end of an experiment, the un-
known amount of splashing gives rise to errors of this magnitude.
It is of interest to point out an implied effect of lance immersion
depth and bath weight on oxygen utilization efficiency. In an effort
to achieve high decarburization rates, shallow molten iron bath (2,000
Ibs, 15 inches deep) and high air injection rates (200 SCFM) were used.
When the lance was immersed to the full depth of the iron bath, con-
siderable splashing and lifting of the molten iron were observed. This
gave rise to a low oxygen utilization efficiency as illustrated in
Figure 6. This test suggests that a deep immersion depth in a shallow
23
-------
O
0'
'
M-l'
w-
4J-
cd
N
g
00
>,
X
o
100
75
50
25
0
High Air Flow
Severe Splashing
Iron Bath-lb.
O 2000
• 3000
O 4000
A 5000
-20 -10 0 +10
Lance Position-Inches Above (+) or Below (-) Iron Surface
FIGURE 6-EFFECT OF LANCE POSITION ON OXYGEN UTILIZATION
24
-------
bath (and for the geometry of the system used in this work) considerable
bed eruption can occur to lower oxygen utilization efficiency. However,
the latter observation has little commercial interest and is discussed
only to explain experimental results. Inasmuch as a commercial combustor
will operate at minimum air injection depth commensurate with high
oxygen utilization efficiencies to keep compression costs down, lance
submersion depth will be maintained at about 5 inches and bath eruption
will not occur.
The data exhibited too much scatter to correlate decarburization
efficiency with gas velocity at constant lance immersion depth.
Nevertheless, it is believed that such an effect should prove to be
substantial. For example in a basic oxygen furnace (EOF), oxygen
injection is carried out at supersonic lance tip velocities. Normally,
the lance is located several feet above the surface of the molten iron.
Even so, at these high velocities, sufficient oxygen transfer exists
to generate oxygen utilization efficiencies of about 90 to 95 percent.
To determine the effect of carbon concentration on decarburization
rates, a number- of experiments were completed in which molten iron
baths of varying carbon content and weight were decarburized at dif-
ferent flow rates. Typical data for the variation of hot metal car-
bon content with air injection time are presented in Figure 7. The
data show that the decarburization rate is a constant and that the per-
cent carbon in the molten iron varies lineraly with time. This tends
to support the conclusion that the decarburization rate is independent
of carbon ceontent and is primarily a function of the transfer of oxygen
from the gas phase to the liquid metal.
If the oxygen transfer were 100 percent efficient, the slopes of each
of the decarburization curves would be directly proportional to the
air input rate. Because of inefficiencies generated by lance position
and other variables, not all of the data are amenable to this simple
translation. However, by incorporating into the rate equation, the
efficiency of oxygen utilization correlation presented earlier, it is
possible to normalize all the rate data so that the decarburization
rate can be presented as a simple function of efficiency, air injection
rate, and bath metal weight. The normalized data are presented in
Figure 8. As expected, the data show a high degree of correlation
and indicate that the decarburization rate increases with increasing
air injection rate and decreasing metal bath weight.
Because of the limitations of the experimental combustor (i.e., small
size), high decarburization rates were obtained when the combustor
contained the least amount of metal. This is to be expected since,
for a given air flow rate, high decarburization rates are achieved by
decreasing the amount of metal contained within the combustor. However,
for the same geometrical reasons, these data tend to exhibit the
25
-------
B
-------
.16
.12
c
•H
e
c
0)
o
M
.08
.04
2468
IQOVE(P)
W
FIGURE 8-DECARBURIZATION RATE AS A FUNCTION
OF OPERATING VARIABLES
10
27
-------
N>
oo
s
-------
27. Carbon
e
Q.
0.
