-------
2. CHARACTERIZATION OF EMISSIONS
FROM FLUIDIZED-BED COMBUSTION OF COAL
AND CONTROL OF SULFUR EMISSION
WITH LIMESTONE
R. D. GLENN AND E. B. ROBISON
Pope, Evans and Robbins
INTRODUCTION
This paper presents a brief review of
emission studies in the area of fluidized-bed
combustion, conducted by Pope, Evans and
Robbins for the National Air Pollution
Control Administration. The studies, in
general, involved a characterization of
emissions from fluidized-bed combustion of
coal, sujfur emission control by limestone
injection into the bed, and sulfur emission
control by combustion of coal in a limestone
bed. This paper also presents an estimate of
the potential for sulfur emission control by
continuous regeneration of a limestone bed
system based on some preliminary experi-
mental data.
CHARACTERIZATION OF EMISSIONS
The initial investigation to characterize the
emission from the coal-fired fluidized bed
indicated levels of sulfur dioxide, sulfur
trioxide, nitric oxide, hydrocarbons, and
particulates, and the extent to which these
could be modified by purely operational
changes; e.g., changes in excess air rate, coal
feed rate, and bed temperature. A summary
of the results of this investigation follows.
Sulfur Oxides
balance was retained by the ash. A very small
percentage, 1 percent or less, appeared as
sulfur trioxide in the flue gas. In the absence
of a sorbent, the sulfur emission was substan-
tially unaffected by changes in the operating
variables.
Nitrogen Oxides
Emission of nitric oxide (NO) was found to
be the dominant form in the NOX group. The
emission level, although low in comparison to
the level of other coal-fired units, was still an
order of magnitude greater than the equi-
librium value predicted from the bed tempera-
ture. Furthermore, the NO level did not
change when bed temperature was changed.
This result led to the conclusion that NO
emission was controlled by local temperatures
higher than the measured bed temperature.
The NO emission responded only to
increase in the oxygen content in the flue gas
(increase in excess air). In one test an NO
concentration of 280 ppm, observed at 1
percent oxygen in the flue gas, increased to
340 ppm at 4 percent oxygen. This variation
is shown in Figure 1. The average concentra-
tion of a test series conducted at 3 percent
oxygen in the full-scale unit was 275 ppm.
Hydrocarbons
About 95 percent of the sulfur in the coal The full-scale boiler could be operated with
appeared as sulfur dioxide in the flue gas; the as little as 5 percent excess air without visible
II-2-1
-------
smoke in the flue gas; however, hydrocarbon
concentrations were found to be as high as
1500 ppm (methane) at this excess air level.
This high concentration was reduced sharply
as the excess air rate was increased. Hydro-
carbon burnup was essentially complete at 24
percent excess air (4 percent C>2 in the flue
gas). The variation in shown in Figure 2.
Particulates
In the test system, particulate emissions
were a function of the cyclone collector ef-
ficiency. Analysis indicated that 90 percent of
the fly ash that passed the cyclone was
smaller than 20 microns in particle size.
EFFECT OF OPERATIONAL CHANGES
Effect of Limestone Particle Size
When limestone (BCR 1359) was added to
the bed in a comparatively coarse particle size
(-7 +14 mesh), the sulfur absorption was
rather poor. Under the most favorable condi-
tions, the limestone utilized in the sulfur
capture was limited to about 20 percent.
Emission of SC>2 was reduced by about 20
percent at a stoichiometric ratio of 1.0 and
about 40 percent at a stoichiometric ratio of
2.0. A dolomitic limestone (BCR 1337) of the
same size was more effective when the
stoichiometric feed ratio was based on the
calcium fraction alone; however, on a total
weight basis, it was equally poor.
At this point in time, it became evident to
many researchers that utilization of sorbent
was limited by the formation of a sulfate shell
on the surface of the sorbent particle.
Improvement required an increase in the
surf ace-to-mass ratio of the particle; i.e., a
reduction in particle size.
When limestone was ground to a fine
powder (-325 mesh) and injected into the
II-2-2
bed, the sulfur capture was markedly
improved. The reduction in SO2 emission
with increase in stoichiometric rates is shown
in Figure 3. The trend shows SC«2 reductions
of 40, 66, and 84 percent at respective Ca/S
ratios of 1, 2, and 3. The corresponding lime-
stone utilizations are 40, 33, and 28 percent.
These results were obtained with a high sulfur
(4.5 percent) coal burning in a 10- to 12-inch
deep fluidized bed operating in the tempera-
ture range of 1500 1600°F, and with a
superficial velocity of 12-14 ft/sec.
Comparable results were observed with a
medium (2.6 percent) sulfur coal.
The lower line on Figure 3 shows the
greater reactivity obtained with a dolomitic
limestone; however, the stoichiometric ratios
are calculated for calcium only. If the weight
of magnesium is included and results are
expressed in practical terms, such as pounds
of stone fed per pound of sulfur removed,
dolomite has no advantage over limestone.
In a short series of tests with calcined lime-
stone, relatively poor results were obtained, as
indicated by the upper line of Figure 3.
A detailed study with the particular lime-
stone used (No. 1359) showed that fine
grinding would be necessary for favorable
sulfur capture. This limestone is one of the
most durable, however, and such fine grinding
may not be necessary with less durable lime-
stones which may decrepitate from thermal
shock during the initial calcination stage.
Figure 4 shows the effects of limestone
particle size, bed depth, and temperature on
SC>2 reduction. All of the data in this series
was obtained while burning a 3-percent sulfur
coal in a fluidized bed of sintered ash. Lime-
stone was injected at a stoichiometric ratio of
2.6.
At particle sizes below about 400 microns,
the beneficial effects of reducing the particle
size of the limestone injected are clearly
evident. At particle sizes above 400 microns,
-------
some other factor—probably increased
residence time in bed—more than offsets the
decrease in surface area, and a slight improve-
ment is obtained with increase in particle size.
From a practical viewpoint, however, only
high reductions are important.
If SOX emissions from a fluidized-bed
boiler are to be controlled by the injection of
limestone, it must be a very finely divided
limestone.
Effects of Bed Depth and Temperature
Figure 4 also shows the effect of bed
depth. As expected, an 18-inch bed shows a
higher absorption than does a 10-inch bed.
Also shown is the effect of bed tempera-
ture; the beneficial effects of operating at a
temperature of 1550°F are clearly indicated.
The poorer results obtained at 1675°F and at
1800°F bring to mind some of the poor
results that have been obtained by injecting
powdered limestone into a conventional
pulverized-coal-fired furnace which must
operate at even higher temperatures. One of
the outstanding advantages of fluidized-bed
combustion is its ability to operate
economically at temperatures of 1500 to
1600°F. This makes the process particularly
well adapted for sulfur emission control
through the use of limestone.
LIMESTONE INJECTION
The injection of limestone for sulfur
control had no observable effect on hydro-
carbon emission. In several instances the ad-
dition of limestone reduced NO emission but
the effect was not reproducible. Sulfur
trioxide emission was reduced to zero when-
ever limestone was injected.
Results of the limestone injection tests
indicate that a 4.5-percent sulfur coal can be
made the equivalent of a 1-percent sulfur coal
with the injection of 1359 limestone at a rate
of 27 lb/1000 Ib of coal. A 2.6-percent sulfur
coal can be converted to a 1-percent sulfur
equivalent with the addition of 1359 lime-
stone at a rate of 10 lb/100 Ib of coal. The
corresponding Ca/S ratios are 1.9 and 1.2.
LIMESTONE BED
As an alternative to injecting pulverized
limestone, limestone was used as a bed
material. The effect on SO2 emission was
remarkable.
Tests conducted with a limestone bed of
-10 +20 mesh particle size and a 3-percent
sulfur coal indicated that SO 2 emission could
be almost completely eliminated for a period
of 2 to 3 hours. After this period, SO2 began
to appear in the flue gas above the bed with
concentrations increasing with time. After
about 4 hours, the SO2 emission represented
about 30 percent of trie input and the con-
centration of sulfur in the bed rose to 7.4
percent. This concentration corresponds to a
limestone utilization of about 16 percent. A
once-through process in a limestone bed was
clearly uneconomical, and attention was given
to the regeneration of the limestone for reuse.
Emissions monitored during both adsorption
and regeneration are shown in Figure 5.
In this test the sulfur content of the coal
was again 3 percent, corresponding to a flue
gas concentration of 0.24 percent SO2- As
before, for the first 2 hours, substantially all
of the SO2 was removed. The concentration
in the flue gas then rose gradually and, after a
little over 3 hours, the SO2 emission rose to
about a third of what it would have been if
coal ash rather than limestone had been used
as the bed material.
The operating conditions were then
changed. The coal rate was increased so that
the temperature, which had been 1550°F, in-
creased to 1950°F. The oxygen concentration
in the flue gas, which had been 3 percent, was
II-2-3
-------
dropped to 1 percent. A rapid desorption of
SC>2 took place, and 90 percent of that which
had been absorbed during more than 3 hours
of operation, was desorbed in a few minutes.
The SC>2 concentration of the flue gas
reached a peak of 8.1 percent—a value which
is more than 30 times higher than the 0.24
percent that would have been experienced if
the coal had been burned without the use of
limestone in the bed.
estimated from the SCb removal efficiency of
the regenerated limestone bed. Experimental
observations of SC>2 emission after regenera-
tion of the bed showed that the emission rate
increased linearly with time. The total sulfur
in the bed also increased simultaneously in a
linear fashion. The rate of sulfur emission
could, therefore, be related to the total sulfur
in the bed for a system operating with a
3.0-percent sulfur coal.
REGENERATIVE PROCESS
The results observed in this test suggested a
regenerative process for using a limestone bed
to control SC>2 emissions from a fluidized-bed
boiler. Absorption might be carried out in the
fluidized bed of the main part of the boiler.
Desorption, and the production of a flue gas
that is very high in SC>2 concentration, may
be carried out in another section. The product
shell limitation—the mechanism by which
absorption of SC>2 on the surface of a lime-
stone particle reduces its effectiveness for the
absorption of additional SC>2—can be
circumvented by repetitive use of the particle
surface. The high concentration of SC>2 in the
small volume of flue gas from the desorption
section can be removed and/or recovered by
any of the variety of processes that have
previously been proposed for the treating of
conventional boiler flue gas.
These processes, presently high in cost, will
become economical as a result of the high
S(>2 concentration and low volume of flue gas
to be treated.
A patent application has been filed on this
"SC>2 Acceptor Process."
The performance of a continuous absorp-
tion and regeneration system has been
A sulfur balance around an absorption cell
employing recycle of regenerated limestone
follows the relation:
(1)
where L is the bed mass, xs the sulfur fraction
in the bed, t time, GQ the sulfur input rate, XQ
the sulfur fraction in the regenerated lime-
stone, fL the fraction of the bed recycled per
unit time, and Ge the flue gas sulfur emission
rate.
Relating Ge to the sulfur in the bed,
equation (1) becomes:
= G0 + (x0 - Xs) L fL - KxsL (2)
where K is the proportionality constant
between the ps sulfur emission rate and the
total sulfur in the bed.
Integrating equation (2), the sulfur fraction
of the bed varies as follows:
_ GQ + xQLfL - [GQ + XQLfL - XQ (KL - LfL) ] e - (K + fL)t
KL +
(3)
II-2-4
-------
The transient term indicates that xs ap-
proaches equilibrium rapidly as the recycle
rate is increased.
At equilibrium,
disappears and
the transient term
xs =
x0LfL Ge
+ I) ~KL
(4)
If the sulfur fraction of the regenerated
limestone (XQ) is small, the term xgLfL can
be neglected. The sulfur emission rate Ge is
then related to the sulfur input rate, the gas
constant, and bed recycle rate by the relation:
K(G0)
= or
K + fL GO
K
(5)
The value of K determined by experiment
was found to be 0.32 hr"^ (Ib sulfur emission
per hr per Ib sulfur in the bed).
The measured values of
— £, — §•, and L
dt dt
were 0.37 lb/hr2, 0.0215 hr1, and 55 Ib,
respectively.
Figure 6 is a plot of the calculated SC>2
removal versus the number of bed changes per
hour. It is indicated that, with only one bed
change per hour, a 76-percent reduction of
SO2 emissions may be obtained. With two
bed changes per hour, an 87-percent reduc-
tion is predicted, and with three bed changes
per hour, a 91-percent reduction is predicted.
These results indicate a very favorable
potential for sulfur emission control by lime-
stone regeneration in a continuous operation.
Further investigation is planned.
II-2-5
-------
400
300
LU
Q
X
O
O
cc
200
100
1500
1.0
I
2.0 3.0
OXYGEN IN FLUE GAS. %
4.0
5.0
Figure 1. Effect of oxygen concentration on NO emission.
4.0
1.0 2.0 3.0
OXYGEN IN FLUE GAS, %
Figure 2. Effect of oxygen concentration on hydrocarbon emission.
5.0
11-2-6
-------
o
o
LU
cc
01
Q
O
Q
GC
U-
20
O
Q 40
LU
X
O
Q
tr 60
80
100
2.0
Ca/S RATIO
Figure 3. SC>2 reduction - - effect of limestone-to-sulfur ratio.
I STATIC BED
DEPTH
•OlO in.
>18 in.
T
I
200 400 600 800 1000 1200
LIMESTONE PARTICLE SIZE, microns
1400
1600
1800
Figure 4. S02 reduction - - effect of limestone particle size, bed depth, and temperature.
II-2-7
-------
0.8
0.7
0.6
0.5
se
w 0.4
UJ
o
Q
oc
0.2
0.1
u- 2000
O
5 1800
1600
8.1% S02
~1 T
S02 INPUT EQUIVALENT
\
XT'
I / \l I
/ \ %°2
.J\f\ TEMP
I iV/T"
0
Figure
II-2-8
234
TEST PERIOD, hours
5. S02 in flue gas - - regenerative
process using a limestone bed.
1.0 2.0 3.0
BED CHANGES PER HOUR BY RECYCLE
Figure 6. Calculated S02 removal by
regenerative process.
4.0
-------
3. POLLUTION CONTROL CAPABILITIES
OF FLUIDIZED-BED COMBUSTION
L. J. ANASTASIA, E. L. CARLS,
R. L. JARRY, A. A. JONKE, AND G. J. VOGEL
Argonne National Laboratory
ABSTRACT
The combustion of a high-sulfur coal has
been studied in a 6-inch dia fluidized-bed
reactor. Basic additives containing CaO (e.g.,
limestone and dolomite) have been used for
SC>2 control during combustion. The parame-
ter having the greatest effect on SC>2 emission
has been found to be the Ca/S ratio in the
additive and coal feed streams. At Ca/S mole
ratios of 2.0 to 2.5, about 80 to 90 percent of
SC>2 emission can be prevented. When coarse
additive particles are used as both the
fluidized bed and additive, the optimum
temperature range for SC>2 removal appears to
be 1480 to 1550°F.
INTRODUCTION
Fluidized-bed combustion is being studied
at Argonne National Laboratory as a method
of reducing the quantity of atmospheric
pollutants (oxides of sulfur and nitrogen)
released during the combustion of fossil fuels.
This work is being carried out under contract
with the National Air Pollution Control
Administration.
The concept of fluidized-bed combustion
involves burning fuel in a fluidized bed of
solids. In an industrial application, boiler
tubes would be immersed in the bed. The
rapid motion of the fluidized particles at the
heat transfer surfaces, together with high
heat-transfer rates between gas and particles,
makes the fluidized bed a highly efficient
heat-transfer medium. If the excellent heat
transfer characteristics of fluidized-bed
combustors allow combustor size to be
decreased, lower capital and operating costs
than for conventional units may be expected.
Another advantage of the fluidized bed in this
application is that it is a highly efficient
contacting medium for carrying out gas-solids
reactions; therefore, use of a bed material that
reacts with gaseous pollutants generated
during combustion offers good prospects for
controlling air pollution. Also, combustion
can be carried out at lower temperatures in a
fluidized bed than by conventional com-
bustion methods.
Disadvantages of the fluidized bed include
a gas velocity limited by the extent of entrain-
ment of unburned carbon from the fluidized
'bed; therefore, relatively large bed areas and a
large number of fuel introduction points are
required. The rapid motion of the fluidized
particles results in some decrepitation of bed
particles; the viscous drag of the gas is suf-
ficient to entrain fine particles from the
reactor, necessitating off-gas cleanup of both
additive and fly ash particles.
To control sulfur oxide emissions, lime-
stone or dolomite additive may be fed
continuously to the fluidized-bed combustor,
where calcination to lime (CaO) occurs
simultaneously with combustion of the fuel
and in situ reaction of sulfur oxides with lime
to form calcium sulfate.
II-3-1
-------
Two alternative modes of operation for a
desulfurizing fluidized-bed combustor have
been considered in which the major differences
pertain to additive particle size and composi-
tion of the fluidized bed. In both modes of
operation, crushed and washed coal (-6 and
-14 mesh) was used; at the fluidizing-gas
velocities employed (3 and 9 ft/sec), essential-
ly all the coal ash generated by combustion of
these coal particles was elutriated from the
fluidized bed. In the first mode of operation,
pulverized additive (<325 mesh) was used for
SC>2 control, and the fluidized bed was
refractory alumina having a relatively coarse
particle size (30 mesh). Since most of the ash
and additive was elutriated from the bed
during a run, the fluidized bed at steady state
was composed largely of alumina particles. In
the second mode of operation, large additive
particles used for SC>2 control were of suf-
ficient size to remain in the fluidized bed. At
steady state in this mode of operation, the
fluidized bed was composed essentially of
partially sulfated additive material.
EQUIPMENT, PROCEDURE, AND
MATERIALS
Figure 1 is a schematic diagram of the
fluidized-bed combustor system. The com-
bustor is a 6-inch dia stainless steel vessel
equipped with electric heaters and an air
cooling system for temperature control. To
start a run, preheated air is passed into the
combustor through a bubble-cap air dis-
tributor mounted on the bottom flange.
Electric heaters raise bed temperature to
1000°F, at which point coal is fed to the bed
at controlled rates. Ignition of the coal
increases the bed temperature to the desired
operating temperature (i.e., M600°F). When
combustion conditions have stabilized, feed-
ing of additive and/or recycle of elutriated
fly-ash/limestone mixture is started.
Variable-drive volumetric screw feeders
mounted on scales allow solids feed rates to
the system (coal, additive, and recycled
solids) to be metered. The solids are fed
pneumatically to the fluidized bed at a point
just above the gas distributor.
II-3-2
Flue gas from the combustor is passed
through two high-efficiency cyclone
separators in series and a fibrous-glass final
filter to remove entrained solids. Downstream
from the cyclones, approximately 5 percent
of the total flue gas is diverted to a gas
analysis system. The water content of the flue
gas sample is reduced to 3000 ppm (by con-
densation and refrigeration) to prevent the
moisture from interfering with gas analysis.
Continuous analyses of the dried gas for NO,
SC>2, and oxygen are carried out by infrared
analyzers and a paramagnetic oxygen
analyzer. Gas chromatography provides inter-
mittent analyses for CC<2. Periodically, the
bed and overhead solids are sampled for
chemical analysis and to obtain material
balances.
A Central Illinois bituminous coal mined in
Christian County has been used in all experi-
ments reported here. The coal contains 4.5 wt
percent S and 12.3 wt percent ash and has a
heating value of 1 1,550 Btu/lb. The additive
materials studied include:
1. Limestone No. 1359, Stephens City,
Virginia (97.8 wt percent CaCC«3, 1.3
wt percent MgCC«3).
2. Limestone No. 1360, Monmouth,
Illinois (78.0 wt percent CaCC«3, 22.0
wt percent
3. Dolomite No. 1 337, Gibsonburg, Ohio
(53.4 wt percent CaCO3, 46-5 wt Per~
cent
4. Tymochtee dolomite, Huntsville, Ohio
(49.3 wt percent CaCO^, 36.6 wt per-
cent MgCO3).
In several experiments with milestone No.
1359, the fuel was natural gas supplied by a
public utility through commercial gas lines.
Table 1 summarizes operating conditions
and variables studied in combustion experi-
ments. (Experiments pertaining to NO
-------
Table 1. OPERATING CONDITIONS AND VARIABLES STUDIED IN
FLUIDIZED-BED COMBUSTION EXPERIMENTS
Fuel
Coal
Coal
Gas
Coal
Coal
Coal
Number of
Experiments
9
4
3
4
5
1
Starting
Bed
Material
Alumina
Alumina
Alumina
Alumina
Partially
sulfated
limestone
Partially
sulfated
limestone
Additive
Type
Limestone No. 1359
Limestone No. 1359
Limestone No. 1359
Tymochtee
dolomite
Limestone No. 1 359
Limestone No. 1360
Dolomite No. 1337
Limestone No. 1359
Average
Particle
Size,
ptm
25
600
1400
25
103
25
<44
615
630
540
490
Combustion
Temperature,
°F
1600
1600
1600
1550-1650
1600
1600
1600
1600
1480-1800
1400-1600
Superficial
Gas Velocity,
ft/sec
3
3
9
3
3
3
3
Variables Studied
a) Gas velocity
b) Recycle of elutriated
solids
c) Additive particle
size
a) Combustion temperature
b) Additive particle size
c) Ca/S mole ratio
a) SO2 removal with natural
gas combustion
b) Additive regeneration
a) Additive type
b) Additive particle size
c) Ca/S mole ratio
a) Limestone fluidized
bed
b) Combustion temperature
a) Combustion temperature
-------
removal, discussed in another paper, are
omitted from this table.) The variables
studied include gas velocity, recycle of
elutriated solids, combustion temperature.
additive type, particle size, fluidized-bed
material, and the Ca/S mole ratio in the feed
streams. As discussed below, the Ca/S mole
ratio in the additive and coal feed streams
appears to have the greatest effect on SO->
control.
S02 EMISSION CONTROL WITH
FINE-PARTICLE ADDITIVE
The most thoroughly studied additive for
SC»2 control has been limestone No. 1359
because:
1. It has a minimum amount of waste
burden (e.g., its MgCO3 content is 1.3
wt percent).
2. In preliminary tests1 this stone
showed a high chemical reactivity
with SO2-
Pulverized limestone No. 1359, with an
average particle size of 25 (im, was studied at
addition rates equivalent to Ca/S mole ratios
in the feed up to ^3; Figure 2 summarizes the
results. The SC<2 content of the flue gas was
reduced by about 80 percent at a Ca/S mole
ratio of 2.5.
Figure 2 includes SC»2 removals for runs in
which recycled solids (a mixture of partially
sulfated limestone and fly ash) were
introduced into the combustor along with the
additive and coal feeds. For runs in which
recycled material was added, the contribution
of unsulfated calcium in the recycle material
was not included in the calculated Ca/S ratio.
The absence of any improvement in SC»2
removal, when using fresh limestone plus
recycled fly-ash/additive mixture, is puzzling
since the recycled solids contain significant
quantities of unreacted CaO. Also, as
discussed below, separate laboratory tests on
II-3-4
the reactivity and capacity of fresh limestone
and the recycle material detected no signifi-
cant difference in reactivity for the two
materials.
The effect of superficial gas velocity (3 and
9 ft/sec) for limestone No. 1359, with an
average particle size of 25 jum. is shown in
Figure 3. At the higher gas velocity, there was
less SO2 reaction with fine additive particles.
The difference is not large but does appear to
be significant; this result is not unexpected
since at the higher gas velocity, elutriation
rates and, therefore, bypassing of solid
particles are increased. At a superficial gas
velocity of 3 ft/sec, the residence time for
additive particles, calculated from the weight
of the fluidized bed and the additive feed
rates, is about 1 hour; however, experimental
evidence indicates that, even at the low gas
velocity, actual residence time for the bulk of
the additive particles is much less.
The effect of combustion temperature in
the range 1550 to 1650°F for 25-jum lime-
stone No. 1359 particles (Figure 4) appears to
be small; the data indicates that SO 2 removal
at 1600°F is slightly better than at either
1550 or 1650°F. Figure 4 also shows that a
90-percent reduction in SO2 emission occurs
at a Ca/S mole ratio
The overall utilization (conversion of CaO
to CaSO4) of 2S-nm limestone No. 1359
ranged from 25 to 40 percent (Figure 5). This
utilization was calculated from the relation-
ship between SO2 removal and the Ca/S mole
ratio in the feed streams. As might be
expected from the relationship between Ca/S
ratio and maximum CaO utilization (solid
curve, Figure 5), higher utilization of CaO was
obtained at the lower Ca/S ratios.
Chemical analyses of various effluent solids
were performed. Significant differences in
CaO utilization and in the extent of calcina-
tion for the different solids streams were
noted. The results (Figure 6) show that
material collected in the primary cyclone,
-------
which had the lowest extent of calcination,
also had the lowest CaO utilization. Material
retained in the fluidized bed, which had the
highest extent of calcination, also had the
highest extent of CaO utilization. Secondary
cyclone solids were intermediate with respect
to both extent of calcination and CaO utiliza-
tion. Since most of the entrained solids in the
flue gas might be expected to have similar
residence tunes in the reactor, the difference
in CaO utilizations for materials from the
primary and secondary cyclones can be at-
tributed to particle size. Larger particles were
collected in the primary cyclone; finer
particles, collected in the secondary cyclone,
had a higher specific surface area and reacted
more quickly and more thoroughly than the
larger particles collected in the primary
cyclone. For additive particles retained in the
bed, residence times are long and a high
degree of utilization and calcination can be
achieved. The Ca/S mole ratios for cyclone
samples and fluidized-bed samples taken
during a single experiment with 25-^tm lime-
stone No. 1359 are shown in Figure 7. In the
fluidized-bed samples, the Ca/S mole ratios
were relatively constant at 1.5, which
represents a CaO utilization of ^70 percent.
CaO utilization of elutriated material
collected in the secondary cyclone ranged
from 30 to 50 percent; CaO utilization for
material in the primary cyclone was 20
percent or less. The data in Figures 6 and 7
suggests that calcination of CaCO3 and
utilization of the resulting CaO are dependent
upon the same factors, which include particle
size and residence time in the combustor.
In an attempt to increase the utilization of
CaO, elutriated solids were recycled in the
fluidized-bed combustor in several experi-
ments. Material collected in the primary
cyclone, having a CaO utilization of 20
percent, was recycled. As indicated in Figures
2 and 5, SO2 removal was not discernibly
improved by adding recycled material. To
explain these results, laboratory tests were
carried out to compare the rates of sulfation
of fresh limestone and recycled solids. Figures
8, 9, and 10 show the results of these tests.
The reaction rates for fresh limestone and
for recycled solids (mixture of fly ash and
partially sulfated limestone) were essentially
the same, both at the start of the experiment
and after 2 hours of reaction (Figure 8).
These tests were performed at an SO2 partial
pressure of 3.2 mm Hg, which is considerably
higher than SO2 pressures in the combustor
(<0.8 mm Hg). Additional tests were
performed (Figure 9) at an SO2 partial
pressure of 0.4 mm Hg: again, the reaction
rates for the fresh limestone and the recycled
solids were similar, both initially and after 2
hours of reaction. Thus, the apparently lower
reactivity of the recycled fly-ash/limestone
mixture in the fluidized-bed combustor
cannot be attributed to a difference in re-
action rate as a function of SO2 partial
pressure. The percentage of CaO utilized
during the laboratory sulfation tests was
determined for recycled solids and fresh lime-
stone (Figure 10). As might be expected from
the similarity of reaction rates, the total per-
centage of CaO utilized was the same for both
materials. Thus, the ineffectiveness of recycle
operation in reducing SO2 levels cannot be
accounted for by a reason intrinsic to the
reaction of recycled material with SO2-
Electron microprobe studies2 were carried
! out in an attempt to explain the anomalous
behavior of the recycled solids in the
fluidized-bed combustor. Cross sections of
typical particles, collected in the primary
cyclone during combustion experiments with
natural gas and with coal, were examined for
homogeneity of sulfur and calcium content.
Figures 11 and 12 show the results of the
examinations. Some differences in particle
composition noted are related to the fuel
used. With coal as the fuel, sulfur concentra-
tion usually decreases at increasing distance
from the surface of the particle; on the other
hand, sulfur concentrations in particles
collected during the combustion of natural
gas appear to be more uniform throughout
11-3-5
-------
the particle. In gas combustion experiments,
SC>2 was introduced into the coal feed port
from an external source and SC>2 concentra-
tion in the fluidized bed may have been more
uniform than during the combustion of coal.
Three of the four particles from the gas
experiment (Figure 12) had a calcium content
quite close to that expected for CaCC>3;only
one particle had a calcium and sulfur content
expected for CaSC>4.
The laboratory tests have indicated that
high utilization of the limestone is possible
and that the reactivity of recycled solid is
essentially the same as that of fresh lime-
stones. A lack of further reaction of recycled
solids in the combustor may be due to insuf-
ficient gas-solids contacting, which, in turn,
may result from short residence times. When
both recycled solids and fresh limestone are
fed to the combustor, the conditions in the
combustor may favor greater initial reactivity
of fresh limestone. Perhaps better utilization
of the recycle material could be achieved with
separate combustor beds for recycled additive
and fresh additive feed.
Effect of Additive Material
The effectiveness of pulverized Tymochtee
dolomite was compared with that of lime-
stone No. 1359 for SC>2 control during coal
combustion in a fluidized bed of alumina.
Figure 13 shows that the difference between
the additives is small, but that Tymochtee
dolomite yields higher 862 reductions at Ca/S
mole ratios smaller than 2. It is generally
thought that the MgO content of calcined
dolomitic additives helps keep the particle
pore structure open so that more of the CaO
can react with SC>2- However, the presence of
MgO in dolomites is a serious disadvantage
because MgO does not react with SO2 and
greatly increases the waste burden for
dolomitic additives; e.g., the stoichiometric
quantity to react with the sulfur in 1 Ib of 4.5
wt percent S coal is 0.29 Ib Tymochtee
dolomite, compared with only 0.15 Ib lime-
stone No. 1359.
II-3-6
The waste burden for any of the additive
materials can be considerably reduced in
operational schemes which include an additive
regeneration step. In such a scheme, sulfur
values could be recovered from the spent ad-
ditive, and regenerated additive can be
utilized for SO2 emission control so that the
net waste burden of the system would only be
due to fresh makeup additive.
SO2 EMISSION CONTROL WITH
COARSE-PARTICLE ADDITIVE
Experiments were performed with additive
particles too coarse to elutriate from the
fluidized bed; fluidized beds of partially
sulfated additive of a composition close to
that expected at steady-state operation were
used. Table 2 shows the experimental results
for limestones No. 1359 and 1360, and
dolomite No. 1337. Data for coarse limestone
is compared with results for pulverized lime-
stone No. 1359 (25 jum) in Figure 14. The
results for coarse additive particles tend to
cluster in a region of smaD area, indicating a
relative independence of the type of additive
material tested and, for limestone No. 1359, a
relative independence of particle size up to
615 jum. Although the coarse additive
particles have lower reactivity toward SO2
than do fine particles, their increased
residence time and the resultant large mass of
material in contact with the gas stream
resulted in the same degree of SO2 removal as
with finer (more reactive) particles, which
elutriate readily from the fluidized bed and
hence have shorter contact times with the gas
stream. The observed similarity of SO2
removals for a variety of stones and particle
sizes may be fortuitous or may indicate that
the Ca/S mole ratio in the feed streams is
highly significant.
The effect on degree of SO2 removal of
combustion temperature in the range 1400 to
1800°F was determined for coarse limestone
No. 1359 and dolomite No. 1337 (Figure 15).
The maximum SO2 removal of ^90 percent
-------
Table 2. SO2 REMOVALS3 ACHIEVED
WITH COARSE-PARTICLE ADDITIVE MATERIALS
Additive
Material
Limestone No. 1359
Limestone No. 1359
Limestone No. 1359
Limestone No. 1359
Limestone No. 1360
Dolomite No. 1337
Average
Particle
Size,
/urn
615
615
615
490
630
540
Ca/S
Mole
Ratio
2.6
2.3
2.5
2.5
2.3
2.2
S02
Removal,
%
79
83
79
86
74
81
aAt combustion temperature of 1600°F; with gas ve-
locity of "v>3 ft/sec.
occurred in the temperature range 1480 to
1550°F. In contrast, for pulverized limestone
No. 1359 (Figure 4), SC>2 removal was little
affected by variation of combustion tempera-
ture in the range 1550 to 1650°F. The
different behaviors of coarse and fine lime-
stone may be explained by recognizing that
reaction of the pulverized limestone with SC>2
depends on the extent of calcination.
Although finely powdered limestone readily
elutriates from the bed, it has a high specific
surface area and reacts at high rates so that
the extent of calcination may be relatively in-
dependent of bed temperature in the range
1550 to 1650°F. Thus the optimum tempera-
ture for reaction with SC>2 from pulverized
limestone is not necessarily the same as that
for coarse limestone.
CONCLUSIONS
Control of SCH emission by reaction at
1400 to 1800°F with basic additives (lime-
stone or dolomite) containing CaO has been
studied in a 6-inch diameter fluidized-bed
combustor. In the combustion of a high-sulfur
Illinois coal, SOo emissions have been
controlled effectively by feeding additive into
the combustor. For each of the additives
tested, Ca/S mole ratio in the feed had the
greatest effect on 862 emission. Emission of
about 80 to 90 percent of the sulfur has been
prevented by feeding additive and coal at a
Ca/S mole ratio of 2.0 to 2.5. For fluidized
beds of limestone No. 1359 or dolomite No.
1337, the optimum combustion temperature
for SC<2 removal appears to be 1480 to
1550°F. Variables that have minor effects on
SO2 removal are additive type, additive
particle size, and superficial gas velocity.
BIBLIOGRAPHY
l.Borgwardt, R. National Air Pollution
Control Administration. Private com-
munication, 1968.
2. Natesan, K. Argonne National Labora-
tory. Electron Monoprobe Studies.
II-3-7
-------
00
PREH EATER
ADDITIVE
FEEDER
I
TRANSPORT M
AIRHXr—'
COAL
FEEDER
SECONDARY
CYCLONE
RECYCLE-
MATERIALS
FEEDER
TO
FILTER
AND
VENTILATION
EXHAUST
TRANSPORT
AIR
COMBUSTOR
Figure 1. Equipment for investigating S02 emission in combustion experiments.
-------
CO
(D
UJ
D
g
O
CM
O
CO
90
80
70
60
50
40
30
20
10
n
I I
1 1 1
8/0
h 7*
@f A TEMPERATURE: 1600°F
~~ 0 COAL FEED: 4.5 wt % S
/ ADDITIVE: LIMESTONE
I
—
—
—
— f FLUIDIZED BED: ALUMINA"
~~ ^ GAS
ft VELOCITY,
— / ft/sec
/ O 2.7
~l • 2.7
/ & 8.6
7 A 8.6
1 1 1
LIMESTONE
NO.
1359,
pm
25
25
25
25
RECYCLE
No
Yes
No
Yes
I I I
—
—
I
2345
Ca/S MOLE RATIO
Figure 2 Effect of calcium/sulfur mole
ratio on S02 removal with fine limestone
additive.
s?
C/J
O
cc
(M
I I
ft/sec
I
2345
Ca/S MOLE RATIO
S?
00
UJ
Of
O
cc
UJ
cc
23
Ca/S MOLE RATIO
Figure 4. Effect of combustion temperature
on S02 removal with fine limestone additive.
Figure 3. Effect of gas velocity on SO2
removal with fine limestone additive.
II-3-9
-------
o
O
(0
O
100
90
80
70
60
50
40
30
20
10
0
I I I I
GAS
VELOCITY,
ft/sec
( 0 2.7
\ • 2.7
\ £ 8.6
\A 8.6
LIMESTONE
NO.
1359
pm
25
25
25
25
RECYCLE
No
Yes
No
Yes
MAXIMUM CaO UTILIZATION
FOR REMOVAL OF 100% OF_
)o (no recycle)
25 \tm ADDITIVE
I I I I I I
O
$
O
<0
O
100
90
80
70
60
50
40
30
20
10
I I I I I I
_• FLUIDIZED-BED SAMPLES
A PRIMARY CYCLONE SAMPLES
.O SECONDARY CYCLONE SAMPLES
8 -
- a
oo
Ca/S MOLE RATIO
30 40 50 60 70 80 90 100
EXTENT OF CALCINATION, %
Figures. Utilization of CaO as a function of Figures. Relationship of CaO utilization to
calcium/sulfur mole ratio. extent of stone calcination.
HI
_l
Q_
z
o
1
UJ
O
CO
\
ID
U
/.u
6.0
5.0
4.0
3.0
2.0
1.0
0
I I I I II I
/\
/ X ^
a _^_ / <>• PRIMARY
/ ^ A
I ' I
/ Ca/S MOLE RATIO IN FEED
— / ^-* '9 SECONDARY
/ .— -• *^^«
/ • •— • •*
/ /n ^ 0 ^ c 0 ruin
v / / W0o
« I I I I I
CYCLONE
CYCLONE —
3IZED BED
-i-
56
TIME, hr
Figure 7. Calcium/sulfur mole ratio in fluidized bed and in elutriated fines during a coal
combustion experiment, 25-um limestone No. 1359.
II-3-10
-------
I
O
I
40
30
20
10
8
6
4
I
• FRESH LIMESTONE
O RECYCLED SOLIDS
O
I
I
20 40
60
TIME, min
80 100 110
Figure 8. Reaction rates at an SC>2 partial
pressure of 3.2 mm Hg for fresh limestone
and recycled solids.
c
3
I
g
I
3.0
2.0
1.0
0.8
0.6
0.4
0.2
0.1
I I I
1 T
O FRESH LIMESTONE NO. 1359
& RECYCLED SOLIDS
A UTILIZATION ACHIEVED FOR FRESH LIMESTONE,
LABORATORY TESTS
D UTILIZATION ACHIEVED FOR RECYCLED SOLIDS,
LABORATORY TESTS
O TOTAL UTILIZATION ACHIEVED FOR RECYCLED
SOLIDS IN BOTH FLUIDIZED-BED COMBUSTION
EXPERIMENTS AND LABORATORY TESTS
100
0.5 1.0 1.5
TIME, hr
Figure 10. CaO utilization for fresh limestone
No. 1359 and for recycled solids.
20 40 60 80 100 120 140
TIME, min
Figure 9. Reaction rates at an SO2 partial
pressure of 0.4 mm Hg for fresh limestone and
recycled solids.
II-3-11
-------
I I
*d ; 165 pm
20 40
DISTANCE FROM SURFACE, pm
60
I
* d = 140 pm
20
40
60
DISTANCE FROM SURFACE, pm ^
(«) d is dimension of particle in the direction of the trace.
Arrow 1 is for Ca in CaCO-j in standard samples.
*d - 70 pm
I
I
20 40
DISTANCE FROM SURFACE, pm
*d- 100 pm
20
40
1 *•
60
60
DISTANCE FROM SURFACE, pm »~
Arrow 2 is for Ca in CaS04 in standard samples.
Arrow 3 is for S in CaS04 in standard samples.
Figure 11. Microprobe traces showing calcium and sulfur levels in typical elutriated par-
ticles collected in the primary cyclone during combustion of coal.
JI-3-12
-------
: 65
0 20 40 60
DISTANCE FROM SURFACE, \>m
0 20 40
DISTANCE FROM SURFACE, pn
T
60
I I T
*d : 100 pm
Ca
I
1
*d = 50
Ca
^/VV^V/WA^^^/N^V
20 40
DISTANCE FROM SURFACE,
3 *-
I
0 20 40 60
DISTANCE FROM SURFACE, pm
I
60
(«) d is dimension of the particle in the direction of the trace. Arrow 2 is for Ca in CaSO4 in standard samples.
Arrow l is for Ca in CaCOg in standard samples. Arrow 3 is for S in CaSO^ in standard samples.
Figure 12. Microprobe traces showing calcium and sulfur levels in typical elutriated par-
ticles collected in the primary cyclone during combustion of natural gas.
II-3-13
-------
100
90 -
80 —
* 70
CO
O
£ 60
50
40
cc
CM
30
20
10
t LIMESTONE NO. 1359,
' 25 gm
52
100
90
80
70
60
E 50
1
i 40
DC
gT 30
20
10
I I
POINTS: COARSE ADDITIVE AND
ADDITIVE FLUIDIZED BED
' CURVE: FINE LIMESTONE NO. 1359
(25 Hm> AND ALUMINA
FLUIDIZED BED
I
I
1 2
Ca/S MOLE RATIO
Figure 13. Control of S02 emission with
Tymochtee dolomite No. 1359 limestone.
100
90
80
70
3?
i eo
>
§ 50
oc
g" 40-
30
20
10
1300
0123
Ca/S MOLE RATIO
Figure 14. SC>2 removal with coarse additive
and inert fluidized beds and with fine additive
and additive fluidized beds.
I I
A LIMESTONE NO. 1359
AVERAGE SIZE 490 \m
Ca/S - 2.5
O DOLOMITE NO. 1337
AVERAGE SIZE 540 gm
Ca/S - 2.2
6-in. DIA BENCH-SCALE UNIT
GAS VELOCITY: ~3 ft/sec
BED HEIGHT: ~2 ft FLUIDIZED
3-4 vol % 02 IN OFF GAS
1
1400
1500 1600
TEMPERATURE, "F
1700
1800
Figure 15. Effect of fluidized-bed combustion temperature on 802 removal.
II-3-14
-------
4. A REGENERATIVE LIMESTONE PROCESS
FOR FLUIDIZED-BED COAL COMBUSTION
AND DESULFURIZATION
G. HAMMONS AND A. SKOPP
Esso Research and Engineering
Linden, New Jersey
ABSTRACT
The desulfurization efficiency attained on a
3-inch ID fluidized-bed coal combustor
operating with a limestone bed is reported as
a function of the independent variables of the
system. A kinetic model was derived for use
in design studies; the reactor length-to-
diameter ratio (L/D) was suggested as a
scaling parameter in this model. High L/D
ratios were found to give relatively poor
•desulfurization efficiency. It was demon-
strated experimentally that the stone main-
tains a relatively high capacity with repeated
cycles of combustion and reductive regenera-
tion. NOX emissions decreased as a com-
bustion run progressed and the average SO2
concentration in the batch reactor increased.
This indicated a possible interaction between
SOX and NOX which resulted in lowered NOX
emissions. Small scale experiments using a
simulated flue gas showed that SO 2 and NO
did react to an appreciable extent in the
absence of oxygen. A different mechanism
appeared to be operative in the presence of
oxygen but a decrease in NOX was still
observed.
INTRODUCTION
Esso Research and Engineering Company is
conducting an experimental program for the
National Air Pollution Administration (under
NAPCA contract CPA 70-19) on a system in
which finely ground coal is combusted in a
fluidized bed of limestone. The Esso study is
a part of NAPCA's overall program to
examine fluidized-bed combustion as a
possible new boiler (power regeneration)
technique. The air pollution control potential
of a fluidized-bed combustor is great because
of the good gas-solid contacting and,
consequently, by appropriate choice of bed
material, the high SO2 removal efficiency
which can be attained. Limestone appears to
be one of the most promising materials for
capture of SO2- The sulfur in the coal is
oxidized to SO2, which is then captured by
the lime as CaSO4- The system proposed by
Esso involves transferring the partially
sulfated stone from the combustor to a
separate regeneration vessel in which the
following reaction occurs:
CaSO4
CO
H2
CaO + SO2 +
C02
H2O
(1)
The regenerated stone (CaO) can then be
returned to the combustor for further use.
This method of operation naturally reduces
the fresh limestone requirements substantial-
ly. The off gas from the regeneration unit has
a high SO2 concentration and can be utilized
as a feed to a byproduct sulfur or sulfuric acid
pknt. Figure 1 is a schematic diagram of this
process.
Finely ground coal Ov200 fji) is being
utilized in the experimental program to
H-4-1
-------
facilitate separation of the coal ash from the
limestone bed material by entrainment of the
coal ash from the bed.
Esso's fluidized-bed combustion program
has four primary objectives:
1. Investigate reactor combustion effi-
ciency when feeding finely ground
coal.
2. Investigate flue gas desulfurization
attainable as a function of the in-
dependent variables of the system.
3. Determine the potential for high
temperature reductive regeneration of
the sulfated bed material.
4. Determine the level of NOX emissions
from the Esso fluidized-bed com-
bustor and investigate possible
catalytic reduction of NOX by lime-
stone.