CM
O
CO
u
Q
CO
-202468
Lance Position, Inches
FIGURE-U.-EFFECT OF IRON BATH CARBON CONTENT AND
LANCE POSITION ON S0? EMISSION
29
-------
greatest amount of scatter as shown on Figure 8. Physical size alons
limits the amount of air that it is possible to inject within the
experimental combustor. It is of interest to point out that commercial
Bessemer converters readily obtain decarburization rates of 0.25 per-
cent carbon per minute and higher . Consequently, we believe that
commercially acceptable decarburization rates not only have been
demonstrated in the samll experimental unit but are supported by data
available on a commercial scale from the steel industry.
SOn NO and Particulate Emissions
The S0_ content of the combustor offgas is determined continuously
during all data runs. This value is averaged over each run and con-
verted to an equivalent stack SO,, content which would result in a com-
mercial power plant burning the combustor offgas in a boiler. Figure 9
shows a plot of the equivalent stack gas S0? content during decarburization.
These results are for experiments performed under various experimental
conditions. Consequently, there is much scatter but the trend is obvious.
The S0_ content of the stack offgas decreases as the carbon content of
the iron increases. One factor of particular commercial importance
should be stressed. The data show that S02 levels of 50 ppm or less
can be achieved by maintaining a minimum tnree percent carbon level in
the combustor. It is expected that in commercial use, the combustor
will operate at a carbon level between three and four percent.
A prime operating variable affecting S09 generation is the position of
the lance relative to the slag-metal interface. The lower the lance is
into the metal, the higher will be the decarburization efficiency and
less unused oxygen will be available to form sulfur dioxide by reaction
with the CaS content of the slag. This is shown in Figure 10 which is
a plot of the equivalent stack gas SO- content vs. lance position
during decarburization. As seen, the S0_ content decreases from about
200 ppm with the lance in the slag (nine inches above the slag-metal
interface) to about twelve ppm with the lance in the iron, four inches
below the slag-metal interface. The results of Figure 10 include
experiments in which the carbon content of the iron varied between 2.5
and 3.75 percent carbon content. Since both carbon level in the molten
iron and lance position affect SO™ production, the S0» emission data
were correlated with regard to both variables. Figure 11 presents the
smoothed data obtained from cross-correlation of these variables. The
data scattered somewhat, but conservative relationships are presented.
This generalized plot shows that S0_ emission decreases with increasing
carbon content and increasing lance immersion depth. This means that
in a commercial combustor operating at 3 to 4 percent carbon with a
lance immersion depth of 4 inches into the iron, less than 50 ppm S0_
will be found in the stack gas.
30
-------
The concentration of NO and NO in the combustor offgas was also
measured continuously and averaged for each run. In general, the con-
centration of these gases remained below 50 ppm. These low levels of
nitrogen oxides (NO ) are to be expected in the combustor's reducing
atmosphere. However, it must be pointed out the secondary combustion
of the combustor offgas in a boiler will occur under oxidizing con-
ditions, and additional NO formation is to be expected. The NO
generation in the boiler will depend on flame temperatures, excess air
used, and residence time in the boiler. It remains to be seen whether
new boilers can be designed and existing boilers can be modified to
take advantage of the gaseous fuels generated by the process to operate
at reduced temperatures, lower excess air, and at decreased residence
time.
Particulate emissions from a combustor operating at steady state can
be (1) undissolved coal particles that escape from the iron bath if the
coal is not injected deep enough, (2) iron oxide fume generated if the
carbon level in the bath should drop below 1.5 percent instead of the
3 to 4 percent design carbon level, and (3) iron and slag particles which
splashed out of the bath and are carried out of the combustor by high
velocity gases. The first two causes of particulate emissions can be
adequately controlled as discussed in conjunction with the results of
carburization and decarburization. That is, by injecting the coal
sufficiently deep, undissolved coal leaving the system is minimized; and
by using iron carbon levels of 3 to 4 percent, carbon will be preferentially
oxidized, thereby, minimizing iron oxide formation and carryover. How-
ever, reducing gas velocities to minimize particulate carryover in the
combustor offgas can affect economics of the process as well as its
practical application in existing plants by causing the combustor
area to become unwieldy.
For this reason, considerable effort has been expended in the attempt
to obtain meaningful and reliable data on particulate matter carry-
over from the combustor. Data gathering has been hampered by the
geometry of the experimental combustor which limits, of necessity, the
free-board of the metal bath and by the intermittent method of operation.
The low free-board height permitted rather heavy coal particles to be
carried out of the combustor and deposited on horizontal piping duct-
work during carburization runs, only to be picked up by the higher
velocity gases created during subsequent decarburization runs. The
intermittent method of operation made it difficult to obtain reasonable
in-flight samples and has forced the reliance on a total dust gathering
technique using a filter bag. This filter bag was located at the exit
of the offgas handling facility and was allowed to gather dust over the
duration of a run. The latter technique was rather recently installed
and is currently undergoing refinement, primarily to minimize the
weighing errors that are introduced by the large weight of the bag
relative to the small weight of particulates.
31
-------
TABLE 1
Typical Combustor Slag/Iron Compositions
Slag Iron
With
Agitation
Without
Agitation
CaO
15
17
16
15
17
12
18
14
15
14
A100
30
27
29
32
28
35
36
31
31
44
SiO
55
56
55
53
54
53
46
55
54
42
S
5.6
4.9
5.0
4.8
4.8
0.16
0.85
0.21
0.21
0.32
Basicity
0.17
0.21
0.19
0.18
0.21
0.14
0.22
0.16
0.18
0.16
S
0.48
0.48
0.48
0.48
0.48
0.84
0.55
1.46
0.40
1.30
Partition
Ratio
11.5
10.1
10.4
9.9
9.9
. 0.19
1.55
0.14
0.53
0.25
Notes:
1. CaO, SiO-, and Al-O. are on a CaS-free weight per-
cent basis.
2. Basicity is the weight ratio CaO/(SiO_ + Al-O-).
3. Partition ratio is defined as the ratio of the per-
cent sulfur in slag to percent sulfur in iron.
32
-------
Combustor Slag Studies
To expedite start-up instead of waiting for slag to form from limestone
additions and coal ash, combustor slags were produced by injecting a
dry mixture of slag components into the bath at the onset of a test
series. Laboratory studies indicated that a basicity weight ratio of
CaO to SiO plus Al 0 , of 0.2 or less would yield a fluid slag. Con-
sequently, most experimental combustor runs were made with slag which
had an initial basicity of 0.2. Carburization runs, in which coal ash
migrated to the slag, lowered the basicity. In general, slag basicities
were in the 0.1 to 0.2 range.
Combustor slag data generally fall into two categories; those collected
when the bath was agitated, maintaining intimate contact between iron
and slag, and those collected when no agitation was used. The bath
was well agitated during carburization and decarburization runs. How-
ever, it is impractical to maintain bath agitation in the between-runs
intervals which sometimes lasted for several hours. In the latter case,
the metal stirring caused by the induction furnace provided a limited
amount of agitation. A list of typical combustor slag compositions and
the iron in contact with the slag is shown in Table I.
When dry slag ingredients containing about 8 percent sulfur were added to
the system and with no subsequent agitation, the sulfur was retained
mostly by the iron. As seen in Table I, under these conditions, the iron
had up to about 1.5 percent sulfur and the slag from about 0.2 .to 1 per-
cent sulfur, yielding partition ratios of 0.2 to 1.6. These low
partition ratios result because without continuous agitation, the slags
tended to be cool relative to the metal—in the order of 2000' F as com-
pared to metal at 2600 F. In fact, the slags tended to crust over on
the surface and became extremely viscous because of radiant heat losses
from the slag to the cooler combustor transition section. The laboratory
study has shown that as the slag temperature decreases, the partition
ratio also decreases. For example, the equilibrium partition ratio
for a 0.2 basicity slag, decreased from about 170 at 2680 F to about
30 at 2490 F. Consequently, the low partition ratios obtained under
nonagitated conditions are expected and confirm laboratory results.