This presentation is concerned primarily with
the air pollution control aspects of the pro-
gram; i.e., the last three objectives.
EXPERIMENTAL EQUIPMENT
Two fluidized-bed units—a combustor and
a regenerator—are being utilized in the Esso
experimental study. Operation in these units
during regeneration studies is on an inter-
mittent basis: a batch of limestone is charged
to the combustor; coal is continuously fed to
the combustor until the desired level of lime-
stone sulfation is achieved; the combustor is
cooled; and the partially sulfated material
from the combustor is transferred to the
regenerator. Following regeneration, the
solids are returned to the combustor, and
another cycle is begun.
Figure 2 is a flow diagram of the Esso
fluidized-bed combustor (FBC). The reactor is
a 3-inch ID incoloy tube. Four continuous
flue gas analyzers are used: NDIR SC>2 and
CO analyzers, and polarographic NOX and ©2
analyzers.
Figure 3 is a flow diagram of the fluidized-
bed regeneration unit. The reactor is a 2-inch
ID alumina-ceramic tube. An NDIR analyzer
is used for continuous measurement of SO2 in
the off gas.
EXPERIMENTAL RESULTS
Range of Variables Examined
Table 1 shows the range of variables
examined in the Esso program. Only one coal
and one limestone have been utilized in this
study.
Table 1. RANGE OF VARIABLES EXAMINED
Variable
Combustor
Bed temperature
Settled bed height
Superficial gas velocity
Excess air
Average stone particle diameter
Average coal particle diameter
Coal
Limestone
Regenerator
Bed temperature
CO
C02/C0
Superficial gas velocity
Range
1500-1800°F
4-16 in.
2-4 f ps
3-50%
460-930/1
200 At
Bit. (A) (3% S)
N-1359
2000°F
10mol?
2
2fps
Reactor Combustion Efficiency
Table 2 shows the effect of the unit
operating conditions upon the reactor com-
bustion efficiency. Perhaps the most unusual
effect was on bed height. Increasing the bed
height might be expected to cause a longer
solid residence time and, consequently, a
higher combustion efficiency. The opposite
II-4-2
-------
effect was observed in this study; i.e., as bed
height was increased, combustion efficiency
decreased. This effect is believed to be the
result of the slugging nature of the small Esso
FBC. This slugging apparently causes a
reduced coal particle residence time as bed
height is increased. Hence, the increased
velocity of the coal particles through the
bubble phase in deep beds apparently more
than offsets the effect of the additional bed
height.
Table 2. EFFECTOR INDEPENDENT VARIABLES
ON COMBUSTION EFFICIENCY
Variable
Increasing this variable
Bed temperature
(1500-1800° F)
Settled bed height
(4-16 in.)
Superficial gas
velocity
(2-4 fps)
Excess air
(3-50%)
Average stone particle
diameter
(460-930 ju)
Increases combustion
efficiency
(91-97%)
Decreases combustion
efficiency
(94-91%)
Decreases combustion
efficiency
(97-91%)
No effect above 10%
excess air
No effect
Desulfurization
The desulfurization reaction was inves-
tigated over the range of conditions previous-
ly shown in Table 1. Figure 4 shows typical
batch desulfurization data obtained. The
capacity of the stone at 20 percent SC>2
breakthrough has been chosen as a relative
measure of the desulfurization efficiency at a
given set of conditions. Table 3 shows the
effect of the reactor operating conditions
upon the desulfurization efficiency of the
Esso FBC. The effect of bed height upon the
desulfurization efficiency again apparently
reflects the slugging nature of the small Esso
fluid bed. Increasing the excess air level
decreased the stone capacity. One possible
explanation which has been advanced1 for a
similar effect is as follows: at high excess air
levels (high C>2 content in the bed), the
following reaction sequence can occur:
CaO + SO2
CaSO3
CaSO3 + 1/2 O2
(2)
(3)
At high C>2 levels, reaction (3) occurs quickly
and apparently causes a pore blockage, re-
stricting the accessible CaO. However, at low
excess air levels, reaction (3) does not proceed
as quickly so that less pore blockage by the
CaSO4 is experienced.
Table 3. EFFECT OF INDEPENDENT VARIABLES
ON STONE CAPACITY
Variable
Increasing this variable
Excess air (3-30%)
Decreases stone capacity
(23-12%)
Average stone particle Decreases stone capacity
diameter (460-930/1) (32-18%)
Superficial gas velocity No effect
(2-4 fps)
Bed temperature
(1500-1800°F)
Settled bed height
(4-16 in.)
Not effective above 1600 F
Decreases stone capacity
(20-7%)
aCapacity at 20% SO2 breakthrough.
Kinetics of Desulfurization
A kinetic model derived in a previous Esso
study2 has been utilized to correlate the
II-4-3
-------
desulfurization results from the present work.
Figure 5 gives the model assumptions and
development. Apparent rate constant k(X) is
a function of CaO utilization X.
Figure 6 shows the experimental kinetic
data from the Esso FBC program. Constant
k(X) is observed to be a function of bed
height, as well as CaO utilization. This bed
height effect is believed to be a result of
poorer gas-solid contacting (because of
slugging) as bed height is increased.
A comparison was made between the
kinetic data obtained in this study and data
obtained in a previous Esso study2 in which a
simulated flue gas was desulfurized. Figure 7
shows that the two sets of kinetic data agree
closely. The implication from this comparison
is that a large fraction of the SC>2 generated
during coal combustion in the Esso FBC is
released near the reactor inlet.
In tins study, the data shown in Figure 6
has been replotted at constant values of the
L/D (aspect) ratio. We believe that by main-
taining a constant L/D ratio in scale-up,
similar contacting efficiencies can be
expected. Figures 8 and 9 show that de-
creasing the L/D ratio at a constant set of
other operating conditions increases the
percentage of available CaO which can be
utilized in achieving a given SO2 removal.
Hence, it appears that high values of the L/D
ratio will give conservative estimates of the
performance of large scale equipment, hi
which the L/D ratio is less than unity. For
design purposes, the desulfurization data
shown in Figure 8 at an L/D of 3.7 should
provide a conservative estimate of commercial
reactor performance.
Stone Regeneration
The partially sulfated stone was re-
generated by reduction at 2000° F. A 2/1
volumetric ratio of CO2/CO was maintained
in the reducing gas to minimize sulfide forma-
tion. The reducing gas contained 10 mol
II-4-4
percent CO. Equation (1) is the primary re-
action occurring in the regenerator.
Figure 10 shows the SO2 concentration in
the regenerator off gas. The dashed line
indicates the SO2 concentration that would
be attained at thermodynamic equilibrium.
The occasional low values of SO2 concentra-
tion in the off gas are believed to represent
poor gas-solid contacting in the regenerator
rather than any kinetic limitations. The
system can be expected to attain thermo-
dynamic equilibrium in a commercial system.
Cycling of the bed material between
combustion and regeneration gave the stone
capacity data shown in Figure 11. Initially the
capacity of the regenerated stone was higher
than that of the fresh stone. This phe-
nomenon is possibly a result of a more
favorably crystal lattice, as compared to the
fresh stone, formed on eliminating 803 from
the sulfated material. Sorbent activity slowly
decreased on repeated regeneration. Also
shown in Figure 11 is cyclic stone capacity
data obtained in a previous Esso study2 using
a simulated flue gas for SO2 sorption.
NOX Emissions
Figure 12 is a typical record of NOX emis-
sions from the Esso FBC. As a run progressed,
and the SO2 level in the bed and flue gas
increased, the NOX emissions decreased. The
indication is that there is some interaction
between the SOX and NOX which causes ,the
NOX emissions to decrease.
In order to investigate what type of inter-
action might be occurring between NOX and
SO2, a series of experiments (using a
simulated gas with varying concentrations of
NO and SO2) was conducted in the 2-inch
regeneration reactor. Table 4 shows the
results of this exploratory series of experi-
ments.
-------
Table 4. NOX-SO2 INTERACTION IN A FLUIDIZED BED3 OF LIMESTONE
Bed Material
Gas phase
Partially sulfated
stone from combustor
Alundum
Gas phase
Partially sulfated
stone from combustor
Alundum
NOX Cone., ppm
Before S02
Intro.
After S02
Intro.
Transport Gas - N2
900
J840
J830
820
900
(440
\180
820
Transport Gas - Air
800
710
620
760
710
380
S02 Cone., ppm
Inlet
1290
( 785
\1510
1000
1290
785
1090
Outlet
1290
(300
(480
1000
1180
0
1090
aBed temperature-1630°F.
The indication is that the NO and S(>2 can
react in a partially sulfated bed in the absence
of 62; no reaction occurs over an inert bed in
the absence of O2- One possible reaction
mechanism which can be occurring is:
CaO + SO2 —
CaSOs + 2NO
CaS03-(NO)2 —
CaSO3 (4)
—• CaSO3 • (NO)2 (5)
-CaSO4+N2+l/2O2(6)
However, in the presence of O2, reaction oc-
curs between NOX and SO2 even over an inert
bed. Although no reaction mechanism has
been identified, the oxidized species, NO2
and 803, could possibly be involved.
Further studies will be made of the exact
nature of SOX NOX interaction in the
presence of 62- Methods of utilizing this
interaction to effect low NOX emissions from
a commercial reactor will also be investigated.
BIBLIOGRAPHY
l.Coutant, R.W. et al. Investigation of the
Reactivity of Limestone and Dolomite for
Capturing SO2 from Flue Gas. Summary
Report to NAPCA under Contract PH
86-67-115, August 1968.
2. Skopp, A. et al. Fluid Bed Studies of the
Limestone Based Flue Gas Desulfurization
Process. Final Report under NAPCA
Contract PH 86-67-130, August 1969.
II-4-5
-------
FLUE GAS
& COAL FLY ASH
t
FLUID-BED
COMBUSTOR
CaO + SO2 + 1/ 02
— -3» CaS04
t I *
LSULFATED
SORBENT
HIGH S02 GAS
FLUID
| | REGENERATED
FRESH FLUIDIZING SORBENT
SORBENT AIR
+ COAL
-^
t
REDUCING
GAS
PLANT
-BED REGENERATOR
DISCARDED
*" SORBENT
CaS04+CO — » SO2-I
C02 + CaO
Figure 1-. Regenerative limestone system proposed by Esso Research.
AIR
REFRIGER-
EACTOR FILTER
I AIR
/ CON DENS
SCALES
NDIR SO2
ANALYZER
NDIR CO
ANALYZER
»/v\rr
POLAROGRAPHIC
NOX ANALYZER
POLAROGRAPHIC
02 ANALYZER
INTERMITTENT
GAS SAMPLER
Figure 2- Esso fluidized-bed combustion unit.
11-4-6
-------
ANALYZER VENT
COMBUSTION
ZONE
THERMO-
COUPLE
ALUMINA
REACTOR
lso:
STEAM-
COMBUSTION AND
FILTER OVEN
BY PASS
LINE
AIR / I WATER CONDENSER
CONDENSER H" AND KNOCKOUT
I
?H,O
5555
_» N2_f C02_f AIR—*
NDIR SO-,
ANALYZER
AND
RECORDER
Figure 3. Flow diagram for the fluidized-bed regeneration unit.
2500
2000
£ 1500
Q.
CO
O
8
i 1000
CM
8
500
I I I
S02 EMISSIONS IN THE
ABSENCE OF REMOVAL
1
MATERIAL BALANCE: U
S02 REMOVED FROM GAS : S02 ABSORBED BY
SORBENT
YIELDS: In (CH/CO) = -k(X) . Ho/U
MODEL ASSUMPTIONS:
COMPLETE MIXING OF SOLIDS
PLUG FLOW OF GAS
ALL S02'RELEASED AT REACTOR INLET
2345
RUN TIME, hours
Figure 4. Typical batch desulfurization data. Figure 5. Esso kinetic model development.
II-4-7
-------
1.8
1.5
1.2
= 0.9
5. 0.6
.*
0.3
0.0
5.0
4.0
T 3.0
c
o
2-0
1.0
0
D SIMULATED FLUE GAS DATA3
O PRESENT EXPERIMENTAL DATA
WITH COAL COMBUSTION
0
30
10
15 20
CaO UTILIZATION, %
25
10 20
X, CaO UTILIZATION, %
T; 1600 "F
EXCESS AIR -10%
(dp) COAL - 200 \i
(dp) STONE : 930 p " aData from reference 2.
Figure 6. Kinetic constants from Esso data. Figure 7. S02 formation occurs primarily near
reactor inlet.
T-1600"F
U = 3-4 FPS
H0 = 4 in.
(dp) COAL ; 200 p
EXCESS AIR ~10%
(dp) STONE : 930 |
O
i.o
0.9
0.8
0.7
0.6
0.5
"T—|~T
X:9
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
u
ao, seconds
L/D i 3.7 u |(dp) STONE : 930 p
EXCESS AIR ~10%l(dp) COAL : 200 M
Figure 8. Kinetic results at L/D = 3.7-
0 0.1 0.2 0.3 0.40.5 0.6 0.7 0.8 0.91.0
LJ
—o seconds
L/D: 10.3 U |(dp) STONE: 930 p
EXCESS AIR ~10% I (dp) COAL = 200 p
Figure 9. High L/D gives lowstone utilizations.
II-4-8
-------
10
O
en
LJJ
QC
X
<
i r
EQUILIBRIUM
U : 2 FPS
CO 10 MOL
C02/C0 : 2
T r
I I I
1
0.5
D Z
CM
O
gs 0.4
O
flj
O
4567
CYCLE NUMBER
8
0.2
0.1
SORPTION 1600 °F
REGENERATION 2000°F
U- 2 FPS
500 H N-1359 -
SIMULATED FLUE GAS£
930 fi N-1359
PRESENT STUDY
VITH COAL COMBUSTION
-------
5. COAL-BASED SULFUR RECOVERY CYCLE
IN
FLUIDIZED-LIME-BED COMBUSTION
G. P. CURRAN, C. E. FINK, AND
EVERETT GORIN
Consolidation Coal Company
ABSTRACT
Sulfated dolomite was produced by the
combustion of coal in a fluidized bed of
dolomite continuously fed with raw dolomite,
A brief experimental study is reported here of
a two-stage process for regeneration of the
€3804; i.e., its conversion to CaO, by use of
coal as a reductant. The composition of the
off gases from this process is controlled so
that just enough CO + H2 is available to
reduce the SC>2 to elemental sulfur.
The results of a computer study on
management of the off gas to obtain
maximum yield of sulfur are also presented.
INTRODUCTION
A previous paper1 from these laboratories
presented data on the efficiency of removal of
sulfur in the combustion of coal within a
fluidized bed of dolomite. The sulfur is fixed
on the dolomite in the form of calcium
sulfate. Regeneration of the calcium sulfate to
reform calcium oxide also was investigated to
obtain data on acceptor life and dolomite
makeup requirements.
The regeneration in the previous study was
conducted by the use of partial combustion
with air of CO as a fuel gas. The essential
regeneration reaction is endothermic and may
be written,
CaSO4 + CO = CaO + CO2 + SO2 (1)
A H= +53,720 cal/g mol at 25°C.
Hence, the need for partial combustion to
supply heat.
Although the procedure was successful as a
laboratory expedient, it would be much more
practical in a commercial situation to use coal
as a reductant and fuel; i.e.,
CaSO4 + C
(2)
AH= +27,520 cal/g mol
Attempts to use char from high-sulfur
Pittsburgh seam coals for this purpose were
unsuccessful. It was found that under
reducing conditions and at the high tempera-
tures (>1900°F) required to effect reaction
(2), ash fusion was an insurmountable
problem.
Accordingly, to avoid this problem a two-
stage regeneration process was devised, using
coal as fuel. This paper describes the process
and presents preliminary experimental data
which demonstrates its feasibility.
This paper also discusses the thermo-
dynamics of recovery of elemental sulfur
from the regeneration off gases, and outlines
II-5-1
-------
the net savings in dolomite and energy
requirements by use of the process.
PROCESS DESCRIPTION
Figure 1 is a schematic sketch of the overall
process. In the first stage, CaSO4 is reduced
to CaS at a relatively low temperature (about
1850°F). In principle, CaSO4 can be reduced
to CaS by a series of reactions which can be
expressed as the overall reaction,
CaSC«4 + 2 C = CaS + 2 CO2
AH =+42,440 cal/gmol
(3)
In practice, however, additional carbon must
be burned to produce CO. The overall re-
action then can be expressed as,
X(2-A)-4(1-A)
02 (4)
= CaS + X CO2 +
AX
(1-A)
CO
where: I^A = mol ratio of C/CaSO4 required
for process heat balance, and
A = CO/ (CO + CO2>, mol ratio in
exit gas to satisfy the reaction
kinetics.
Another feature of the first-stage process is
that it is so conducted that sufficient reducing
gas is produced to reduce the SO2 formed in
the second stage to sulfur by reactions such
as,
= 1/2S2 + 2CO2 (5)
AH = -48,900 cal/g mol
Thus, AX/(1-A) = 2 and reaction (4) becomes
CaSO4 + 3.890 C + 0.890 O2
II-5-2
(4a)
= CaS + 1.890CO2 + 2CO
AH = O
The high CO/CO2 ratio in reaction (4a)
ensures that the CaSO4 reduction will occur
rapidly. With the use of coal, combustion of
its hydrogen content produces additional
heat, which is needed to preheat the incoming
coal and air. With respect to the coal-air re-
action, the first stage is a fluidized-gas
producer in which ash slagging will not occur
at nominal bed temperatures of 1850°F
because the combustion takes place in the
presence of heat sink created by the en-
dothermic reduction of CaSO4 to CaS - re-
action (3).
In the second stage, the acceptor, now in
the form of CaS, is contacted with air in a
fluidized bed at temperatures in the range of
1900-1950° F. No external fuel is used. The
sulfur is rejected by a series of reactions
which may be expressed as the overall re-
action,
CaS + 3/2 O2 = CaO + SO2 (6)
AH = -108,970 cal/g mol
A portion of the
oxidized to CaSO4 v*
incoming CaS first is
CaS + 2 O2 = CaSO4
(7)
Then the CaSO4 reacts with the residual CaS
via,
3 CaSO4 + CaS = 4 CaO + 4 SO2 (8)
to give CaO and SO2- At a nominal pressure
of 1 atm, the incoming air must be diluted
with recycled tail gas from the sulfur recovery
section in order to provide a AP SO2 driving
-force for reaction (8). A mol ratio of about
-------
1 /1 recycle-gas/air is required. The exo-
thermic reaction provides the preheat duty
for the incoming air and recycled gas.
The combined off gases from both stages
are treated in the sulfur recovery section
where the SC>2 produced in the second stage
is reduced to elemental sulfur by CO (and H2
formed by the water-gas shift reaction)
produced in the first stage. Since both CaSC>4
and CaS are present in the first stage, some
SC>2 inevitably will be formed by reaction (8).
In the presence of CO and H2, this SO2 will
be reduced to H2S, COS, and 82 with the last
material predominating. This situation is not
detrimental, since all sulfur compounds are
nearly completely converted to elemental
sulfur in the recovery section.
The purpose of the work reported here was
to study briefly the salient features of the
two-stage regeneration process and the
severity of the possible processing problems
outlined below.
First Stage
1. At the very low air/fuel ratios
required, the incoming raw coal may
form coke.
2. To demonstrate freedom from ash
slagging at reducing conditions using
Eastern steam coal.
3. To determine the operating tempera-
ture and coal/CaSO4/air ratios
required to generate sufficient CO
to satisfy reaction (4a).
4. To determine whether deposits form.
Second Stage
1. To determine the O2/CaS ratio re-
quired for optimum rejection of sulfur
as SO2, as a function of temperature
and APg02 driving forces.
2. Previous experience with the CO2
acceptor process, in which CaS is
converted to CaO and SO2 in the re-
generator under conditions similar to
those of the second stage, showed that
when CaS and CaSO4 exist simulta-
neously in the acceptor, a transient
liquid formed which led to massive
deposits and/or cementing together of
the acceptor particles. In this work, it
was vital to determine the nature and
extent of deposit formation.
EXPERIMENTAL
Generally, the equipment and experimental
procedure were the same as reported previous-
ly.1 Figure 2 is a schematic diagram of the
equipment.
Briefly, the reactor had a 4-inch I.D. and
was made from Type 316 stainless steel. The
fluidized-bed height was controlled at 36
inches by an overflow weir. The original re-
actor was modified by replacing the per-
forated disc, baffle, and plenum chamber at
the bottom with a cone having an included
angle of 40 degrees. All solids were fed to the
apex of the cone. The same reactor was used
for all steps of the process, including coal
preoxidation; i.e., no attempt was made to
operate all the steps simultaneously.
After reaching the programmed flows and
bed temperature, solids feeding was continued
nominally for three bed inventory changes
and the data was recorded over a 1-hour
balance period. By adjusting the electrical
power input to each of the three sections of
the reactor heater, the maximum temperature
gradient across the fluidized bed was held to
within 20° F.
The data workup was straightforward,
being based on measured values for the input
II-5-3
-------
and output streams, the dry exit gas rates,
product gas analyses, and the acceptor com-
positions as determined by a special assay for
the CaS and CaSC>4 content. In the first-stage
runs, considerable sulfur was rejected to the
gas through reaction (8). At the conditions
used, most of the SO2 was reduced to $2
which condensed as fog in the gas recovery
system and escaped with the product gas. The
amount of elemental sulfur formed was
obtained by a forced sulfur balance.
The acceptor was the same Tymochtee
dolomite (Western Ohio), sized to 16 x 28
mesh, used in the previous work.
RESULTS AND DISCUSSION
Preoxidation of Coal Feed
All work reported here was carried out
with Ireland Mine coal, a high-sulfur, highly
caking Pittsburgh seam coal. Early work on
feeding raw coal to stage 1 showed that this
was not practical due to a small amount of
coke formation. Accordingly, it was found
necessary to preoxidize the feed coal to
reduce its caking propensity.
A level of preoxidation of 5.3 wt percent
(defined as pounds of oxygen reacted with
100 pounds MF coal) was found sufficient to
prevent coke formation. The preoxidation
was carried out by continuous feed of raw
coal (28 x 100 mesh) to a fluidized bed at
700°F. A mixture of air and nitrogen was used
as fluidizing gas. The composition of the pre-
oxidized coal is:
Hydrogen
Carbon
Nitrogen
Oxygen (by diff.)