To provide agitation between the slag and iron and thereby approximate
a commercial operation, a small flow of nitrogen was continuously
bubbled into the iron to agitate the slag during several runs. As
expected, agitation resulted in increased mass transfer of the iron sul-
fur to the CaO-bearing slag to form CaS with a corresponding increase
in slag sulfur content and decrease in iron sulfur content. Agitation
also kept the slag hot at near the iron temperature and, as shown in
the laboratory study, higher partition ratios result at higher operating
temperatures.
With the slag and iron agitated in this manner, partition ratios in the
order of 10-12 were obtained (Table I). The iron contained about 0.5
percent sulfur and the slag about 4.8 to 5.6 percent sulfur. This con-
33
-------
firmed laboratory results in which partition ratios of approximately
10 to 20 were obtained for 0.1 and 0.2 basicity slags (see Appendix).
In the actual combustor operation, higher partition ratios are expected,
because agitation between slag and iron will be much more pronounced
than was obtained in the experimental combustor. These findings are
economically significant because it is desirable to have the slag
contain as much of the coal sulfur as possible. Sulfur remaining in
the iron is a by-product credit loss to the process.
Combustor slags containing as much as 12 percent sulfur were prepared.
However, these slags tended to be viscous and contained semi-solid slag
lumps. This is expected since the laboratory study showed that
increasing the CaS content of 0.2 basicity slags beyond 10 percent sul-
fur yielded extremely viscous slags (1000 poise).
In summary, the laboratory and experimental combustor slag studies
indicate that partition ratios of at least 10 can be expected under
actual combustor operating conditions.
Materials Problems
There are two main materials problems involved in combustor design and
operation. The obvious problem is that of refractories for the combustor
proper and the ductwork between the combustor and the boiler. Operating
at temperatures in the range of 2500 to 2700 F, the combustor and duct-
work will require some of the highest quality refractories available at
this time, but will not require any advance of the current state of
refractory art. Containment of the iron is not expected to present any
difficulty, provided that coal and air injection lances are located
so that coal and air impingement on the refractory bottom and sides
of the combustor is minimized. In this manner, the erosion effects
become insignificant. Refractory endurance with regard to containing
molten iron is backed up by blast furnace experiences where hearth
refractories last for several years.
The main problem is that of slag attack on refractories within the
cpmbustor. Refractory performance in the experimental combustor and
steelmaking experience indicate that refractory choice should emphasize
protection from slag attack. Since the experimental combustor has
operated with acid slags (slags in which the ratio of lime to silica
and alumina is relatively low), the combustor has been lined with acid
linings consisting primarily of alumina. The study of refractory
linings is not complete at this time, and quantitative data that can
be extrapolated to prolonged operation of a very large combustor cannot
be obtained without extremely expensive tests (which must include a study
.of vessel design as well as refractories). However, the experience
with the current brick lining of the combustor is very encouraging.
The current lining consists of a dense (porosity of 14 to 18 percent)
34
-------
hi-gh alumina (90 percent A1-0-) bricks. This lining has been in use
over four campaigns covering a total operating time of five weeks, and
is still useable. In between these campaigns, the lining has been patched
but the deepest penetration of slag attack on the brick has been well
under two inches with the average being less than one inch. It is
expected that the forthcoming refractory test will provide data of a
more quantitative nature. Even if no better refractory is found as a
result of the test, it is reasonable to expect that dense high-alumina
brick can be used in a refractory design that will permit combustor
operation in the order of one year's duration.
A materials and design problem which is now under intensive study is
that of lances for the injection of coal and air under the surface of
the iron in the combustor. As indicated in the preceeding discussions
of carburization and decarburization, it is necessary to release the
air and the coal under the surface of the iron. To accomplish this,
it is necessary to develop tubes (lances) that are capable of operating
continuously in contact with the molten iron and slag. Further, the
coal injection lances must be cooled to prevent excessive heat transfer
to the coal which would cause the coal to coke and possibly plug the
lance. Because accidental release of water under the iron surface
would release very large quantities of gases, creating a safety hazard,
water-cooled lances are not considered desirable.
Accordingly, an experimental study is now under way to design and
evaluate non-aqueous cooled submersible lance systems for both coal
and air injection. Oils, molten salts, and low-melting point alloys
are candidate cooling materials that can be used «in conjunction with
metal lances.
Slag Desulfurization Kinetics
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:
35
-------
where w = weight of the slag
p = density of the slag
C = sulfur content of the slag
tS= time
c..= stoichiometric conversion constant
k = mass transfer coefficient
o
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 generaP, 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
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 (2)
where D = diameter of the slag particle
G = flow rate of the fluid
Substituting Equation (2) into (1), re-arranging and integrating for a
constant flow system with slag particles having initial sulfur con-
centration C , the following relationship is obtained:
Ci
where C? = constant defines the sulfur level in the slag as a
function of reaction. time.
A detailed description and discussion of the experimental equipment and
procedure can be found in the Appendix and will not be discussed here.
The results of the experimental work—show that the slag desulfurization
36
-------
100
60
A
B
o
t-l
•a
ai
I
D
CO
0)
a
75
50
~ O
Synthetic Slags
® 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
a
Combustion
2100°F -40 +60
25
10
20
30
.40
50
^f
100
500
FIGURE 12-EFFECT OF PROCESS PARAMETERS- ON SLAG DESULFURIZATION
-------
e
0)
o
I
CJ
•o
0)
o
•H
•O
c
c
0)
o
• 3
CD
a)
Q
75
•H
CO
0)
00
50
o 25
M
0)
N
Synthetic Slag
O
Sulfur
SO
10
20
30
40
50
60
Time Minutes
FIGURE 13-VARIATION OF OFFGAS COMPOSITION WITH TIME AT 2000°F
38
-------
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.
Nevertheless, Figure 13 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 SO
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 for-
mation 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 produced
an offgas whose cumulative elemental sulfur content was 53 percent and
the cumulative H?S to S0_ ratio was 1.6 at the end of two hours.
Results thus far indicate that slag desulfurization is mass transfer
controlled. Fortunately, it presents no chemical reaction kinetics
limitations that would prohibit the design of reasonably sized reaction
vessels. Although complete design information is not yet available,
the engineering design must take into consideration tendencies for
slag sintering to occur. The slag must be simultaneous desulfurized and
rapidly cooled in the desulfurizer to below about 2000 F so that
sintering effects will not occur. Based on the tendency for the slag
particles to adhere to each other (as found in the experimental effort)
it is estimated that the time at temperatures about 2000 F in the
desulfurizer must be about 5 minutes or less to prevent sintering. Con-
sequently, the design of the upper portion of the desulfurization tower
will be limited by the rate at which the largest slag particle will
cool to the desired temperature range at a specified time. This implies
that for the production of a road aggregate material, the cooling rate
will determine the largest size lump that can be produced. Since the
desulfurization tower will be a continuous countercurrent moving packed
bed, simultaneous heat-mass transfer considerations will govern the
geometry of this system.
Obviously, design criteria should be based on the largest particle
diameter to be processed through this system. It is tentatively
estimated that depending upon the particle size requirement of the
end product, a residence time of about 30 to 150 minutes will be
required in the reactor.
39
-------
COMBUSTOR DESIGN BASIS
The size of the combustor(s) required in the Two-Stage Coal
Combustion Process depends on a number of process parameters such
as the temperature, pressure, and composition of the offgas leaving
the bath surface, and the coal composition which affects the combustion
air rate requirement, the coal rate into the process,, and the offgas
composition. Once these factors are specified, the volumetric offgas
flow rate and the coal rate are known or can be calculated.