Sulfur
Ash
II-5-4
4.52, wt percent, MF
basis
72.44
1.29
6.71
4.34
10.70
Gross Btu, MF basis, by Dulong formula,
12,995 Btu/lb.
Stage 1 Runs
After preliminary runs were made to deter-
mine the extent of preoxidation required, a
series of runs were made with the above pre-
oxidized coal. Five runs were made with fresh
sulfated acceptor at 1825 and 1875°F and at
various input acceptor, coal and air rates.
Another run, to help determine any kinetics
effects of acceptor activity loss on the rate of
CaSO4 reduction, was made with an acceptor
exposed previously to seven combustion-
regeneration cycles. Run conditions and
results are shown in Table 1.
The objective of these runs was to deter-
mine conditions under which three simulta-
neous conditions could be satisfied: nearly
complete reduction to CaS, a ratio of CO +
H2/CaS = 2 in the product, and process heat
balance. For each run, a complete heat
balance was calculated with the results shown
in the last row of Table 1. The heat balance
was based on air and preoxidized coal fed to
the process at 100 and 700°F, respectively.
Comparisons of Runs A7-A7A and
A10-A10A show that, by increasing the
temperature, the desired ratio, (CO -*• H^)!
(total S in acceptor product), can be achieved
easily. Substantially complete reduction of
CaSO4 occurred in all the fresh acceptor runs,
showing that the reduction reactions are rapid
at temperatures of 1825°F and above, even
with (CO + H2) concentrations as low as 9.9
percent in the exit gas, as in Run A8. The
acceptor retention times in these runs ranged
from 0.8 to 1.1 hours. In run A9, made with
deactivated acceptor, the somewhat lower
level of CaSO4 reduction may have been
caused by the decreased retention time
brought about by the increased acceptor feed
rate which was used in order to keep the
CaSO4 input roughly comparable with that in
the other runs. Thus, deactivation has no
-------
Table 1. CONDITIONS AND RESULTS FOR FIRST-STAGE RUNS3
Run Number
A7
Temperature, ° F 1 825
b
Acceptor Feed i
Feed rate, Ib/hr 9.44
Comp., mol & of total Ca
CaO 10.50
CaS 0
CaS04
89.50
Coal Feed Rate, Ib/hr 4.79
Inlet Gas, SCFH
Air
N2 purges
Exit Gas Rate, SCFH Dry Gas
Exit Gas Comp. at Top of Bed
H20, mol %
H2
CH4
CO
C02
S2
H2S
COS
N2 (diff.)
Product Acceptor
Rate, Ib/hr
Comp., mol % of total Ca
CaO
CaS
CaS04
Unreacted Char, Ib/hr
Outlet Fluidizing Velocity, ft/sec
Bed Density, Ib/ft3
Bed Weight, Ib
Solids Retention Time, hr
.. _ „ Lb C Gasified
% Carbon Burnout 100LbCFed
% S in Acceptor Rejected to Gas
CaS
„ , 0 . CO + H2
Mnl RTfift
IVHJI naiiu, _, _ f\^r*f\
C3o * CaSOA
Process Heat Balance,
Btu Absorbed/lb Mol CaS04 Fed
152
3.6
231
12.72
3.88
0.48
9.79
18.81
1.03
0.29
0.03
52.97
6.27
29.8
69.6
0.6
1.99
2.19
24.6
6.15
0.98
60.3
21.5
99.2
2.20
-67,700
A7A
1875
b
9.44
10.50
0
89.50
4.79
152
3.6
235
12.06
4.65
0.20
12.30
17.55
1.05
0.34
0.03
51.82
6.25
31.3
68.1
0.6
1.88
2.27
23.9
5.98
0.96
63.3
23.3
99.1
2.84
-48,400
A8
1825
b
11.26
10.50
0
89.50
4.79
152
3.6
232
13.80
2.71
0.46
7.21
21.51
0.96
0.25
0.03
53.07
7.54
25.3
73.8
0.9
1.99
2.20
24.5
6.13
0.81
60.7
16.5
98.8
1.26
-27,400
A10
1825
b
11.01
10.50
0
89.50
4.93
125
3.6
207
15.07
3.98
0.55
9.29
20.61
1.06
0.26
0.02
49.16
7.35
27.0
72.2
0.8
2.31
1.96
27.0
6.75
0.92
54.8
18.4
99.0
1.57
-4840
A10A
1875
b
9.44
10.50
0
89.50
4.93
125
3.6
203
13.76
5.88
0.52
13.02
16.74
1.18
0.19
0.03
48.68
6.27
29.8
69.6
0.6
2.33
1.96
27.0
6.75
1.08
53.5
21.6
99.2
2.67
+7800
A9
1825
c
23.72
65.50
0
34.50
4.22
139
3.6
218
13.61
2.20
0.43
6.58
23.86
0.81
0.10
0.02
52.39
20.08
69.4
28.6
2.0
1.52
2.06
30.7
7.68
0.38
68.1
10.6
93.5
0.94
-19,200
aFuel was 5.3% preoxidized Ireland Mine coal, at an 8-psig system pressure.
Fresh sulfated dolomite.
cSulfated dolomite after 7 cycles of regeneration.
11-5-5
-------
obvious effect on the rate of CaSC>4 reduc-
tion.
No deposits of any kind were found, nor
was there evidence of ash slagging in any of
the runs.
In none of the runs was process heat
balance achieved simultaneously with the
desired ratio, (CO + H2>/(total sulfur) = 2.
Four of the six runs were strongly exo-
thermic. Table 2 shows calculated process
conditions which lead to thermoneutrality
and a (CO + H2)/(total sulfur) ratio of 2. For
the process calculations, a carbon burnout of
70 percent (versus 60.3 percent in Run A7)
was assumed, to allow the effects of a deeper
fluidized bed. The process calculations show
that the following CaSO4/coal/air ratios will
be required:
Coal/CaSO4 = 72 Ib coal/mol CaSO4fed.
Air/Coal = 27 SCF/lb coal.
Table 2. CALCULATED PROCESS CONDITIONS FOR HEAT
BALANCE IN STAGE 1a
aBasis: 1 Ib mol CaS04.
bHeat formation at 25° C.
(CaS04> in. - (CaO + CaS +
Preoxidized Coal
Sulfated Acceptor
Inert
CaO
CaSO4
Air
Heat of Reaction15
Total
MAP Char
Ash
Product Acceptor
Inert
CaO
CaS
CaS04
Product Gas
H2O
C02
CO
N2
H2
S2
Heat Losses
Total
Temp,°F
Lbs
Input
700
1800
1800
1800
100
72.45
56.66
6.56
136.14
145.4
Output
1850
1850
1850
1850
1850
1850
1850
1850
1850
1850
1850
1850
15.74
7.74
56.66
18.68
56.13
0.83
21.84
83.71
32.38
111.0
0.83
9.68
Mols
-
0.1173
1.000
5.032
—
—
0.3331
0.7781
0.0061
1.212
1.902
1.156
3.962
0.412
0.151
A'H, Btu
17,600
27,210
2,440
61,070
1,400
120,570
230,290
10,850
3,520
27,990
7,170
18,040
390
43,440
40,480
15,220
52,900
5,080
2,330
2,880
230,290
+ CO2 + CO + H20) out.
II-5-6
-------
The breakdown of the total oxygen required
at process conditions is:
Air
Coal
CaSC>4
33 percent
4 percent
63 percent
Small deposits caused by the CaSO4-CaS
transient liquid occurred in all the runs. Their
location, appearance, and composition were
the same as the deposits which were formed
in the CC>2 acceptor process regenerator.3 In
Run B2 the extent of deposit formation was
0.6 percent of the acceptor fed.
The air requirement corresponds to 22 per-
cent of stoichiometric air for combustion of
the 5 percent preoxidized Ireland Mine coal.
Stage 2 Runs
The feedstock used for these runs was, for
experimental convenience, prepared by
reducing the CaSC>4 to CaS, using CO both as
a fuel and reductant at 1750°F. Part of the
CO was burned with air to provide the pre-
heat duty for the incoming gas and solids. In
the second stage, a portion of the incoming
CaS is- oxidized via reation (7). The CaSO4
reacts with residual CaS via reaction (8),
thereby rejecting sulfur. The overall reaction
is highly exothermic and provides sufficient
preheat duty for the incoming air and diluent
recycled tail gas. During the program, about
twice as much diluent (N2> was used as is
required for process heat balance in order to
provide a conservative AP SO 2 driving force
with respect to reaction (8). Run conditions
and results are shown in Table 3.
In Run B2 the input ah- rate was varied to
determine the ©2 requirement which gives the
optimum sulfur rejection at 1950°F. The
optimum conditions appear to be about 95
percent of the theoretical ©2 needed to reject
all the sulfur. At higher ©2 inputs, the
product acceptor contains CaSO4_
It is not clear whether the lower level of
sulfur rejection in run Bl at 1900°F was
caused by the lower reaction rates for re-
action (8) and/or (7) or by inhibition due to
the lower AP SO2 driving force. More data is
needed on the kinetics of the second-stage re-
actions.
After repeated combustion-regeneration
cycles, it is likely, as in the CO2 acceptor
process, that the extent of deposit formation
will decrease drastically, and that the deposit
problem can be tolerated by shutting down
periodically to descale the reactor walls.
Sulfur Recovery
Sulfur is recovered by so blending the re-
generation off gases from the two stages that
the ratio of (CO + H2 + H2S + COS)/(SO2) is
equal to or slightly greater than 2. The gas is
first passed through a catalytic reductor to
effect reaction (5) and the corresponding re-
action,
SO2= 1/2S2 + 2H2O.
(9)
Small amounts of "trimming" air may be
added to the feed gas so that the product gas
from the reductor step continues a proper
feed to a multistage Claus plant. The equi-
libriums in reactions (5) and (9) are so fa-
vorable that substantially no CO or H2 remain
in the product gas. The Claus feed gas now
contains a ratio of (H2S + COS)/(SO2) = 2.
Figure 3, a schematic flow diagram of the
sulfur recovery section, also shows a typical
feed gas composition.
A thermodynamic analysis of the sulfur
recovery section was made with the basic
premises given below.
11-5-7
-------
Table 3. CONDITIONS AND RESULTS FOR SECOND-STAGE RUNS
Run Number
System Pressure, psig
Bed Temperature, °F
Acceptor, Mol % of Total Ca
CaS
CaSO4
CaO
Feed rate, Ib/hr
Inlet Air, SCFH
Inlet N2, SCFH
N2 Purges, SCFH
Exit Gas Rate, SCFH
Exit Gas Composition, Mol %
N2
S02
Outlet Fluidizing Vel., ft/sec
% Sulfur Rejected
Mol % Total Ca in Product
CaO
CaS
CaS04
Input O2, % of Theoretical
Outlet S02 Partial Press., atm
AP Driving Force, atm
Bed Density, Ib/ft3
Solids Retention Time, hr
B1
5
1900
73.3
2.4
24.3
4.64
70.5
147.0
3.6
215.0
95.76
4.24
2.42
82.1
86.5
7.3
6.2
92.2
0.0560
0.0680
21.2
1.22
Equilibrium is assumed to be established in
all stages with respect to the following re-
actions:
CO+1/2S2 = COS (10)
H2+1/2S2 = H2S (11)
2H2 + SO2=1/2S2 + 2H2O (9)
B2(l) B2(ll) B2(lll)
555
1950 1950 1950
72.3 72.3 72.3
1.8 1.8 1.8
25.9 25.9 25.9
4.95 „ 4.95 4.95
64.4 76.4 85.5
130.6 127.0 129.0
3.6 3.6 3.6
194.0 202.0 208.0
95.21 94.63 95.52
4.79 5.37 4.48
2.23 2.33 2.39
79.4 93.8 80.1
84.8 94.8 85.2
14.7 4.1 2.0
0.5 1.1 12.8
79.4 94.4 102.8
0.0632 0.0709 0,0592
0.0978 0.0901 0.1018
22.7 22.4 21.7
1.25 1.21 1.15
2CO + SO2=1/2S2 + 2CO2 (5)
3Sg = 4S6 (12)
S8 = 4S2 (13)
The equilibrium relationships in the above si>
reactions, at a given temperature and pressure
along with the four elemental balances
completely define the system in each stage.
II-5-8
-------
Experimental data4 obtained in our labora-
tories and by others5 shows that, at least
under laboratory conditions, it is very easy to
establish equilibrium in all of the above re-
actions even at very low temperatures,
provided an active alumina catalyst is
employed.
It is necessary, in order to maintain catalyst
activity, to operate above the dew point of
the sulfur to prevent its deposition on the
catalyst. Each stage was so operated, there-
fore, that the vapor pressure of sulfur at the
outlet temperature was 20 percent greater
than its partial pressure in the outlet gas.
Sulfur is condensed from the product gas
from each stage before feeding the gas to the
succeeding stage.
A computer program was set up using an
iterative trial-and-error procedure to solve for
sulfur recovery and product gas compositions
with the above restraints imposed on the
system.
The free energy data for reactions (5), (9),
(10), and (11) were taken from the Janaf
tables.6 Experimental data7 was used to
define the equilibrium constants for reactions
(12) and (13).
The outlet temperatures from each stage,
for the illustrative example given, are shown
in Figure 3. The total sulfur recovery
potential is 98.2 percent. Addition of a fourth
stage, operated at 270°F, increases the'
potential sulfur recovery to 99.3 percent.
Equilibrium acceptor
activity
Second stage
Fluidized boiler
Sulfur recovery section,
1 st reactor
= 0.41, equivalent
to 8 percent
makeup rate
= 1950°F
= 1800°F
= 900°F
Table 4. COMPARISON OF RELATIVE ENERGY
AND RAW MATERIAL REQUIREMENTS IN FLUI-
DIZED BOILERS
With Re-
generation
and Sulfur Once-
Requirement Recovery Through
Energy as Coal Equivalent3
Coal fed to regeneration 9.16 —
Sensible heat for fresh
dolomite 0.39 1.92
Heat to calcine fresh
dolomite 0.61 3.01
Less
Sensible heat regenerated
acceptor
Residual char from re-
generation burned in
boiler
Steam and boiler feed
water duty from sul-
fur recovery
Net 5.53 4.93
0.21 —
2.22 —
2.20 —
Energy and Raw Material Requirements
The relative energy and raw material
requirements were evaluated for the two
types of fluidized-boiler operations using
once-through dolomite and acceptor regenera-
tion with sulfur recovery, respectively. The
evaluation is based on the following process
conditions:
Sulfur in coal = 4.5 percent
Removal of SC>2 from
stack gas = 90 percent
Cost - iflon Coal Burned
I ncremental coal at
$6/ton +33 +30
Fresh dolomite at
$2.50/ton +28 +138
Worth of recovered sul-
fur at $25/ton -85 —
Total cost -24 +168
Relative cost 0 +192
alb/100 Ib coal burned in boiler.
II-5-9
-------
Table 4 compares the two systems. The net
difference in material costs between the once-
through and regeneration processes, about
$1.90 per ton coal burned in the boiler, is a
powerful incentive to continue the develop-
ment of the two-stage regeneration process.
BIBLIOGRAPHY
1. Zielke, C.W., H.E. Lebowitz, R.T. Struck,
and E. Gorin. Jour. Air Pollution Control
Assoc. 20: 164-169, March 1970.
2. Skopp, A.A. Proceedings of First Inter-
national Conference on Fluidized-Bed
Combustion, Hueston Woods State Park,
Oxford, Ohio. November 18-22, 1968.
3. Curran, G.P., C.E. Fink, and E. Gorin. U.
S. Dept. of Interior, OCR R&D Report
No. 16, Interim Report No. 3 on CO2
Acceptor Process.
4. Consolidation Coal Laboratories. Un-
published Data.
5. Pasternak, R. Brennstoff-Chemie, 50:
200, 1969.
6. Dow Chemical Co., Janaf Tables. Clearing-
house for Federal Scientific and Technical
Information, U. S. Dept. of Commerce.
7. Preuner, G. and W. Schupp. Zeit. Physik.
Chem. 68: 129, 1909.
II-5-10
-------
STEAM
1800°F
Ca SO4
FROM BOILE
FLUIDIZEC
BED
COAL
700 <>F
AIR
100 °F
CaO
TO BOILE
"I
t
1st
STAGE
REACTOR
~1850°F
t
1
r ./"A .
• HO
ASH
CaS
I
2nd
STAGE
REACTOR
~1950°F
I {^
co2
so2
100 op
V/^X
H2O
SULFUR
RECOVERY
ELEMENTAL
SULFUR
TO BOILER
,
,
t A Am C02 RECYCLE BLOWER \J)
I *AIR N 2±S
R 100°F
Figure 1. Flow diagram of two-stage regeneration of CaSO^ to CaO:
COAL ACCEPTOR
FUEL
FEED
HOPPERS
(2)
SOLIDS
FEEDERS
D-®
CARRIER
GAS
JAL AUUtriUK I.
I i i
•S AC5E,1TOR^-----T
FEED
HOPPERS
(2)
MOISTURE
CONDENSER!
•-COOLING WATER
DUST
RECEIVER
REACTOR
FLUIDIZING
GAS "^
CONDENSATE
RECEIVER WITH Err:
SIGHT GLASS
ACCEPTOR FINES
UNBURNED CARBON
& FUEL ASH
•COOLING WATER
WATER
SIGHT
GLASS
BPCV T0 GAS
METERING
& SAMPLING
ACCEPTOR |
RECEIVER
USED
ACCEPTOR
Figure 2. Simplified flow diagram of equipment.
11-5-11
-------
"TRIMMING" AIR
FEED GAS
Gas
co2
CO
N2
H2
H20
so2
2
COS
H2S
Vol
15.
6.
64.
2.
6.
4.
%
30
37
44 ]
27
68
01
0.022
0.204
,
STAGE
1
900 °F
(REDUCTOR)
I
' '
STAGE
3
330 °F
STAGE
2
450 °F
n
TOTAL S RECOVERY
98.2%
TAIL GAS
TO FLUIDIZED BOILER
Gas
H2S
SO,
Vol. %
0.068
0.034
11-5-12
Figure 3. Sulfur recovery from off gases - - two-stage regeneration of
-------
6. THE FLUIDISED-BED
DESULPHURISING GASIFIER
GERRY MOSS
Esso Petrolleum Company, Ltd.
England
ABSTRACT
Explanations are offered of the optimum
sulphur-absorption temperatures which are
observed when fuel oil is combusted in
fluidised beds of lime particles, under both
oxidising and reducing conditions. In the
gasifying case, information is given concerning
the effect on desulphurising efficiency of
variations in stoichiometric ratio, bed replace-
ment rate, and the mean particle size of the
bed material.
Consideration is given to some of the
control problems which arise when
continuous gasification is attempted; pre-
ferred solutions are described.
INTRODUCTION
In a paper presented at the First Inter-
national Conference on Fluidised-Bed Com-
bustion it was shown that sulphur-containing
oils may be desulphurised during combustion
within a fluidised bed of lime particles under
both oxidising and reducing conditions. In the
first case the sulphur is fixed as CaSC«4; in the
second case it is fixed as CaS. It was reported
that in both cases the optimum absorption
temperature was in the region of 850°C. It
had also been found that regeneration of the
reacted bed material was possible in both
cases and that high concentrations of SO 2
could be obtained at temperatures in the
region of 1050°C, by reducing CaSO4 and by
oxidising CaS. Several possible applications of
this principle were described and some experi-
mental results were presented relating largely
to gasifying conditions.
This paper reviews the progress which has
been made in this field since then at the Esso
Research Centre at Abingdon.
OPTIMUM TEMPERATURE FOR SULPHUR
RETENTION
Thermodynamic considerations indicate
that sulphur should be fixed efficiently by
lime under both oxidising and reducing condi-
tions at temperatures up to 1100°C. Re-
generation is accounted for by the decomposi-
tion of calcium sulphite, which is produced
(in one case) by the reduction of calcium
sulphate, and (in the other case) by the oxida-
tion of calcium sulphide. Figure 1 shows the
effect of temperature on the equilibrium
partial pressure of SO2, for the reaction of
oxygen with calcium sulphide; in other words,
the SO 2 equilibrium concentration with
CaSO3- It is quite easy to obtain this curve
experimentally and the important point
which emerges is that very high SO2 con-
centrations may be obtained at temperatures
above 1000°C.
Even at 865°C, the equilibrium SO2
concentration is as high as would be obtained
by burning a 4 percent by wt sulphur fuel oil
with 5 percent excess air. It follows from this
that, under oxidising conditions at tempera-
tures above about 850°C, sulphur cannot be
acquired by lime via calcium sulphite but
must be fixed as calcium sulphate by direct
II-6-1
-------
reaction of calcium oxide with sulphur
trioxide.
Figure 2 shows the pronounced optimum
temperature which is obtained when sulphur
is fixed in a fluidised bed of lime under
oxidising conditions as calcium sulphate. It is
important to note that this curve relates to
lime in the size range 100-200 microns which
was 31 percent reacted to calcium sulphate.
When a fresh bed was used, the initial
absorption efficiency was in the region of
90-95 percent over the whole temperature
range of the experiment. It follows from this
that we are dealing.with a question of kinetics
rather than thermodynamics; indications are
that the reaction is ash-diffusion limited. Even
so some further explanation is required to
account for the optimum temperature.
As we have already seen, although there are
two possible routes to the formation of
calcium sulphate, one of them (via calcium
sulphite) may be ruled out on thermodynamic
grounds at temperatures much higher than
850°C. This leaves the direct reaction
between 803 and lime to be considered.
Figure 3 shows the equilibrium partial pres-
sure of 863 in the combustion products of 4
percent S fuel oil burned with 5 percent
excess air, as a function of temperature.
Although the formation of 863 is favoured
by low temperatures, the rate of formation at
low temperature can be extremely slow in the
absence of a catalyst. It follows from this that
the rate of formation of 863 and, con-
sequently, its concentration could well reach
a maximum level in the region of 850°C.
Under ash-diffusion controlled conditions this
could result in the optimum SC>2 absorption
temperature which has been observed.
Turning now to gasifying conditions, there
is in this case a very much simpler explanation
of the optimum absorption temperature. It
will be appreciated that even under gasifying
conditions, the air which enters the bed
contains its normal quota of oxygen. Con-
sequently regenerating conditions can occur
II-6-2
immediately adjacent to the distributor. As
we have seen regeneration is favoured by high
temperatures. Therefore it appears that the
fall off in the absorption efficiency of the
gasifier at high temperatures is due to an in-
creased tendency to regeneration at the dis-
tributor, with reabsorption of SCH higher in
the bed resulting in an internal refluxing of
SO-; within the fluidised bed. Table 1 shows
the concentrations of SCb in parts per million
which may be observed in gas samples drawn
from within the bed at various heights above
the distributor. The first set of readings relate
to a temperature of 850°C; the second, to a
temperature of 950°C. It is clear from these
results that the sulphur content of the lime in
a bed sample can, under some conditions,
vary according to the position within the bed
from which the sample is taken.
Table 1. CONCENTRATION OF SO2 AT STATED
HEIGHTS ABOVE DISTRIBUTOR8 AT
30 PERCENT STOICHIOMETRIC AIR
Height, in.
1/2
4
8
At 850° C, ppm
20
10
10
At 950° C, ppm
1000
50
20
Stone sulphur content 6 percent by weight.
INFLUENCES OF OPERATIONAL VARIA-
BLES ON GASIFIER PERFORMANCE
Stoichiometric Ratio
i
At the first Conference, information w/as
presented concerning the effect of
Stoichiometric ratio on the desulphurising
efficiency of a gasifier. At that time the
gasifier was being operated with some degree
of pre-combustion; about 25-35 percent of
the incoming air was being bumed before
reaching the distributor. This procedure
results in a bed of uniform sulphur content at
-------
temperatures in the region of 850UC and
enables the sulphur balance to be checked by
stone analysis. Since then the apparatus has
been operated without underfiring and it has
been found that this method of operation is
markedly superior.
Figure 4 shows the effect of changing
stoichiometric ratios on desulphurising ef-
ficiency, with and without underfiring, with 4
percent by wt of sulphur in the 800-micron
mean-particle-size bed material. It can be seen
that, in the absence of underfiring, the desul
phurising efficiency is maintained even at a
stoichiometric ratio of 20 percent; whereas,
when the bed is underfired, there is a marked
drop in desulphurising efficiency at stoichio-
metric ratios lower than 30 percent. Further-
more the results which were obtained without
underfiring were obtained at a superficial gas
velocity of 4 ft/sec, as against the 3 ft/sec
used with underfiring. This enhanced per-
formance naturally has a considerable impact
on the size of the gasifier and therefore on the
investment required for a given fuel through-
put. The reason for the poor performance in
the presence of underfiring is not entirely
clear, but the particles get very sooty under
these conditions at low stoichiometric ratios
whereas, in the absence of underfiring, they
remain clean.
Bed Material Replacement Rate
A very important factor bearing on the
running costs of the desulphurising gasifier is
the rate at which the bed material has to be
replaced in order to maintain desulphurising
efficiency. In order to obtain information on
this subject, we ran several series of cyclic
tests each at a different bed replacement rate,
both with and without underfiring. The
results of just one of these test series are
shown in Figure 5. In this case the bed
replacement rate was 0.5 wt of lime per unit
weight of sulphur; i.e., about a third of the
stoichiometric requirement. It can be seen
that under these conditions the desulphurising
efficiency rapidly bottomed out at about 60
percent. The effect of varying bed replace-
ment rates on desulphurising efficiency can be
seen in Figure 6. Although these results were
obtained using underfiring, the efficiencies
measured without underfiring fall on the same
line. It is clear from this curve that virtually
100-percent desulphurising efficiency is
obtainable with a bed replacement rate which
meets the stoichiometric requirement of the
sulphur; i.e., about 1.75 wts of lime per unit
weight of sulphur. This result relates to a
maximum stone sulphur content of
approximately 4 percent by weight.
Particle Size
The effect of particle size on desulphurising
efficiency has also been examined. No dis-
cernible difference in performance was found
when the same stone was compared at mean
particle sizes of 810 and 635 microns.
Indications are that, due to the highly porous
nature of the lime particles, the internal sur-
face contributes most of the reactive surface
area; consequently, there is little to be gained
by making the mean particle size smaller than
the diameter at which bed containment
becomes a problem.
An interesting feature of the system is its
ability to fix vanadium as well as sulphur.
Vanadium may be fixed with virtually 100-
percent efficiency under both oxidising and
reducing conditions.
DESIGN OF A CONTINUOUSLY
OPERATING GASIFIER
With the feasibility of the desulphurising
gasifier established, consideration was given to
the design problems posed by a practical
system; an experimental unit is currently
being constructed which will deliver 7 MM
Btu of desulphurised fuel per hour on a
continuous basis.
II-6-3
-------
Figure 7 is a schematic flow diagram for a
continuous unit. On the left-hand side is the
gasifying bed in which lime is converted into
calcium sulphide; on the right-hand side is the
regenerator in which calcium sulphide is
oxidised by air to lime and SO2- bed material
is circulated between the two units. If the
gasifier is to function efficiently, it is neces-
sary to maintain the temperature of the
gasifying bed in the region of 850°C and the
temperature of the regenerating bed in the
region of 1050°C. Should the temperature of
the regenerating bed fall below 800°C, re-
carbonation will occur, lowering the tempera-
ture of the regenerator. If the gasifying
temperature rises above 900°C, the desul-
phurising efficiency will be reduced. In the
case of the regenerator, too low a temperature
will result in the formation of calcium
sulphate and a buildup of sulphur in the bed
material; too high a temperature may tend to
deactivate the lime.
A method of controlling the temperatures
of the two beds has been chosen which
involves no heat losses and no heat-transfer
surfaces. In the case of the gasifying bed, it is
proposed to control the temperature by
recycling some of the flue gas formed by
complete combustion back to the inlet of the
air blower. This necessitates a separate air
supply for the regenerator. In the case of the
regenerator, the temperature will be
controlled by varying the circulation rate of
the bed material. The rate of heat release in
the regenerator is determined by the rate at
which air is supplied: it will normally match
the rate at which sulphur is acquired by the
gasifier. The effect of increasing the flow rate
of bed material will be to reduce its sulphur
content and, hence, the amount of heat
generated within the regenerator per unit
weight of lime. Since the lime enters the re-
generator at a temperature of 850°C and is
heated to a temperature of 1050°C, there
must clearly be a specific bed transfer rate for
any particular set of operating conditions at
which the regenerator temperature is in equi-
librium at the desired level. It follows from
this that an accurate and reliable method for
controlling bed transfer rates is an essential
requirement for successful operation.
The method which has been chosen is
gravitational dense phase transfer. Each bed
container is provided with a catchment
pocket slightly above the surface of the bed;
each of these pockets communicates via an
almost vertical duct with the side wall of its
neighbor at a point slightly above the level of
the distributor. There is a short horizontal
section at the bottom of each duct so that, in
the absence of any external stimulus, the two
transfer ducts simply fill with static bed
material. An arrangement of this type is
shown in Figure 8 which is a schematic layout
of the first bed transfer test rig. Here you can
see the two transfer ducts and their horizontal
delivering sections. Each horizontal con-
nection has a suitable perforated tube running
parallel to its axis. When gas is introduced via
one of these tubes, the particles in the
horizontal duct are fluidised and expelled;
they are replaced by gravitational flow down
the vertical duct. As soon as the gas ceases to
flow, so do the solids. By pulsing the gas flow
and varying the frequency of the pulses, it is
possible to closely control the rate at which
bed material is transferred. Very little gas is
required to operate the system. In the original
test rig it was possible to shift about 2 pounds
of bed material per cubic foot of gas. In view
of the possible variation of sulphur content
throughout the depth of the bed, the transfer
of material from the top of each bed to the
bottom of the other is advantageous from an
operational point of view. The dense packing
in each duct, together with the continuous gas
bleed which is required to prevent blockage of
the perforations in the activating rtubes,
ensures that the leakage of gas between
compartments is minimal.
The experimental gasifier has been designed
as a monolithic structure in castable refrac-
tory which is insulated, from the walls of the
mild steel casing which contains it, by a layer
of vermiculite fondu. The overall dimensions
II-6-4
-------
of the unit are about 4 feet square by 8 feet
high. In order to check the operation of the
bed transfer system and the cyclones, the full-
scale cold rig (shown in Figure 9) has been
built in mild steel and perspex. The large tank
represents the gasifying compartment: the
central cylinder flanked by the two cyclones
represents the regenerator. The perspex ducts
allow observations to be made of the flow of
bed material which can of course be measured
by weight. The bed material which is being
used in these tests is crushed brick with the
same size distribution and bulk density as the
lime. The results which have been obtained
have been used as a basis for the experimental
gasifier which is now being constructed as
part of a contract with NAPCA for develop-
ment of the chemically active fluid-bed
gasifier.
II-6-5
-------
CM
o
0.5
0.4
0.3
0.2
0.1
EQUILIBRIUM SO,
MAXIMUM S02 LEVEL
ATTAINABLE USING AIR
70
3?
Q
LU
CD
DC
oc
D
a.
_i
CO
60
50
900
1100
1000
TEMPERATURE, °C
Figure 1. Effect of temperature on the
equiliDnum 862 partial pressures for the re-
action of oxygen with calcium sulphide.
X
CO
5
CO
LU
OC
a>
c/>
LU
DC
a.
800
950
850 900
TEMPERATURE, °C
Figure 2. Optimum temperature for
sulphur oxide absorption by lime (31%
of bed reacted).
oc
O
DC
D
OL
_i
CO
100
80
60
40
20
5% BY WT S IN BED —
• WITHOUT UNDERFIRING
(Gas velocity 4 ft/sec)
O WITH UNDERFIRING —
(Gas velocity 3 ft/sec)
I I I
600 800 1000 1200
TEMPERATURE, °C
1400
15 20 25 30 35
STOICHIOMETRIC AIR, %
40
Figure 3. Equilibrium partial pressure of S03
in combustion products of 4% S fuel oil
burned with 5% excess air.
Figure 4. Stoichiometric air rate with and
without underfiring.
o
z
LU
LL
U.
LU
•^
O
t
or
m
oc
I
i
100
80
60
40
20
0
*U 1 1 1
-V.^
•
—
_
1 1 1
4 P 12
I I !
•• •
• •»
—
—
I I I
16 20 24
<*
EMOVED,
oc
oc
i
D
CO
_l
LU
2
100
80
60
40
20
TEST CONDITIONS
JLPHUR LOADING 0.03 Ib/lb BED/hr
AIR RATE 33%OFSTOICHIOMFTRIC
GAS VELOCITY 2.5 ft/sec
FEED PART. SIZE 850 - 1200'u_
BED TEMP 800 - 900 °C
I
I
I
RUN NUMBER
Figure 5. Bed ageing effect replacement rate
0.5 wt CaO/wt S.
1.0 2.0
WT CaO PER WT SULPHUR
Figure 6. Effect of makeup rate on
desulphurising efficiency.
II-6-6
-------
SULPHUR-FREE SULPHUR DIOXIDE
FUEL GAS RICH STREAM
t !
FUEL >*•
Ca S
t
Ca O (LI ML)
Ca S
SOLIDS
TRANSFER
ra n
Ca S
!
Ua u
A ,>
AIR taJ ».
GASIFIER
REGENERATOR
Figure 7. Chemically active fluid-bed gasifier.
CYCLONE OUTLET
CYCLONE OUTLET
P : PULSED-AIR FEEDER
V : VIBRATORY FEEDER
Al R SUPPLY-
Figure 8. Test rig for bed-transfer trials.
11-6-7
-------
SESSION HI:
Gasification to Desulfurize Coal
SESSION CHAIRMAN :
Mr. J. W. Eckerd, USBM, Morgantown
-------
1. PRODUCTION OF LOW-SULFUR BOILER FUEL
BY TWO-STAGE COMBUSTION —
APPLICATION OF CO2 ACCEPTOR PROCESS
G. P. CURRAN, C. E. FINK, AND
E. GORIN
Consolidation Coal Company
ABSTRACT
A modification of the CC>2 acceptor proc-
ess is presented as a method for the produc-
tion of low-sulfur boiler fuels; i.e., low-sulfur
char and/or low-sulfur producer gas. Partic-
ular emphasis is given to total gasification of
Eastern coals for the production of low-sulfur
producer gas, which could be utilized for the
production of clean power in new combined-
cycle plants. Experimental data is given on
acceptor life under projected operating
conditions
INTRODUCTION
Considerable interest is developing in the
production of clean power from coal by
means of a combined-cycle power plant. The
first step necessary in such a plant is the,
generation of high-pressure producer gas by
the gasification of coal. Recently1 construc-
tion of a 170,000 KW plant of this type in
Germany was announced by Steinkohlen-
Electrizitat AG. The plant will adapt the well
known Lurgi pressure gasification process to
production of high-pressure producer gas.
Surprisingly, although no process is provided
for desulfurization of the gas, the incentive
for construction of the plant is stated1 to be
its lower investment cost in comparison with
that of a conventional power plant. If the
economics can be confirmed, the combined-
cycle plant opens up the prospect of pro-
ducing clean power from coal at a possible
lower cost in new plants than in conventional
plants equipped with stack gas scrubbing
equipment.
The most critical factor in the development
of such a new power cycle is the availability
of a satisfactory process for the production of
low-sulfur, high-pressure producer gas. The
Lurgi process, which must include facilities
for desulfurization of the gas, is the only such
process commercially available today. Eastern
bituminous coals, however, for the most part,
offer difficulties as feedstocks to the Lurgi
process because of coking problems. Further-
more, it would be desirable to purify and
desulfurize the gas while it is hot, as part of
the gasification process, in order to improve
thermal efficiency of the cycle. High thermal
efficiency in the combined-cycle plant can
only be obtained if the gas can be cleaned and
desulfurized hot.
The CC>2 acceptor process is now under-
going intensive development2 as a means of
producing high-Btu pipeline gas from coal. It
is the purpose of this paper to show that
many of the features of this process and the
experience gained in its development can be
utilized to speed the development of an anal-
ogous process for the production of clean,
high-pressure producer gas.
A contract, with the same title as this
paper, has been signed with NAPCA to accel-
erate research on this application of the CC«2
acceptor process. The work reported here
III-l-l
-------
concerns both some precontract work carried
out by Consolidation Coal Research and some
initial results of work carried out under the
contract.
DESCRIPTION OF THE PROCESS
The proposed process is in reality a two-
stage, air-blown, fluidized gas producer. The
simple air-blown, pressurized, fluidized gas
producer itself is not a developed process;
many problems have been associated with its
development. These potentially can be over-
come by use of the present process. The
advantages of the present process are:
1. In situ desulfurization of the generated
gases.
2. Improved efficiency of carbon utili-
zation.
3. Improved operability by control of the
ash-fusion problem.
The first advantage stems from the well
known sulfur-acceptor properties of the lime
acceptor. The second and third advantages
arise because of the two-stage nature of the
process. The final combustion of carbon is
conducted, separate from the gasification
reaction, in the presence of the endothermic
lime calcining reaction.
Figure 1 shows one version of the process
presently being analyzed as part of the feasi-
bility study under the NAPCA contract. The
coal is preoxidized in a separate air-blown
fluidized vessel operated at 750-800°F.
A computer program, consistent with
thermodynamic restraints imposed on the
process, gives the heat and material balance of
the system illustrated in Figure 1. The pro-
gram is broad enough to encompass variations
in the process; e.g., producing low-sulfur char
as a co-product with the low-sulfur producer
gas. The char-to-gas ratio can be varied over
wide limits.
Table 1 is a condensed version of the heat
and material balance for one-spot condition.
III-1-2
This corresponds to the situation where
complete conversion to producer gas is
effected; i.e., no char byproduct occurs. The
heat and material balance relationships given
for the gasifier correspond to gasification of
65 percent of the fixed carbon in the feed
coal.
The process consists of two main process
steps— gasification and acceptor regeneration-
connected in series. The gasifier is operated at
15-20 atmospheres (atm) pressure at
1 700-1 775°F. Preoxidized coal is fed to the
gasifier along with excess air. The char bed is
fluidized by steam and recycle gas from the
regenerator operation after it has been proc-
essed for sulfur recovery in the modified
Claus plant that is described later. Calcined
dolomite from the regenerator is fed to the
top of the gasifier.
The endothermic heat of gasification in the
gasifier is supplied, both by the partial com-
bustion of carbon with air, and by the exo-
thermic heat of absorption of the CC"2 by the
calcined dolomite.
It is necessary, in order to supply chemical
reaction heat to the gasifier, that a driving
force for absorption of CC>2 by the lime be
present. Because the acceptor falls through
the fluidized char in a "showering" action, it
is only necessary that the driving force be
present in the bottom part of the bed. At the
gasifier temperature chosen in the Table 1
illustration, 1700°F, such a driving force for
CC>2 does exist at the bottom of the gasifier:
PCCH = 2-9 atm versus the equilibrium value
of 1 .6 atm.
Another essential feature of the gasifier
operation is simultaneous desulfurization of
the producer gas in situ, via the acceptor
reaction:
CaO + H2S = CaS + H2O
(1)
-------
Table 1. SUMMARIZED HEAT AND MATERIAL BALANCE3
Material
Input
Coal
C
H (as H2)
O
S
Ash, Ib
Steam
Coal moisture
Dry air
Moisture
Claus plant tail gas
CO
CO2
N2
H2O
H2
Acceptor
MgO-CaO
Output
Char
C
H
Ash, Ib
Gas
CH4
H2
CO
C02
H20
N2
NH3
H2S
Acceptor*3
MgO-CaO
MgO-CaCO3
MgO-CaS
Preoxidizer and producer
°F
60
1200
60
398
398
1200
1885
1700
1700
1700
mol
—
5.812
2.381
0.475
0.1341
12.30
3.745
0.354
7.174
0.031
0.105
1.958
4.284
0.075
0.002
7.0861
1.223
0.054
12.30
0.277
3.288
3.523
1.956
2.617
9.951
0.047
0.0050
(0.0303) 6.0318
(0.0045) 0.8904
(0.0006) 0.1285
Material
Char
C
H
Ash (Ib)
Dry air
Moisture
Lift gas
CO
CO2
N2
H2O
Acceptor
MgO-CaO
MgO-Ca03
MgO-CaS
MgCO3-CaCO3
(make-up)
Char
C
Ash (Ib)
Gas
CO
C02
N2
H20
H2
S02
S2
H2S
COS
Acceptor
MgO-CaO
Regenerator
°F
1700
398
398
280
1700
1885
1885
1885
mol
—
1.223
0.054
12.30
5.422
0.023
0.011
0.124
0.281
0.002
6.0318
0.8904
0.1285
0.0354
0.122
12.30
0.172
2.024
4.565
0.075
0.004
0.0300
0.0487
0.0005
0.0006
7.0861
a Basis: 100 Ib dry coal
15 atm system pressure
14 atm clean gas delivery pressure
b Numbers in parentheses are for acceptor discarded in order to maintain acceptor activity.
Gross heating value of cold gas _. _„.
Cold gas effaency = Gross heatin^va,ue of coa, feed = 74'7/0'
III-1-3
-------
The gas composition given in Table 1 pre-
sumes that equilibrium is established in reac-
tion (1). The potential extent of desulfuriza-
tion under the operating conditions given in
Table 1 is thus 96 percent; i.e., the sulfur in
the producer gas product is only 4 percent of
that in the feed coal.
Acceptor regeneration is the second prin-
cipal process step. The acceptor is regenerated
by burning fuel char in an air-fluidized bed.
The fuel char is the gasification residue and its
combustion provides the heat to carry out the
endothermic calcination reaction:
CaCC>3 = CaO + CO2
(2)
The regenerator thus serves as a receptacle for
gasification residue that can be efficiently
used as fuel. This procedure significantly
improves the total carbon gasification effi-
ciency over that of an air-blown fluidized
producer.
Another feature of the regenerator opera-
tion is that combustion of char is conducted
concurrently with the endothermic calcina-
tion reaction (2). This feature minimizes the
difficulties due to ash fusion that are usually
encountered in combustion of char in flui-
dized beds and that are particularly severe
when the operation is conducted under
pressure.
Another important process that occurs in
the regenerator is rejection of sulfur via the
overall roasting reaction:
CaS + 3/2 O2 = CaO + SC>2 (3)
Reaction (3) is in reality the composite of
two separate reactions:
CaS + 2 O2 = CaSC-4
3CaSO4 =
(4)
(5)
The chemistry of the sulfur cycle in the
regenerator2 c has been extensively studied in
IH-1-4
work done for the Office of Coal Research in
development of the CO2 acceptor process.
The material balance of Table 1 provides
for a driving force sufficient to eliminate SO2
via reaction (5); i.e., PSO? = 0-065 atm versus
the equilibrium value of PSO-> = 0-090 atm at
the regenerator operating temperature of
1885°F This driving force is sustained by
operating at slightly reduced conditions in
order to generate enough CO to reduce SO2
by the reaction:
2CO + SO2= 1/2S2 + 2CO2
(6)
Note that more than half the sulfur issuing
from the regenerator is in the form of ele-
mental sulfur.
There appears to be a theoretical difficulty
in this method of operation; i.e., at the CO
levels used, calcium sulfate would not seem to
be present to effect reaction (5) because it
would be reduced by the reaction:
CaSO4 + 4 CO = CaS •*• 4 CO2
(7)
In practice, however, an environment of
oxidation sufficient to form CaSO4 is present
throughout most of the regenerator bed, and
reaction (7) occurs too slowly to reduce all of
the calcium sulfate at the bed outlet, perhaps
because of a slow rate of diffusion of CO into
the acceptor pores. Verification of this
method of operation is described later in this
paper.
Sufficient CO is also present in the regener-
ator off gases to permit reduction of the SO2
to elemental sulfur by a modified multistage
Claus process. The process is essentially' the
same as that described in another paper4
presented to this conference. Table 1 gives the
composition of the Claus plant tail gas.
Important considerations in the develop-
ment of a process of this type are acceptor
life and cost. Some information has been
developed experimentally under conditions
-------
similar to those proposed for this process:
however, more is required. The data available
to date will be presented later in the paper.
Table 2 gives the estimated acceptor makeup
rate and cost under the conditions proposed
for the process. The makeup cost is approxi-
mately 1.3^/MM Btu of product gas.