The diameter of the combustor is controlled by one of two design
criteria. The superficial offgas velocity must be below a specified
value, tentatively set at 25-35 ft/sec.-based on basic oxygen furnace
technology, and there must be sufficient iron in the combustor to ac-
cept coal at an attainable solubility rate. The depth of iron has
been established at 3-4 feet to totally solubilize the coal. Con-
sequently, for any given combustor diameter and set of process con-
ditions, both the offgas velocity and coal solubility rate can be
established when the volumetric flow rate of offgas, the coal rate,
and carbon content of the coal have been determined.
In general, when air is used for combustion, the offgas velocity
criteria will control the diameter of the combustor. Figure 14 pre-
sents a plot of combustor diameter vs. offgas velocity (2500°F) for
a process consuming 361 TPH of coal to supply the energy requirements
of a 1000 MW power plant. The coal has a 12,500 BTU/lb. higher heating
value with the following ultimate analysis:
total carbon - 68.1
hydrogen" - 5.0
oxygen - 7.3
nitrogen - 1.5
sulfur - 3.6
ash - 14.5
2
For this process case, one-60 ft. diameter combustor (2820 ft ) will
be required to yield an offgas velocity of 25 ft/sec, requiring a
coal solubility rate of 0.11 percent C/min. However, to insure total
reliability, three-38 ft. diameter combustors would be used. Thus,
when one combustor is down, the other two would handle the full power
plant load. Three-38 ft. diameter combustors would have a surface
area of 3400 ft and two would have a surface area of 2270 ft . Under
a full power plant load-two combustor operation, the offgas velocity
would be increased to about 31 ft/sec and the coal solubility rate
would be about 0.18 percent C/min.—both values within specified
design criteria.
The lances used in the process carry either an air-coal mixture, pre-
heated combustion air, or a preheated combustion air-flux mixture,
into the combustor. For the sake of this discussion, it is assumed
that the total air requirement of the process flows through a number of
41
-------
100'-
4J
0)
0)
4J
4)
M
O
4J
m
I
o
u
80
60
Expected
|0ffgas
"~^
40
20
10
20
30
40
50
60
Offgas Velocity, ft/sec.
FIGURE 14-COMBUSTOR DIAMETER VERSUS OFFGAS VELOCITY
42
-------
single sized lances. The 361 TPH of coal discussed above requires
1222 tons per hour of air, which is equivalent to about 1.6 MM ACFM
or 26,600 ACFS. For lances designed to carry air at a rate of 200
ft/sec., a 134 ft lance I.D. area is required. If 18 inch I.D. lances
are used with a 1.75 ft I.D. area, then 76 lances will be required—
or 25 for each 38-foot diameter combustor. These lances would have an
outside diameter of 30 inches ( 5 ft ); therefore, for each 38 diameter
combustor (1135 ft area) about 125 ft would be used for lances. Thus,
only about 10 percent of the available bath surface is taken up by the
lances. When three 38 foot diameter combustors are in operation, each
will have a superficial offgas velocity of about 21 ft/sec, which would
increase to about 23 ft/sec, because of the area taken up by the lances.
When two of the three 38 foot diameter combustors are operating at full
load, the velocity would be 31 ft/sec, which would increase to about
34 ft/sec, if the decrease in available area due to the lances is
accounted for. Thus, even under the two combustor operations, all
design criteria are met.
43
-------
ECONOMIC EVALUATION
.To evaluate the economics of the Two-Stage Coal Combustion Process,
a process simulation computer program was prepared for the process
as shown in Figure 15. This program determines the energy and
material balance for a given power plant size and set of operating
conditions. Operating conditions such as coal composition, slag
basicity (ratio of CaO to Si02 + A120 ), the fraction of sulfur in
the slag, and the flux composition can be varied in the program.