Table 2. DOLOMITE MAKEUP COST AND SULFUR
REJECTION
Makeup rate, mol MgCO3'CaCO3/mol
MgO-CaO circulated, % 0.5
Lb raw dolomite/100 Ib dry-coal feed 7.41
Makeup cost, il\ O6 Btua at $3.25/ton 1.27
Sulfur distribution
Product gas, %
Acceptor, %
Spent acceptor composition
CaO, mol, %
CaC03
CaS
3.7
96.3
85.55
12.63
1.82
a Gross heat of combustion of product gas.
The literature does not provide suitable
kinetic data to properly size and define the
optimum operating conditions in the gasifier.
Some limited integral rate data is presented
later in this paper; however, more is required
to further develop the process.
EXPERIMENTAL PROCEDURE
The equipment and operating procedure
have both been fully described in the report
to the Office of Coal Research.20 Briefly, the
gasifier was a 70-in. long, thin-walled vessel,
with a 4-in. ID. The fluidized bed height was
controlled at 40-in. by an overflow weir. The
vessel was contained, along with the electrical
heaters and insulation, inside a pressure shell.
CC>2 was used as pressure-balancing gas. The
38-in. long regenerator was of similar con-
struction, with a 2-in. ID; the fluidized-bed
height was controlled at 22 in. by the over-
flow weir. These two reactors were designed
to operate at temperatures up to 1950°F and
pressures up to 300 psig. All vessels and
internals were Type 310 stainless steel except
the regenerator thermowell, which was Hastel-
loy X. The acceptor, sized to 16 x 28 mesh,
was dolomite from the Tymochtee formation
in Western Ohio.
Calibrated rotary feeders controlled the
flow rates of three of the solids: (1) gasifier
feedstock (char or coal), (2) regenerator fuel
char, and (3) acceptor feed to regenerator.
These solids were fed from dual lockhoppers
through pneumatic transfer lines. The accep-
tor circulated continuously between the two
vessels. Acceptor flowed from the regenerator
to the gasifier by gravity, through a water-
cooled standleg. The acceptor particles
showered down through the fluidized bed of
char and segregated as a separate fluidized bed
in the 2-1/2 in. ID "boot" of the gasifier. The
acceptor was then withdrawn through a
water-cooled standleg and a rotary device
similar to the solids feeders, and was dis-
charged periodically from dual lockhoppers
situated below the rotary device.
A fixed investory of acceptor, correspond-
ing to 18 pounds of raw dolomite, was used in
each run. No fresh makeup was added. The
acceptor discharged from the bottom of the
gasifier was sampled periodically; its activity
was determined by a special assay. Thus, the
relationship between activity and the number
of calcining-recarbonation cycles undergone
by the acceptor could be determined.
In the regenerator, the heat required to
calcine the acceptor was supplied by the
combustion of char, which was partially gasi-
fied material from a previous run. The fuel-
char ash was elutriated from the fluidized bed
of acceptor and was removed by an external
cyclone.
The inlet gas to both vessels contained
recycled product gas, which was used to
III-1-5
-------
simulate the partial pressures that would exist
near the top of the tall fluidized beds in the
commercial process. Hydrogen and CC>2 were
added to the gasifier recycle gas to adjust the
exit partial pressures to the desired values.
This inlet dry gas was then passed through a
saturator in which the water temperature was
controlled to give the desired inlet-steam
partial pressure. The regenerator inlet gas was
primarily recycled gas and air. CC>2 or N2 was
added to adjust the outlet CC>2 partial pres-
sure to the desired value. The product gas
from both vessels was analyzed frequently by
gas chromatography.
RESULTS AND DISCUSSION
The experiments reported here were carried
out under conditions similar to those for the
gasifier of the process under discussion, al-
though they were originally designed for
investigation of the process as a "CO maker."
The acceptor life and integral rate data gen-
erated are therefore similar, but not identical,
to what might be expected in practice with
the process described above. In general, con-
ditions (from the point of view of acceptor
life) were more severe than those projected
for the process under discussion; i.e., higher
temperatures were used in the gasifier and
regenerator, and lower steam pressures were
used in the gasifier.
Tables 3 and 4 show the conditions and
results from the runs made during the pro-
gram. Table 3 shows the conditions and
results of Run Kl, which was made: (1) to
demonstrate the operability of the process,
(2) to produce fuel char, and (3) to obtain
some practical gasification kinetics data. The
conditions and results for a typical acceptor
life study run are shown in Tables 3 and 4 for
the gasifier and regenerator, respectively.
Table 3. GASIFIER RUN CONDITIONS AND RESULTS
Cycle
System pressure, atm
Acceptor
Size, mesh
Activity
Circulation rate, Ib/hr (raw basis)
Gasifier
Temperature, ° F
Feedstock
Feed rate, Ib/hr (MF basis)
Product char rate, Ib/hr
Recycle fluidizing gas, scfh
Inlet steam, scfh
H2, scfh
C02, scfh
N2, scfh
Total inlet gas, scfh
Inlet fluidizing velocity, ft/sec
Acceptor layer
Bottom of char bed
Run No.
K1
18
—
—
—
—
1770
Colstrip 1 1
Subbituminous
Dried at 400° F
5.74
3.20
161.4
2.5
0
51.2
0.5
215.6
—
0.165
L2
15
18
Tymochtee-6 Dolomite
16x28
0.39
8.18
1775
Disco char
Carbonized
Dried at 950° F
2.95
1.86
232.3
26.5
0
54.3
7.2
320.3
0.540
0.246
III-1-6
-------
Table 3 (continued) GASIFIER RUN CONDITIONS AND RESULTS
Inlet gas partial pressures, atm
H20
H2
CH4
CO
CO2
N2
Others
Product gas, scfh
N2 purges, scfh
Condensate, scfh
Total outlet gas, scfh
Outlet fluidizing velocity, ft/sec
Outlet gas partial pressures, atm
H2O
H2
CH4
CO
C02
N2
Others
Product gas composition, mol %
H2
CH4
CO
CO2
N2
Others
Product char particle density, Ib/ft3
Char bed density, Ib/ft3
Char bed weight, Ib
Retention times
Solids, hr
Vapor, sec
^ ... . Ib carbon gasified ...4
(Ib C in bed x mm)
Carbon burnout, %
Equilibrium CO2 partial pressure, atm
C02 partial pressure driving force, atm
Char product
H,wt%
C
N
O
S
Ash (diff.)
Run No.
K1
0.21
2.09
0.56
7.86
6.57
0.68
0.03
116.6
4.9
20.4
293.4
0.225
1.25
2.63
0.70
9.86
2.95
0.50
0.11
15.54
4.14
58.37
17.01
4.75
0.19
52.7
25.0
6.21
1.94
15.1
43.8
32.2
2.53
—
0.35
83.23
0.41
0
1.00
15.01
L2
1.49
1.24
0.08
6.01
6.24
2.90
0.08
103.7
12.7
27.3
352.5
0.271
1.39
1.63
0.11
7.89
4.23
2.64
0.11
9.51
0.65
45.98
24.37
19.15
0.34
71.8
27.4
6.87
3.69
11.4
25.7
36.2
2.62
1.61
0.30
76.17
0.64
0
0.29
22.60
III-1-7
-------
Table 4. REGENERATOR RUN CONDITIONS AND RESULTS
Run No. L2
Cycle
System pressure, atm
Acceptor
Size, mesh
Circulation rate, Ib/hr (raw basis)
Activity
Regenerator
Temperature, °F
Fuel char
Size, mesh
Feed rate, Ib/hr
Overhead solids rate, Ib/hr
Recycle gas, scfh
Inlet gas
N2, scfh
CO2, scfh
Air, scfh
Total inlet gas, scfh
Inlet fluidizing velocity, ft/sec
Inlet gas partial pressure, atm
02
CO
C02
N2
Outlet gas
Product gas, scfh
N2 purges, scfh
Total outlet gas, scfh
Outlet gas partial pressures, atm
02
CO
CO2
N2
Recycle gas composition, mol %
02
CO
C02
15
18
Tymochtee dolomite
16x28
8.18
0.39
1940
Colstrip
Subbituminous char
28x150
1.58
0.82
182
45
28
100
355
1.15
1.06
0.44
3.73
12.77
197
15
364
0
0.89
4.69
12.42
0
4.75
25.01
70.24
III-1-8
-------
Table 4 (continued). REGENERATOR RUN CONDITIONS AND RESULTS
Equilibrium CO2 partial pressure, atm
C02 partial pressure driving force, atm
Inlet
Outlet
Run duration, hours
Run duration, cycles
Lb/fuel char fed during run
Carbon burnout, %
Overhead material
H, wt %
C
N
0
S
Ash (diff.)
Input char
H, wt %
C
N
0
S
Ash (diff.)
Run No. L2
7.60
3.87
2.91
26.5
25.3
41.9
63.2
0.29
54.97
3.39
41.35
0.62
77.55
0.37
0.0
0.64
21.44
Operability
No serious operability problems were en-
countered in using the CC>2 acceptor process
to make CO, using sub-bituminous coal as
gasifier feed-stock. In Run Kl, raw Colstrip
sub-bituminous coal that had been dried at
400°F was fed to the gasifier at 1770°F: no
coking occurred. The char product particles
were not rounded or swollen. In contrast,
slight preoxidation (4 percent) of the sub-
bituminous coal was required to avoid coking
in the hydrodevolatilizer of the conventional
CO2 acceptor process, where H2 partial pres-
sure was much higher and CC>2 partial
pressure was much lower.
However, the process was not operable when
preoxidized Pittsburgh seam bituminous coal
was used as feedstock.
Two attempts were made to feed Ireland
coal that had been preoxidized to levels of 5.5
and 13 percent, respectively, into a bed of
Disco char: no massive coking occurred with
either feedstock, but the coal became plastic
as it entered the bed, and cemented together
the char particles. The agglomerates that were
formed, although tenuous, plugged the boot
and forced a shutdown.
In the actual process as proposed (Figure
1), coking is prevented by feeding the preoxi-
dized coal into the gasifier entrained in a
stream of gas containing free oxygen, allowing
additional pretreatment to take place at the
point of introduction into the gasifier. The
efficacy of this method was demonstrated on
a 30 ton-per-day pilot-scale unit at the Con-
solidation Coal Co. Research Center in 1950.
III-1-9
-------
During Run Kl, where the vapor retention
time was 15 seconds, there was no overt tarry
material in the gasifier product gas; however,
there was some residual naphthalene. Most of
the naphthalene condensed as fog and was
collected in glass-wool filters installed down-
stream of the primary and secondary product
gas coolers. Some of the naphthalene con-
densed in the primary cooler.
No deposits of any kind appeared in the
gasifier during Run Kl. In the acceptor life
study runs where Disco char was used as feed-
stock, however, slight gasifier deposits
occurred, as they had in the earlier work at
1600°F.
Gasification Kinetics
The gasifier was operated with recycled
product gas to simulate an upper section of a
tall fluidized bed. The feedstocks were either
raw coal or partially devolatilized Disco char
(produced from Pittsburgh seam coal). Thus,
the measured masification rates, shown in
Table 2, include carbon-bearing gases formed
during pyrolysis and devolatilization. Com-
parison of the carbon burnout for Run Kl
with that for an earlier run, in which Colstrip
coal was fed to the conventional CC>2 gasifica-
tion process,2 c shows that at the CC>2 and CO
partial pressures used, essentially no gasifica-
tion of fixed carbon occurred in Run Kl and
that the yield and product distribution
represent only the pyrolysis and devolatiliza-
tion reactions at the modified process con-
ditions.
The relatively poor kinetics in Run Kl is
apparently due to very strong inhibition by
CO at the high CO/CO2 ratios employed in
this run. The equilibrium constant in the
"Boudouard" reaction is:
(Pco)2
= K= 100
at the conditions used. The Kexp at the outlet
conditions of Run Kl is about 33. It is clear,
therefore, that total inhibition by close
approach to equilibrium is not in itself the
reason for the poor kinetics.
The integral rates observed with Disco char
are representative of what may be achieved
with bituminous coal or char feedstocks.
They are marginally acceptable for the proc-
ess under discussion. Because the steam
partial pressure would be higher in the pro-
ducer-gas operation, higher gasification rates
are anticipated.
Operation of the Regenerator
The regenerator was operated at a nominal
level of 4.5 percent CO in the outlet gas to
ensure that no difficulty would be caused by
the CaS-CaSO4 transient liquid. To achieve
this conservative CO level with fuel chars
having a wide range-of particle diameter—28 x
150 mesh versus 48 x 65 mesh used in earlier
work —a higher carbon investory was
required in the regenerator bed. Therefore,
the carbon burnout was lower than that at
comparable bed expansions in the earlier
work.
Char Desulfurization and Sulfur Cycle
In Run Kl the total sulfur rejected from
the Colstrip coal in the gasifier was 35 per-
cent. This low level is typical of the conven-
tional CO2 acceptor process using low-ranking
Western coals. The high lime content of the
ash in these coals causes "self-acceptance,"
with the result that most of the sulfur is
carried to the regenerator in the form of CaS.
In the life study runs with Disco char, which
has low lime content in the ash, sulfur rejec-
tion was at the high level of 83-84 percent.
The sulfur content of the product char was
only 0.29 wt. percent versus 1.16 wt. percent
in the feed. The above desulfurization was
effected with a hydrogen partial pressure of
IH-1-10
-------
1.6 atm. The results are roughly in accord
with prior work3 on desulfurization of low-
temperature chars from bituminous coals.
The sulfur content of the gasifier off gases
was not determined. However, if equilibrium
were approached in the acceptor reaction,
CaO + H2S = CaS +
the H2S content of the gas would have been
about 0.02 mol percent.
In the life study runs, the rejected sulfur
was equivalent to the conversion to CaS of
1.8 mol percent of the CaO content of the
acceptor on each pass through the gasifier, a
level considerably higher than in any run
during the conventional CC»2 acceptor process
studies. Analyses of the acceptor, sampled
periodically from the gasifier, showed that all
of the sulfur had been rejected in the regen-
erator, even at the relatively high level (4.5
percent) of CO in the exit gas.
Acceptor Activity
Figure 2 shows acceptor activity, measured
preiodically during the runs, as a function of
the number of calcining-recarbonation cycles.
The activity is defined as the mol percent of
CaO that is active as a CO2 acceptor in the
recirculating inventory. Equilibrium activities
(R) calculated from the data in Figure 2, are*
shown below as a function of fresh acceptor
makeup rate. The method for calculating the
equilibrium activity has been given pre-
viously.20
Makeup, % R
1
2
3
0.143
0.244
0.320
Note that the figure used for acceptor
makeup in Table 2 was only half of the
figures above, derived from the experimental
data presented here. The lower value is justi-
fied by the fact that the process is projected
to operate under milder conditions than those
under which the above data was obtained.
Correlations developed in prior work20 on
development of the CO2 acceptor process
were used to justify the lower makeup cost.
The relative influences of gasifier and
regenerator bed temperatures, of steam and
H2 partial pressures, and of the extent of
conversion of CaS on acceptor activity remain
to be determined over wider ranges of these
variables than were used in the brief study
reported here.
BIBLIOGRAPHY
1. Rudolph, P.F.H. Coal Combustion for
Present and Future Power Cycles. (Pre-
prints of Division of Fuel Chemistry,
ACS.) Presented at Toronto, May 23,
1970.
2. Office of Coal Research, U.S. Department
of the Interior. Research and Develop-
ment Report No. 16:
a. Interim Report No. 1 — Feasibility
Study.
b. Interim Report No. 2 — Low-Sulfur
Boiler Fuel via Consol CO2
Acceptor Process.
c. Interim Report No. 3 — Bench Scale
Research on CO2 Acceptor
Process—Books 1, 2, and 3.
d. Interim Report No. 4 - Current
Commercial Economics.
3. Batchelor, J.D. and E Gorin. Desulfuriza-
tion of Low-Temperature Char. Ind. Eng.
Chem. 52: 162. 1960.
4. Curran, G.P., C.E. Fink, and E. Gorin.
Coal-Based Sulfur Recovery Cycle in
Fluidized-Lime-Bed Combustion. Paper
presented at the Second International
Conference on Fluidized-Bed Combus-
tion, Hueston Woods, Ohio, October 6,
1970.
III-l-ll
-------
PLANT
STEAM
GASIFIER PLANT
STEAM STEAM
TO COMBINED
STEAM AND GAS-*
TURBINE CYCLES
AIR
Figure 1. Simplified flow diagram of clean gas at 14 atm pressure.
1.0
0.9
0.8
0.7
\v
% 0.5
tr
O
LU
O
O
0.4
0.3
0.2
0.1
0
0
10
15 20 25
NUMBER OF CYCLES
30
35
40
Figure 2. Acceptor activity versus number of cycles.
IH-1-12
-------
2. COAL DESULFURIZATION
ASPECTS OF
THE HYGAS PROCESS
H. L. FELDKIRCHNER AND
F. C. SCHORA, JR.
Institute of Gas Technology
INTRODUCTION
One obvious alternative method for avoid-
ing air pollution in the generation of energy
from coal is to convert the coal to gas, de-
sulfurize the gas, and utilize the low-
pollution, sulfur-free gas to produce energy.
Among the gasification processes now under
study is the Hygas process being developed at
the Institute of Gas Technology (IGT) under
the joint sponsorship of the U.S. gas industry
(via the American Gas Association) and the
U.S. Government (via the U.S. Department of
the Interior, Office of Coal Research). The
former is seeking to secure a supplemental
supply of gas equivalent to natural gas so that
it will be assured of a future gas supply; the
latter is vitally concerned with ensuring the
availability of adequate energy supply and
stimulating the utilization of our large coal
reserves. This process not only converts coal
to gas, but produces byproduct sulfur in
elemental form.
The Hygas process has been under develop-
ment at IGT since 1946. Until 1964, this
work was supported almost entirely by the
American Gas Association (AGA), at which
time the U.S. Department of the Interior
joined with the AGA in developing the proc-
ess concept via a large-scale pilot-plant pro-
gram. This pilot plant, now in the final con-
struction and shakedown phase of develop-
ment, is designed to produce about 1.5
million cu ft/day of gas from 80 tons of coal.
It is hoped that an 80-million cu ft/day
demonstration plant will be completed by
1974. Work, some of it still in progress, has
been done in several smaller development
units.
PROCESS DESCRIPTION
The Hygas process has been discussed in
detail in a number of recent publications.6 -9
Figure 1 gives the basic steps. This process is
operable with high- as well as low-rank coals,
at a thermal efficiency of about 70 percent.
However, it is at present necessary to pretreat
agglomerating coals, such as the Eastern
bituminous coals, before hydrogasification.
Pretreating ensures that the coal will not
agglomerate and hinder solids flow in the
hydrogasifier. Nonagglomerating coals, such
as those in the lignite and subbituminous
class, do not require pretreatment. A substan-
tial fraction of the sulfur may be removed,
from some high sulfur coals, by pretreatment.
Hydrogasification is carried out in a two-
stage fluidized-bed reactor. As now planned,
coal or char is fed to the gasifier in a light oil
slurry to avoid the use of lock hoppers. Small
quantities of light oils are formed in the
process; they vaporize in a drying zone at the
top of the reactor, leave the reactor along
with the raw product gas, and are recycled for
reuse. Temperatures in the hydrogasifier range
from about 1300°F to about 1800°F; reactor
III-2-1
-------
Table 1. TYPICAL ANALYSES OF FEEDSTOCKS
Coal rank
Mine
County
State
Coals (before pretreatment)
Ultimate analysis, wt%
Carbon
Hydrogen
Nitrogen
Oxygen
Sulfur
Ash
Total
Proximate analysis, wt %
Moisture
Volatile matter
Fixed carbon
Ash
Total
Chars (after pretreatment)
Ultimate analysis, wt %
Carbon
Hydrogen
Nitrogen
Oxygen
Sulfur
Ash
Total
Proximate analysis, wt %
Moisture
Volatile matter
Fixed carbon
Ash
Total
Coal seam
Pitts. No.8
HVBA
Ireland
Marshall
W. Va.
67.90
4.91
1.18
6.38
4.38
15.30
100.00
3.9
32.6
48.8
14.7
100.0
69.90
3.09
1.46
10.78
3.33
11.44
100.00
1.4
24.2
63.1
11.3
100.0
Ohio No.6
HVBB
Broken Aro
Coshocton
Ohio
74.40
5.64
1.23
9.64
3.36
5.73
100.00
2.1
39.2
53.1
5.6
100.0
75.00
4.18
1.30
10.80
2.31
6.41
100.00
1.4
26.8
65.8
6.0
100.0
III. No.6
HVBC
Crown
Montgomery
Illinois
70-10
4.88
1.10
9.82
4.00
10.10
100.00
13.3
33.0
44.9
8.8
100.0
70.30
3.58
1.20
10.11
3.33
11.48
100.00
1.4
22.7
64.6
11.3
100.0
W. Va.
Ind. No.6
HVBC
Minnehaha
Sullivan
Indiana
73.20
5.23
1.29
8.73
2.34
9.21
100.00
12.0
32.7
47.2
8.1
100.0
73.00
3.43
1.55
10.11
2.03
9.88
100.00
0.5
22.7
67.0
9.8
100.0
Block 5
HVBA
Kanawha
Kanawha
W. Va.
79.00
5.84
1.34
6.04
0.80
6.98
100.00
1.5
35.7
55.9
6.9
100.0
76.10
3.97
1.51
9.96
0.77
7.69
100.00
0.6
23.2
68.6
7.6
100.0
111-2-2
-------
USED IN HYGAS PROGRAM
Coal seam
Sewell
MVB
Dante
Nicholas
W. Va.
83.20
5.04
1.52
5.58
0.62
4.04
100.00
1.4
29.4
65.2
4.0
100.0
80.20
3.65
1.59
9.85
0.59
4.12
100.00
1.8
19.0
75.2
4.0
100.0
Sewell
LVB
Lockgelly No. 2
Fayette
W. Va.
87.40
4.85
1.52
2.86
0.60
2.77
100.00
1.1
20.6
75.6
2.7
100.0
85.10
3.32
1.49
7.84
0.58
2.67
100.00
1.0
16.2
80.2
2.6
100.0
Pocahontas No .4
LVB
Stotesbury No. 10
Raleigh
W. Va.
85.10
4.20
1.12
2.81
0.73
6.04
100.00
0.7
16.1
77.2
6.0
100.0
85.30
2.23
0.84
5.24
0.59
5.80
100.00
1.2
14.0
79.1
5.7
100.0
Laramie
No. 3
Subbit.
Eagle
Weld
Colorado
75.10
2.56
1.33
15.90
0.38
4.73
100.00
3.7
35.2
56.5
4.6
100.0
73.10
4.27
1.40
15.62
0.27
5.34
100-00
0.3
31.1
63.3
5.3
100.0
—
Subbit.
Colstrip
—
Montana
67.9
4.52
1.01
17.86
0.81
7.90
100.00
3.0
36.6
52.7
7.7
100.0
—
—
—
—
—
__
—
—
Lignite
Glenharold
Mercer
N. Dakota
66.0
4.37
0.56
20.00
0.99
8.08
100.00
5.0
40.8
46.5
7.7
100.0
—
_•_
—
—
—
—
—
Lignite
Savage
Sidney
Montana
64.8
4.17
0.95
21.22
0.68
8.18
100.00
4.3
39.3
48.6
7.8
Tf)Q&
—
—
—
—
—
—
III-2-3
-------
pressures are up to about 1500 psig. Nor-
mally, sulfur removal is greatest in this proc-
essing step.
After the two stages of hydrogasification,
the partially gasified char is gasified further
with steam at 1800-1900° F in an electro-
thermal gasification stage. The hydrogen and
carbon oxides produced in the electrothermal
gasifier are fed directly to the hydrogasifier.
The char from the electrothermal gasifier is
then used as fuel for conventional power and
steam generation. The sulfur content of this
spent char is normally minimal.
The raw hydrogasifier effluent gases are
water-quenched and the light oil and excess
water are removed along with ammonia and
coal fines. Next, the gas passes through an
acid-gas-removal system that removes carbon
dioxide and sulfur compounds primarily
hydrogen sulfide. After a final caustic and
water wash, the gas is catalytically meth-
anated to reduce the carbon monoxide
content of the gas to less than 0.1 mole per-
cent. The final gas has a sulfur level well
below 1 ppm, a dew point of about -40°F,
and a heating value of over 950 Btu/scf.
SULFUR REMOVAL
As was mentioned earlier, most of the
sulfur is removed from the coal or char in
three of the Hygas process steps: (1) coal
pretreatment, (2) hydrogasification, and (3)
electrothermal gasification. These steps are
discussed in greater detail below.
Information on the amounts and types of
sulfur compounds removed in these three
stages is somewhat limited at this time. This
limited information is the result of the
research program .on the Hygas process having
been directed primarily toward developing a
process for producing high-Btu gas from coal,
with emphasis on hydrogasification, electro-
thermal gasification, and pretreatment—in
that order. Sulfur removal aspects have been
considered, but to a lesser degree. In addition,
a wide range of coal feeds has been studied,
which limits the amount of data on any one
coal and makes it difficult to draw conclu-
sions for some feedstocks.
Table 1 lists the coals that have been used
as feedstocks thus far in this program. Most of
the work has been with the Pittsburgh No. 8
seam, HVBA char; consequently, more is
known about sulfur removal from this feed.
For example, over 60 pilot-plant hydrogasifi-
cation tests have been conducted on the
Ireland mine coal, whereas only a few have
been conducted with Illinois No. 6 and Ohio
No. 6 coals.
Table 2 lists the pyritic and organic sulfur
contents reported12 by the U.S. Bureau of
Mines for each type of coal studied. This data
should only be considered as approximate,
however, because we did not analyze in our
laboratories samples of coals tested in the
development unit.
Pretreatment
Agglomerating coals are pretreated by
making them contact air in a fluidized bed at
about 800°F. This process has been described
in detail in earlier papers.1'3 The sulfur
removed is primarily as sulfur oxides, al-
though there are undoubtedly small amounts
of organic sulfur and other sulfur compounds
present. The pretreater off gas could be
treated with one of the many developing
sulfur-oxide-removal processes to recover
sulfur, or it could, after concentration of
SO2, be combined with the hydrogen-sulfide-
containing gases removed from the hydro-
gasifier effluent and sent to a Claus plan>t for
elemental sulfur production. The large pilot
plant now being completed will contain a
Claus-type plant.
Figure 2 shows the results of typical pro-
treatment tests for the coals listed in Tables 1
and 2. Although only a limited amount of
III-2-4
-------
data is available and the data exhibits con-
siderable scatter, it is clear that sulfur removal
in pretreatment increases with increases in
total sulfur content. Accurate measurement
of the fraction of sulfur removed is difficult.
especially for low-sulfur coals, because of the
small differences in the sulfur contents of
most of the coals and chars. Inaccuracies in
material balances, variations in pretreatment
levels, and sampling errors all contribute to
the overall error in this measurement.
Figure 3 shows the effect of organic sulfur
content of the coal on sulfur removal.
Apparently pyritic sulfur is more easily
removed from the bituminous coals than
organic sulfur. The higher sulfur removal
measured for the lower rank subbituminous
coal is not considered significant in view of its
very low total sulfur content.
We expect that more information will be
obtained from the results of the operation of
the large pilot plant now being completed. In
general, however, we can expect to remove
from 20 to 40 percent of the sulfur present in
bituminous coals in the pretreatment step.
Table 2. TYPICAL SULFUR BREAKDOWNS RE-
PORTED FOR TYPES OF COALS TESTED8
Coal
Pittsburgh No. 8
Ohio No. 6
Illinois No. 6
Indiana No. 6
W. Va. Block 5
W. Va. Sewell
W. Va. Sewell
Pocahontas No. 4
Laramie No. 3
Montana
N. Dakota
Montana
Type
HVBA
HVBB
HVBC
HVBC
HVBA
MVB
LVB
LVB
Subbit.
Subbit.
Lignite
Lignite
Pyritic
sulfur
wt%b
2.20
1.80
2.84
1.15
0.05
0.06
0.04
0.09
0.06
0.39
0.02
0.37
Organic
sulfur
wt%b
2.27
1.61
2.17
0.87
0.55
0.51
0.50
0.51
0.22
0.41
0.85
0.36
Total
sulfur
wt%b
4.48
3.48
5.09
2.06
0.61
0.57
0.55
0.60
0.28
0.82
0.88
0.77
aReported by U. S. Bureau of Mines, 1C 8301 -12
bDry basis.
H y drogasif ication
More data is available on sulfur removal
during hydrogasification than during pretreat-
ment because this program has been con-
cerned primarily with the hydrogasification
step. Details on this step have been published
previously.7-8 When coals and chars are
hydrogasified with either hydrogen-steam
mixtures or synthesis gas, sulfur is removed
primarily as hydrogen sulfide and carbonyl
sulfide, which can be removed from the
hydrogasifier effluent by a number of well-
developed processes. One such approach, for
example, is to scrub out r^S, COS, and €03
by the hot carbonate process and then con-
vert the H2S to elemental sulfur in a Clans
plant, as was discussed above. COS is hy-
drolyzed to H2S in the gas-cleanup operation.
The Claus plant feed could contain (on a dry
basis) from about 3 mole percent (for lignite
feeds) to about 32 mole percent (for bitu-
minous coal chars and high-hydrogen-content
feed gases to the hydrogasification section), if
the pretreater off gas were combined with the
H2$, and CO2 were removed from the
hydrogasifier effluent.
Figure 4 shows that, as in the pretreatment
test data, considerable scatter exists in the
sulfur removal data for hydrogasification. The
major trend observed is that the degree of
char sulfur removal varies directly with the
MAF char-gasification level. In contrast to the
pretreatment results, sulfur is removed more
readily from chars from coals with the highest
organic-sulfur contents. The lignite feeds,
which were not pretreated and which hydro-
gasify much more readily than higher rank
feeds, gave results comparable with the other
feeds. In addition, the results with North
Dakota lignite compare favorably with those
for Montana lignite, even though their
organic/pyritic sulfur ratios differ markedly.
The West Virginia coals, all of which have a
high organic/pyritic sulfur ratio, allowed even
higher sulfur removal. This is just the opposite
III-2-5
-------
of the pretreatment results, but is in agree-
ment with the coal hydrodesulfurization work
of Vestal and Johnston.1 J
The large amount of scatter in the data
shown in Figure 4 suggested that other vari-
ables (e.g., temperature, pressure, and gas
composition) might also be important. The
only approach available at that point seemed
to be regression analysis becuase the test pro-
gram to date had been concerned with hydro-
gasification variables, not with those variables
influencing sulfur removal. We attempted,
therefore, to correlate the results with
Ireland-mine coal char (the majority of testing
was done with this feed material) using mul-
tiple linear regression analysis.
There were three main sets of data: those
using hydrogen-steam feed-gas mixtures,
those using hydrogen-methane steam, and
those using synthesis gas. The results of the
regression analysis showed that the only
statistically significant variable for the syn-
thesis gas "and hydrogen-methane-steam feed-
gas runs was temperature. The hydrogen-
steam run results showed a dependence on
both MAF-char gasification and feed-gas/char
ratio. We then decided that these results were
subject to too great an experimental error for
conventional techniques. We did try to force-
fit the results by standard techniques; Figure
5 shows that the predicted results agreed, in
general, quite well with the experimental
data. It appears that further work will be
required to establish precisely the variables
affecting sulfur removal. It is possible,
however, to estimate sulfur removal from the
coals tested to within about ±10 percent.
Using this information, we can expect to
remove anywhere from 40 percent to essen-
tially all of the feed-char sulfur content in the
hydrogasification step.
Electrothermal Gasification
At the present time, less data is available
from our electrothermal gasification program.
Again, details of this program have been
III-2-6
published elsewhere.4-s Because the gas
produced in the electrothermal gasifier passes
directly into the hydrogasifier, it is important
to be able to predict its sulfur content. Any
sulfur compounds evolved here will ultimately
end up in the hydrogasifier effluent.
The data available correlates better for
these tests (Figure 6) than for those discussed
previously. It should be pointed out that the
variable is "char" gasified (in the electro-
thermal gasifier), not "MAF-char" gasified.
Also, the sulfur removal data is based on the
sulfur content of the feed to the electro-
thermal gasifier; this feed has had most of its
sulfur removed in the two previous processing
steps—pretreatment and hydrogasification.
The closer agreement of the data from the
various feeds tested may stem from their
having similar properties: they are all highly
gasified materials. Earlier work2 has shown
that the reactivities of coals, ranging in rank
from lignite to anthracite, approach one
another at high levels of gasification.
Economics of Sulfur Removal
Figure 7 shows a typical breakdown of
sulfur removal from the coal in each stage of
the Hygas process. It is difficult to extract the
true cost of sulfur removal from available data
because of the uncertainties in the future
value of sulfur. It is also difficult to separate
sulfur-removal costs from the total costs of
gas purification. However, if elemental sulfur
could be sold for $25/long ton, the Hygas
process economics might be more favorable
for higher sulfur coals; the cost of sulfur
removal would then be less than the sulfur
byproduct credit. For low-sulfur coals and a
lower sulfur price, the reverse would be true.
These figures assume a final gas cost of
5-6^/therm (based on AGA accounting
procedures), a coal cost of 16^/million Btu,
and present economic conditions in the
United States. A detailed study1 ° on sulfur
recovery in the manufacture of pipeline gas
from coal has recently been presented.
-------
BIBLIOGRAPHY
1. Channabasappa, K.C. and H.R. Linden.
Fluid-Bed Pretreatment of Bituminous
Coals and Lignite—Direct Hydrogenation
of Chars to Pipeline Gas. Ind. Eng. Chem.
50: 637-44, April 1958.
2. Feldkirchner, H.L. and H.R. Linden.
Reactivity of Coals in High-Pressure Gasi-
fication With Hydrogen and Steam. Amer.
Chem. Soc. Div. Fuel Chem. Preprints, p.
191-208, September 1962; also I&EC
Process Design Develop. 2: 153-62, April
1963.
3. Kavlick, V.J. and B.S. Lee. Coal Pretreat-
ment in Fluidized Bed. Amer. Chem. Soc.
Div. Fuel Chem. Preprints. 10: 131-39,
September 1966; also Advances in Chem-
istry Series No. 69, p. 8-17. American
Chemical Society, Washington, D.C.,
1967.
4. Kavlick, V.J. and B.S. Lee. High Pressure
Electrothermal Fluid-Bed Gasification of
Coal Char. Paper presented at 64th
National Meeting of the A.I.Ch.E., New
Orleans, March 16-20, 1969.
5. Kavlick, V.J., B.S. Lee and F.C. Schora.
Electrothermal Coal Char Gasification.
Paper presented at the Third Joint Meet-
ing of the I.I.Q.P. and the A.I.Ch.E., San
Juan, Puerto Rico, May 17-20, 1970.
6. Lee, B.S. Synthetic Pipeline Gas From
Coal by the HYGAS Process. Paper pre-
sented at the American Power Con-
ference, Chicago, April 21-23, 1970.
7. Pyrcioch, E.J. and H.R. Linden. Produc-
tion of Pipeline Gas by High-Pressure
Fluid-Bed Hydrogasification of Char.
Amer. Chem. Soc. Div. Gas Fuel Chem.
Preprints, p. 59-69, September 1969; also
as Pipeline Gas by High-Pressure Fluid-
Bed Hydrogasification of Char. Ind. Eng.
Chem. 52: 590-94, July 1960.
8. Pyrcioch, E.J., B.S. Lee and F.C. Schora.
Hydrogasification of Pretreated Coal for
Pipeline Gas Production. Amer. Chem.
Soc. Div. Fuel Chem. Preprints. 10:
206-23, September 1966; also Advances
in Chemistry Series No. 69, 104-27.
American Chemical Society, Washington,
D.C., 1967.
9. Schora, F.C. and B.S. Lee. Hydrogasifi-
cation Process. Paper presented at 65th
National Meeting of the A.I.Ch.E., Cleve-
land, May 4-7, 1969.
10. Tsaros, C.L., J.L. Arora, W.W. Bodle.
Sulfur Recovery in the Manufacture of
Pipeline Gas From Coal. Amer. Chem.
Soc. Div. Fuel Chem. Preprints. 13:
252-69, September 1969.
11. Vestal, M.L. and W.H. Johnston. Chem-
istry and Kinetics of the Hydrodesulfuri-
zation of Coal. Amer. Chem. Soc. Div.
Fuel Chem. Preprints. 14: 1-11, May
1970.
12. Walker, F.E. and F.E. Hartner. Forms of
Sulfur in U.S. Coals. U.S. Bureau of Mines
Information Circular 8301. Washington,
D.C., 1966.
III-2-7
-------
FUEL GAS
HYDRO-
GASIFIER
PRETREATER
RAW
HYDROGEN-
RICH
GAS
•.'••'.'.'•'•'"•'
. •...".
(
1^>
(HIGH-
BTU GAS
GAS PURIFICATION
AND METHANATION
STEAM
ELECTROTHERMAL
GASIFIER
ELECTRICITY
CHAR1
AIR:
,1,1,1
POWER
GENERATION
I'l'i' 1
ASH
50
40
30
DC
Q_
a
DC
D
Q
LU
o
cc
D
20
10
0
I f
Figure 1. Schematic flow sheet of Hygas process.
50
O BITUMINOUS COALS
* SUBBITUMINOUS COAL
012345
SULFUR CONTENT OF COAL FEED, wt % (dry basis)
Figure 2. Sulfur removal from coals during
pretreatment increases with increasing sulfur
content.
III-2-8
tr.
o.
cc
u.
40
30
cc
Q
O
£ 20
O
10
^*u**i 'v •' • '•; \ •'
^^ •'•
° BITUMINOUS COALS
. A SUBBITUMINOUS COAL
40 50 60 70 80 90
ORGANIC SULFUR CONTENT OF
COAL FEED, % of total sulfur
100
Figure 3. Sulfur removal from coals during
pretreatment decreases with increasing
organic sulfur content.
-------
100
OC
o:
u.
o
30
10 20 30 40 50 60
MAP CHAR GASIFIED, %
o PITTS NO. 8 HVAB CHAR OCOLORADO SUBBIT CHARS
AND COAL
• NORTH DAKOTA LIGNITE OMONTANA SUBBIT COAL
6OHIO NO. 6 HVBB CHAR AMONTANA LIGNITE
^7 INDIANA HVBC CHAR
DILLINOIS NO. 6 HVBC CHAR
OW. VA HVBA
MVB CHAR
LBV
Figure 4. Char sulfur removal increases
with char gasification in hydrogasifier.
Or
CC
D
O
100
90
80
70
60
50
40
20 30 40 50
CHAR GASIFIED, %
60
O IRELAND MINE HYDROGASIFIED CHAR
AIGT-PRETREATED CHAR
D INDIANA HYDROGASIFIED CHAR
tfHV BIT. HYDROGASIFIED CHAR
O FMC CHAR
Figure 6. Char sulfur removal in electro-
thermal gasifier increases with increasing
char gasification.
O
OC
OC
Q
LU
I
O
<
O
100
90
80
70
60
50
40
I
I
40 50 60 70 80 90 100
OBSERVED SULFUR REMOVAL, %
o HYDROGEN-STEAM
A SYNTHESIS GAS
O HYDROGEN-METHANE-STEAM
Figure 5. Agreement between measured
sulfur removal data and forced-fit data.
SULFUR
IN RAW
COAL
TO HYGAS
PROCESS
Figure 7. Typical sulfur removal in each
stage of the Hygas process.
HI-2-9
-------
3. STEAM-OXYGEN GASIFICATION
OF
VARIOUS U. S. COALS
A. J. FORNEY, S. J. GASIOR,
R. F. KENNY, AND W. P. HAYNES
U. S. Bureau of Mines
ABSTRACT
The Bruceton process for making a supple-
mental natural gas from coal consists of gasifi-
cation, purification, and methanation steps.
The advantages of the system are that caking
coals can be used in the fluid-bed gasifier, and
more than half of the methane can be made in
the gasifier instead of in the methanator.
Furthermore, the final product contains no
sulfur and has a heating value exceeding 900
Btu/cu ft.
INTRODUCTION
The need for a supplemental natural gas is
becoming more evident each year. The
reserves-to-production ratio, now about 14:,!,
has been declining. If the cost of natural gas
continues its present upward trend, a
substitute natural gas will be able to compete
with the natural product in the immediate
future.
One source of this substitute gas is to make
it from coal because of its widespread abun-
dance and low cost. The steam-oxygen proc-
ess for making supplemental gas from coal not
only makes the product inexpensively, but
also decreases the sulfur pollution to a tolera-
ble level. The process has been previously
reported in the literature.1 -2 -3
All of the work described here was done at
the Bruceton, Pa., laboratories of the Pitts-
burgh Energy Research Center of the Bureau
of Mines.
THE OVERALL PROCESS
The Bruceton process for making a supple-
mental natural gas is shown schematically in
Figure 1. The main components are: (1) the
gasifier; (2) the shift converter, which changes
the gas ratio from approximately IK^ ICO to
3H2: ICO; (3) the purification system, which
removes most of the CO2 and all of the sulfur
compounds from the gas; and (4) the catalytic
methanator, in which 3H2 + ICO reacts to
form CH4 + H2O. Because the shift and puri-
fication systems are commercial, research at
the Bureau is concentrated on the gasifica-
tion and the methanation steps. This paper
will concentrate on the gasification step; it
should be emphasized, however, that the gas
is 400-500 Btu/cu ft before methanation and
afterwards should be of pipeline quality—at
least 900 Btu/cu ft. Because of the nickel
catalyst used in the methanation step, the
sulfur level cannot exceed 1 ppm; therefore,
the final product may be considered sulfur-
free.
The gasifier is shown schematically in
Figure 2 and photographically in Figure 3.
III-3-1
-------
The Bructon gasifier combines the operations
of pretreatment, carbonization, and gasifica-
tion in one reactor. The coal, 70 percent
through 200 mesh, is dropped through the
pretreater with oxygen and steam (or CC>2) at
400°C, where it is rendered noncaking. The
decaked coal then falls into the carbonization
zone and finally into the gasification zone,
where it is gasified with steam plus oxygen at
900-1000°C. The gasifier is operated at 40
atmospheres pressure.
Advantages of the Bruceton gasifier are: (1)
it can handle both caking and noncaking
coals, (2) gases from pretreatment add to the
product gas from the gasifier, and (3) it makes
a purified product gas with a very high (20-30
percent) methane content (the balance is H2 -
CO).
This high methane content means the
volume of gas is less than if it were H2 + CO
only. Thus, the size of all process vessels and
lines downstream from the gasifier are re-
duced as much as 30-40 percent; the final
methane reactor is reduced about 50 percent
from the size it would need to be if only H2 +
CO were in the feed gas.
The Bruceton process has been estimated
to have an overall thermal efficiency of 63
percent. The selling price for the product gas
has been estimated at 54^/M cu ft, a price
that is competitive with that of other coal-to-
gas processes.4
Table 1 shows the results of selected tests
with four coals. The free-swelling index (FSI)
is used to indicate the caking or coking prop-
erty of each coal. Pittsburgh seam, a highly
caking coal, has a maximum FSI of 8 to 9;
Illinois No. 6, a weakly caking coal, has an
FSI of about 4.5; and North Dakota lignite
and Montana subbituminous coal have an FSI
of 0 (i.e., they are noncaking).
Table 1. RESULTS OF TESTS WITH VARIOUS COALS GASIFIED
AT 40 ATMOSPHERES
Test
Free-swelling index
Carbon conversion, %
Product gas, H2+CO
+CH4, scf/lb
MAF coal
CH4, scf/lb MAF
Gas analysis, %
H2
CO
CH4
CO2
Tar, % feed
Type coal
Pittsburgh
Seam
8 to 9
67
17
3.9
34
19
16
32
2.9
Illinois
No. 6
4.5
72
18
3.7
36
19
14
31
3.4
North
Dakota
Lignite
NCa
88
25
4.2
35
23
15
29
2.0
Montana
Subbit.
NCa
71
16
3.0
38
11
11
40
3.3
Goals
65
17
4.7
-
aNC = Noncaking.
III-3-2
-------
The carbon conversion for the tests shown
in Table 1 ranges from 67-88 percent; our
goal is 65 percent. This degree of conversion
permits the complete utilization of carbon in
the overall process because the char and tar
can be burned to raise the steam needed for
compression and other steps in the process.