Incorporated in the simulation program are subprograms which also
allow the capital and operating costs of the Two-Stage Coal Combustion
Process to be estimated.
For cost estimating purposes, process equipment has been grouped into
various equipment complexes—coal preparation, slag preparation', flux
preparation, desulfurization, combustor, and air preparation. The
equipment breakdown for these complexes and the capital and operating
cost estimating procedure are presented in the Appendix.
For a 1000 MW power plant operation, the important process stream
rates are shown in Table II for a 3.6 percent sulfur, 14,5 percent
ash, 12500 BTU/lb heating value coal. The estimated fixed capital
requirement for the Two-Stage Coal Combustion Process, retrofitted
into a existing facility, is shown in Table III and seen to be $23.33
MM or $23/KW (1980).
Because the combustor is essentially a coal gasifier, it can be used
to produce a hot combustible gas from coal as a feed for a gas-fired
boiler.
TABLE 2
*
Stream Quantities for Two-Stage Coal Combustion Process
(1000 MW Power Plant)
Stream Ton/Hr
IN
Coal from Coal Preparation 361
Combustor Air (Injection plus Combustion Air)(1182°F) 1222
Limestone 24
Slag to Combustor (From Desulfurization) 104
OUT
Sulfur Bearing Slag to Desulfurization 166
Combustor Offgas to Steam Generation (2500°F) 1418
Iron 6
Sulfur 13
Desulfurized Slag (To Storage and Sale) 60
Operating Conditions:
14.5% Ash, 3.6% Sulfur in Coal; 0.1 Slag Basicity, 8% Sulfur Slag
45
-------
Combustor Air
Stack
s Secondary Combustion
1 Air
Coke Transport
Air
Flue Gas
t
Air
Compression
Steam
Generation
Steam
Turbine
Electricity
Combustion
Air —
r
Air
Air
Preheater
Combustion 1 Air
Combustor Offgas
Coke
Jl
Limestone
Combustor
Slag
Slag to Combustor ti
Storage
Desulfurized Slag
Desulfurization
Steam
Sulfur
FIGURE 15 - PROCESS FLOW DIAGRAM
46
-------
TABLE 3
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 2.48
Combustor 1.71
Desulfurization 0.60
1. Total Purchased Equipment Cost 7.04
Installation, Piping, Electrical
Instrumentation, Utilities
(70% of 1) 4.92
2. Physical Plant Costs • 11.96
Engineering and Construction
(30% of 2) 3.59
3. Direct Plant Cost 15.55
Contingency (10% of 3) 1.56
Contractor's Fee (6%'of 3) 0.93
4. T 0 T A L 18.04
Escalation to 1980 (14.75% of 4) 2.66
5. T 0 T A L 20.70
Interest During Construction
(12.6% of 5) 2.63
6. TOTAL FIXED CAPITAL (1980) 23.33
47
-------
For this reason, a comparison was made of the investment required
and operating costs for two types of grass root plants: (1) coal-
fired boiler without S0_ control, and (2) gas-fired boiler with
qombustor supplying hot gas from a coal feed. It is expected that
the hot combustor offgas can be fired in a natural gas boiler with
a minimum of modifications.
To estimate the fixed capital requirement for a grass roots 1000 MW
Two-Stage Coal Combustion Power Station, power plant fixed capital
costs were obtained . These costs are shown in Table III for coal-
fired and gas-fired boiler operations producing 3500 psig - 1000°F
steam. Costs were determined using the standard FPC costing procedure.
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. To establish the .-cost of
a Two-Stage Coal Combustion Power Station, the capital cost of Table
III was combined with the gas-fired fixed capital requirement. As
seen in Table IV, the capital requirement if $163.2 MM and $162.8 MM
for the coal-and gas-fired facilities. On a dollar per kilowatt basis,
the Two-Stage Coal Combustion Power Station results in a net cost
increase of only $1.5/KW.
The estimated operating costs (shown in Table V) for the above power
plants are 6.57 and 6.40 mills/KW hr, if by-product credits are taken
in the ATC process. Without by-product credits for sulfur (420/ton),
iron ($20/ton) and slag ($0.5/ton), the operating costs are respectively
6.57 and 6.82 mills/KW hr. The conventional coal.burning plant is,
however, introducing S02 into the atmosphere. Incorporation of S0_
abatement equipment would probably add at least $25/KW to the cost of
the conventional coal plant bringing the operating cost to 7.14 mills/KW
hr. Thus, the Two-Stage Coal Combustion Process becomes extremely at-
tractive in comparison—even if no by-product credits are taken (operating
cost = 6.83 mill/KW-hr).
Table VI 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 VI
are:
1. The air preheat temperature range is 1000 - 1200°F.
2. The capital requirement for a Two-Stage Coal Combustion
Process plant is essentially independent of coal compo-
sition and is about $23/KW.
3. The net capital cost increase of a Two-Stage Coal
Combustion Process - power plant over a conventional
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.
Other important economic conclusions which have resulted from the
process simulation are:
1. Coals with as low a moisture content as practical (1-1 1/2%)
should be used to reduce operating cost and air preheat
temperature requirements.
2. Limestone, lime or combinations of the two can be used to
flux the coal ash; however, a minimum amount of lime should
be used to minimize operating cost.
3. Both process operating cost and air preheat temperature
increase as the slag basicity increases. Based on experi-
mental results to date, a minimum slag basicity of 0.1
yields operable slags and was selected to minimize operating
cost.
4. Experimental results indicate that operable combustor
slags containing as high as 8 percent sulfur can be used.
Since both the operating cost and air preheat temperatures
increase with decreasing slag sulfur content, 8 percent
sulfur slags are preferred. However, even down to as low
as 4 percent sulfur in the slag, the operating cost is not
significantly increased.
49
-------
TABLE 4
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 (1270 of 2)
3. TOTAL
Escalation (1980)
(14.75% of 3)
4. TOTAL
Interest During Construction
(12.67o of 4)
5. TOTAL
Standby Oil Facility
6. TOTAL
Two-Stage Coal Combustion
Plant Cost
7. TOTAL
Net Mtf 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
Note:
All cost and factors, except coal combustion process cost, from
Reference 4.
50
-------
TABLE 5
Estimated Operating Cost
1000 MW Power Plants
All Figures in Mills per Killowatt-Hour
Capital Charge (14% Rate)
Operating, Supplies and Maintenance
Coal ($0.3/MM BTU)
Limestone ($3/Ton)
Iron Credit ($20/Ton)
Slag Credit ($0.5/Ton).
Sulfur Credit ($20/Ton)
TOTAL POWER COST
Mills per Kilowatt Hour
TOTAL POWER COST
Without By-Product Credits
Conventional
Coal Burning
3.683
0.323
2.567
0
0
0
0
6.572
6.572
Natural-Gas-
Coal Combustion
Process
3.717
0.326
2.710
0.068
(0.129)
(0.025)
(0.263)
6.403
6.820
51
-------
TABLE 6
Effect,of Coal Composition on Operating and 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 Burning
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
52
-------
REFERENCES
1. McGannon, H. E., "Making Shaping and Treating of Steel",
U. S. Steel Corp. 9th Edition, 1972, pp 494-496.
2. Op. Cit, pp. 479.
3. Perry, J. H., Chemical Engineers Handbook, McGraw-Hill, New
York, 1950, p. 547.
4. Robson, F., Etal, "Technical and Economic Feasibility of
Advances Power Cycle and Methods of Producing Non-Polluting
Fuels for Utility Power Stations", NAPCA, Contract CPA 22-69-114,
1970.
53
------- |