Other goals required to make the process
economical are listed in Table 1. It is desirable
to make about 17 scf of product gas per Ib
MAP coal because this results in a final
high-Btu gas volume of 8 to 9 scf per Ib MAP
coal. We want to make 4.7 scf CH4 per Ib
MAP coal, but are slightly low on this goal.
The gas analyses are similar even though
different coals are used.
Table 2 shows that the gas leaving the gasi-
fier contains most of the sulfur in the coal.
Some sulfur is left in the char and some in the
tar; however, both will be burned.
It is desirable—even necessary-that the sul-
fur content- of both the char and tar be below
1 percent so they can be burned without dis-
charging unacceptable amounts of SOX into
the air. Table 3 shows that this limitation is
achieved with all coals tested except Illinois
No. 6. This coal, therefore, and similar high-
sulfur coals may present a problem. If the
carbon conversion is increased to lower the
Table 2. DISTRIBUTION OF SULFUR DURING GASIFICATION3
Sulfur distribution, Ib
Coal
Char
Tar
Gasb
Total S in gas, %
Pittsburgh
Seam
26.0
3.1
0.5
22.4
86
Type
Illinois
No. 6
78.0
11.2
1.1
65.7
84
coal
N. Dakota
Lignite
16.0
4.4
0.4
11.2
70
Montana
Subbit.
28.0
4.4
0.3
23.3
84
aBasis: per ton of coal feed to gasifier.
bBy difference.
Table 3. SULFUR CONTENT OF COALS AND PRODUCTS
Sulfur content, wt %
Coal
Char
Tar
Gas
Type coal
Pittsburgh
Seam
1.3
0.5
0.8
0
Illinois
No. 6
3.9
1.4
1.7
0
N. Dakota
Lignite
0.6
0.6
1.0
0
Montana
Subbit.
1.4
0.9
0.5
0
III-3-3
-------
sulfur in the char, the amount of char is
decreased. This means low-sulfur supple-
mental coal must be burned to raise the steam
required for the overall process. Another
possibility is that the coal could be washed to
achieve a lower level of sulfur before gasifica-
tion.
CONCLUSION
The high-Btu gas made in the Bruceton
process is low enough in sulfur that it causes
no problem with air pollution requirements.
The chars and tars from most coals can be
burned to raise steam and can be made so the
total sulfur in the feed to the power plant is
less than 1 percent. Some high-sulfur coals
may present a problem, however, because the
sulfur in the char and tar may be too high.
More research will be required to resolve this
problem.
BIBLIOGRAPHY
1. Field, J. H. and A. J. Forney. High-Btu Gas
via Fluid-Bed Gasification of Caking Coal
and Catalytic Methanation. Proceedings of
Synthetic Pipeline Gas Symposium.
American Gas Association, Pittsburgh, Pa.
p. 83-94, November 1966.
2. Forney, A. J., S. Katell, and W. L. Crentz.
High-Btu Gas From Coal via Gasification
and Catalytic Methanation. Proceedings of
American Power Conference. Chicago,
Illinois, April 1970.
3. Forney, A. J., S. J. Gasior, W. P. Haynes,
and S. Katell. A Process to Make High-Btu
Gas From Coal. BuMines Technical
Progress Report 24. 6 p. April 1970.
4. Tsaros, C. L. and T. J. Joyce. Comparative
Economics of Pipeline Gas frorri Coal
Processes. Proceedings of Second Synthetic
pipeline Gas Symposium. American Gas As-
sociation, Pittsburgh, Pa. p. 131-146,
November 1968.
HI-3-4
-------
COAL
STEAM f-*-|
OXYGEN *H
SPRAY
TOWER
SHIFT
CONVERTER
HOT
CARBONATE
SCRUBBER
GASIFIER
STEAM
OXYGEN
TAR
AND
/DUS
PIPELINE
GAS
RESIDUE
Hjs
cos
C02
FINE
SULFUR
REMOVAL
METHANATOR
Figure 1. Flow diagram of system used to make high-Btu gas from coal.
COAL FEED HOPPER
STEAM GENERATOR
PRETREATERr
WATER
OXYGEN
WATER
STEAM GENERATOR
GASIFIER
OXYGEN
OXYGEN CHROMATOGRAPH
ASH HOPPER' ' ANALYZER
Figure 2. Diagram of a 40-atmosphere fluid-bed gasifier.
III-3-5
-------
Figure 3. 40-atmosphere gasifier.
-------
4. DESULFURIZED FUEL
FROM COAL
BY INPLANT GASIFICATION
E. K. DIEHL AND R. A. GLENN
Bituminous Coal Research, Inc.
ABSTRACT
The growing demand for electrical power
will continue to draw on fossil fuels as a
source of energy. At the same time, the need
to control and reduce environmental pollu-
tion will dictate the consumption of fossil
fuels in a manner that will minimize their
contribution to atmospheric pollution.
For coal, a solution lies in the conversion
of the solid fuel into a clean, sulfur-free fuel
gas. This fuel gas can be subsequently utilized
in a variety of downstream systems, many of
which promise overall improvements over
conventional power-generating systems.
Bituminous Coal Research, Inc. (BCR) is
developing, for the Office of Coal Research,
U.S. Department of the Interior, a two-stage,
super-pressure, oxygen-blown coal gasifier for
the production of pipeline gas. Results
obtained thus far have led to the conclusion
that a similar air-blown system would be ad-
vantageous for the production of a fuel gas
suitable for desulfurization and pollution-free
combustion in power plants.
This paper describes the BCR gasifier and
suggests several alternate uses of the clean-fuel
gas so that a major reduction in atmospheric
pollution can be realized with continued use
of coal in the growing energy market.
INTRODUCTION
As a result of the increased concern over
environmental pollution, strict limitations and
control requirements are being imposed on
processes emitting atmospheric contaminants.
This is especially true of fossil-fuel-fired
electric generating stations that may emit fly
ash and/or sulfur oxides. Flue gases can be
processed to minimize fly ash emission at
reasonable costs. Currently, the development
of equally satisfactory methods for the
control of sulfur oxide emission is the
principal objective of major research programs
being conducted by both Government and
industry.
The electric power generating industry, in
particular, finds itself on the horns of a
dilemma. The demand for electricity within
our society continues to grow at a minimum
rate of 7 percent each year, with a doubling
time of about 10 years; in 1968 and 1969, the
growth rate was 9 percent. Generating
capacity to meet this demand is being
constructed, planned, and projected on a scale
never before experienced.
Although there is a wide difference of
opinion concerning how this mushrooming
generating capacity will be fueled, there is
general agreement that for the next several
decades coal will remain a major source of
energy fuel. Thus the dilemma for the
industry: huge coal resources will continue to
be tapped to keep pace with the electric
power demand, but environmental pollution
control will dictate their consumption in a
manner that will assure minimum atmospheric
contamination.
IH-4-1
-------
While there may still be some minor gains
to be realized in conventional fuel-to-
electricity cycles, research is rapidly shifting
toward the development of new uncon-
ventional concepts. Basic to the development
of these new concepts is the realization that
they must include features that reduce or
eliminate environmental pollution problems.
Further, new concepts, in order to be
accepted and adopted by the power genera-
tion industry, must be economically competi-
tive with conventional systems equipped with
pollution control devices.
In the recent years, terms such as magneto-
hydrodynamics, electrogasdynamics, ad-
vanced-cycle power systems, fuel cells, flu-
idized-bed combustion, and gasification have
become commonplace wherever technical
people gather to contemplate the future of
power generation. It is obvious that uncon-
ventional concepts are receiving serious atten-
tion, and that they will, in time, take their
place in the total power generation picture.
For coal-fired installations, a promising
approach is the conversion of the coal to
sulfur- and dust-free gas prior to its ultimate
use as an energy fuel. The careful selection of
specific operating parameters and the appli-
cation of new technology can lead to the
production of gas having a gross heating value
ranging from about 200 Btu/scf to as high as
950 Btu/scf.
Such a system holds promise as a means of
generating power not only without air pollu-
tion, but also with decreased overall cost,
especially if recovered sulfur can be sold as a
credit against the cost of the process. In addi-
tion, a reduction of nitric oxide emission is
conceivable because of the more precise
combustion control possible with gaseous
fuels.
GAS PRODUCERS
The only commercial fixed-bed process for
the gasification of coal at elevated pressure
111-4-2
available today fulfills the requirement for
gasifying coal with high thermal efficiency.
However, this process is not suitable for the
gasification of highly caking coals and is
limited to the use of lump fuel. The permissi-
ble gas-flow velocity limits the size of the
individual gasifiers, as well, and for a large
utility power plant, costly multiple units
would be required. The production of tar
inherent in this process also adds to the
process cost because purification of both the
liquid and the gaseous effluents then becomes
necessary.
Coal-gasification processes have also been
developed, using fluidized beds. The Winkler
process, operating at atmospheric pressure,
has been used in large units burning predomi-
nantly brown coal. Operation of the process
at elevated pressure has been carried out in
pilot plants only, and with only noncaking
fuels. Fluidized-bed gasification processes
leave a residue with a high carbon content
that requires separate utilization and thus
complicates the process economics. The
operability of fluidized beds at elevated
pressure using caking coals has not, moreover,
been established. This system requires devel-
opment work on a large scale, and may im-
pose limitations on such a process.
The simplest process is the gasification of
coal in the entrained state. Very large units
can be built and complete carbon conversion
assured by operating under slagging condi-
tions. These processes use coal at any rank,
regardless of caking properties. Many com-
mercial plants, operating at atmospheric
pressure, have been built to use peat, brown
coal, and bituminous coals as raw material.
Operability of this type of process at elevated
pressure has been proved in pilot plants.
BCR Two-Stage Gasifier
BCR is developing, under sponsorship of
the Office of Coal Research, a two-stage
process for the gasification of coal for the
production of high-Btu pipeline gas.1 -2 »3 The
-------
process combines the advantages-high gas
yield and low exit temperature-of the more
complex fixed- and fluidized-bed processes
with the simplicity of entrained processes and
their ability to use any coal.
The basic component of the process under
development by BCR is the two-stage gasifier
(Figure 1). Fresh coal and steam are intro-
duced into the upper section (Stage 2) of the
gasifier, where the fresh coal is contacted and
entrained by a rising stream of hot synthesis
gas produced in the lower section (Stage 1).
The fresh coal is rapidly heated and partially
gasified to methane and additional synthesis
gas in Stage 2. Residual char, swept out of the
gasifier along with the raw gas, is separated,
then returned to Stage 1. Here the char is
completely gasified under slagging conditions
to produce both the synthesis gas and the
heat required in Stage 2v
For the production of pipeline gas, Stage 1
is blown with oxygen and steam, and the
product gas from Stage 2 undergoes purifica-
tion, partial shift reaction, and catalytic
methanation to raise the heating value above
900 Btu-/scf. The gasifier is operated at about
lOOOpsi.
Experimental results obtained so far in the
development of the two-stage, super-pressure
process have led to the conclusion that, if
blown with air rather than with oxygen, the
two-stage principle would also be advanta-
geous for the production of a fuel gas suitable
for desulfurization and pollution-free combus-
tion in power plants. Thus, the air-blown
gasifier could be used to generate fuel gas
from which sulfur, as well as particulates,
could be removed by existing cleanup
systems. The resultant clean, sulfur-free gas
could then be utilized in a variety of power-
generating systems.
Air-Blown Gasifier
Using information generated in the pipeline
gas project, studies have been made to verify
the technical feasibility of using the two-stage
gasifier to supply a fuel gas. Figure 2 is one
example of a flow diagram and material
balance for a conceptual fuel-gas system. A
basis of 100 Ib of moisture and ash-free coal is
used for the material balance, the gasifier is
operated at 20 atmospheres, and the gasifier
outlet temperature is 1700° F.
Steam and air are preheated to 1700°F.
For the purpose of the example, it is assumed
that progress in materials development and
the small pressure differential between
products and feeds can make possible this
preheat step.
Table 1 gives analyses of the feed coal and
resultant fuel gas; Table 2 shows a simplified
heat balance.
Mathematical Model
The material and heat balances for the
example system were estimated using a
mathematical model developed at BCR.4
Figure 3 shows the general scheme of the
model. In Stage 1, all of the materials are
mixed thoroughly and the makeup char is
gasified. The resulting mixture comes to
water-gas shift equilibrium at the Stage 1
temperature that results from the Stage 1
energy balance based on an assigned heat loss.
This gas moves to Stage 2 and contacts the
coal, where the following reactions occur:
C+2H2-»CH4
C + H2O -»• CO + H2
(1)
(2)
In the gasifier model, several general assump-
tions were made pertaining to the thermo-
dynamics of the process:
1. The water-gas shift equilibrium ade-
quately describes the gas composition
from Stages 1 and 2.
2. The enthalpies of mixing of the gases are
zero.
3. All gas fugacities and activity coeffi-
cients are unity.
4. The effect of pressure on the heat of
reaction is minor.
III-4-3
-------
Table 1. ANALYSES OF COAL AND PRODUCT GAS
Coal analysis
Constituent
Carbon
Hydrogen
Oxygen
Nitrogen
Sulfur
Ash
Moisture
Total
H.V., gross
H.V., net
H.V., net,
MAP coal
Percent
70.1
4.9
6.6
1.4
3.0
12.7
1.3
100.0
1 2,800 Btu/lb
1 2,300 Btu/lb
14,350 Btu/lb
Fuel-gas analysis
Constituent
CH4
CO
C02
H2
H2O
N2
H2S
Total
H.V., gross, dry
Mo I percent
3.3
22.7
7.0
24.9
10.9
30.7
0.5
100.0
210 Btu/scf
Table 2. HEAT BALANCE, AIR-BLOWN GASIFIER3 Effects of Major Operating Variables
Heat input
Heat output
Constituent
Coal, H.V., net
Coal, sensible
heat
Steam
Air
Char
Total
MBtu
1435.2
22.8
87.0
90.8
34.4
1670.2
Constituent
Gas, H.V., net
Gas, sensible
heat
Slag
Char
Loss
Total
MBtu
1332.1
263.4
12.4
58.6
3.7
1670.2
aBasis: 100 Ib MAF coal; 20 atm; exit gas at 1700°F;
steam/coal ratio 1.06.
Kinetic equations for methane yield and
carbon oxide formation, together with
appropriate material and energy balances,
were programmed for a digital computer. The
resultant simulation was used to obtain both
the foregoing balances and the relationships
that follow.
III-4-4
Using the mathematical model, the system
was tested for the effects of changes in steam-
coal ratio, operating pressure, and gasifier
outlet temperature. The resulting influence of
these variables on the heating value of the
product gas, the producer-gas volume, and the
thermal efficiency of the system is shown in
Figures 4, 5, and 6.
The strongest effect is that of the gasifier
outlet temperature of the heating value of the
product gas. If a higher temperature is re-
quired by a certain fuel-utilization system, the
heating value of the gas can be increased by
changing one or more of the other operating
variables.
Although the system study presented here
is by no means exhaustive, it shows the opera-
tional flexibility of the gasifier. Thus, depend-
ing upon downstream use of the fuel, such
operating variables as steam-coal ratio,
-------
temperature, and pressure may be selected to
optimize costs, to provide for convenience of
operation, or to reflect downstream needs.
Sulfur Removal
Because sulfur appears in the gasifier
product gas as H2S, its removal is not a major
problem. Reduction of both gas temperature
and pressure, if required, can be accomplished
without undue penalty to the system. Such
processes as the Benfield and iron oxide fit
the system quite well.
Current Status
For the further development of an air-
blown two-stage gasifier for commercial-scale
use, the next step is considered to be the
operation of the system at the pilot-plant
level, A-technical evaluation and cost estimate
of a 5-ton/hr pilot plant was recently com-
pleted with the assistance of Koppers
Company, Inc., as an outside engineering
subcontractor. In addition to the basic gasi-
fier, the pilot plant would include coal prepa-
ration and handling facilities, gas desulfuriza-
tion units, effluent waste treatment, and
other general facilities needed for the conver-
sion of coal to sulfur-free, low-Btu fuel gas.
Cost evaluation is continuing. It has been
expanded in scope to include an estimate of
the cost of a commercial plant as well as tne
ultimate cost of the clean fuel gas.
FUEL GAS UTILIZATION
The clean, desulfurized, high-pressure fuel
gas could be utilized through several alternate
routes. Perhaps the least efficient use of the
gasifier product would be its combustion in a
conventional steam boiler. Although the ash-
free quality of the gas would eliminate the
need for boiler-cleaning equipment, and
freedom from sulfur would solve a major air
pollution problem, it is doubtful if these
benefits would ordinarily offset the coal con-
version process. It is conceivable, nevertheless,
that gasification of available coal could be the
best solution in a particularly vital energy
situation. Low-pressure gasification could, for
example, supply fuel gas directly to a con-
ventional boiler. If more efficient gasification
would be attained at higher pressures, the
gasifier product gas could be expanded
through a gas turbine before being burned.
A more efficient use of the gasifier product
would be its combustion in a supercharged
boiler. With the addition of a topping gas
turbine to utilize the work available in the
high-pressure flue gas, the system shows some
advantages over conventional operation. The
cycle efficiency can be improved if, in addi-
tion to driving a combustion air compressor,
the topping turbine is coupled to a generator.
While steam remains the major generating
fluid, the gas turbine can add to the net gen-
eration and at the same time eliminate other
fan power requirements.
Although present-day supercharged boilers
do not operate much above 5-7 atmospheres,
it is conceivable that higher pressures could be
utilized. Little economic gain, however, can
be realized above 7 atmospheres. Neverthe-
less, Figure 5 shows that the gasifier can be
operated at 5 atmospheres without appre-
ciably affecting gas quality or thermal effi-
ciency.
Combined Cycles
Over the past several years, many examples
of conceptual combined cycles have been
studied. A recent paper5 by Robson and
Giramonti evaluates a number of combined
cycles and establishes a timetable based on gas
turbine development. Known by various
names, most cycles have in common the
major generation of power by a gas turbine,
which then exhausts into a steam boiler at
essentially atmospheric pressure. Net genera-
tion includes that obtained from a steam
turbine.
I1I-4-5
-------
There are many variations on the com-
bined-cycle concept. Some publicized cycles
limit themselves to current commercial gas-
turbine inlet temperatures, approximately
1800°F. Others, however, expressing firm
belief in advancing gas turbine technology,
assume higher temperatures in order to
demonstrate the potential of the combined
cycle.
Direct coal-fired gas turbines, the object of
some past and current research, are faced with
a major problem area: erosion and corrosion
from coal ash and certain ash constitutents.
The coal gasifier, on the other hand, because
it produces a clean product gas, is an excellent
means for using coal as a source of energy for
the combined cycle.
Figure 7 shows a conceptual 1000-MW
combined-cycle power generating system.
This system, conceived6 by Curtiss-Wright
Corporation, uses the BCR two-stage gasifier
to generate fuel gas at 22 atmospheres and
2000°F. (Note that the cycle incorporates
liquid-metal heat exchange to reclaim the
sensible heat in the product gas before it is
cleaned.) Further, the gas turbine inlet
temperature is shown as 3196°F. While both
of these innovations are not considered to be
currently commercial, they are not believed
to be beyond grasp. Sufficient incentive for
their development may be fostered by the
promise that the combined cycle holds.
The Curtiss-Wright plan shows that a
system heat rate of 6740 Btu/KWH can be
obtained using 12,800-Btu/lb coal. Although
this gain in cycle efficiency may be somewhat
optimistic, because it assumes successful
development of the innovations mentioned
above, it does indicate that the combined-
cycle concept deserves further attention.
With gasification and gas cleanup at ele-
vated pressures, the size and possibly the cost
of the gas producer would be significantly less
than with atmospheric pressure systems. In
III-4-6
addition, the power cycle's increased effi-
ciency, coupled with the air pollution control
aspects of the system, should help to supply a
solution to the growing need for clean power.
Engineering development of the BCR two-
stage gasifier for the production of pipeline
gas is entering the pilot-plant stage. Current
projections indicate that a commercial plant
can become a reality in about 10 years.
In the meantime, it is believed that use of
the two-stage system as a producer-gas gen-
erator for inplant production of clean fuel gas
from coal can be realized earlier. Much will
depend upon the incentive provided by
current evaluation of the types of end uses
described.
New power-generation systems will be
developed to help meet the growing require-
ments of our energy-consuming way of life.
Coal, transformed into a clean, non-polluting
fuel gas, can continue to help fuel those new
systems.
ACKNOWLEDGEMENT
Parts of this paper are based on work
carried out at Bituminous Coal Research, Inc.,
with support from the Office of Coal Re-
search, U.S. Department of the Interior,
under Contract No. 14-01-0001-324.
The assistance of Dr. R. L. Zahradnik,
Carnegie-Mellon University, and of Dr. E. E.
Donath, Consultant, is gratefully acknowl-
edged.
BIBLIOGRAPHY
1. Donath, E. E. and R. A. Glenn. Pipeline
Gas From Coal by Two-Stage Entrained
Gasification Proc. AGA Production Con-
ference. Buffalo, New York, 1965.
2. Glenn, R. A. Progress in the Development
of the Two-Stage Super-Pressure Entrained
Gasifier. Am. Gas Assoc. Proc. Synthetic
-------
Pipeline Gas Symposium, New York: p.
67-75, 1966.
3. Glenn, R. A. and R. J. Grace. An Inter-
nally-Fired Process Development Unit for
Gasification of Coal Under Conditions
Simulating Stage Two of the BCR Two-
-Stage Super-Pressure Process. Presented at
the 1968 AGA Synthetic Pipeline Gas
Symposium, Pittsburgh, Pa., 1968.
4. Zahradnik, R. L. and R. A. Glenn. The
Direct Methanation of Coal. ACS Div. Fuel
Chem. Preprints 13 (4): 52-70, 1969.
5. Robson, F. L. and A. H. Giramonti.
Advanced-Cycle Power Systems Utilizing
Desulfurized Fuels. ACS Div. Fuel Chem.
Preprints 14(2): 79-96, 1970.
6. Curtiss-Wright Corporation. 1000 MW
Combined-Cycle Power Generating System.
(Curtiss-Wright brochure).
lll-4-l
-------
COAL
STEAM
STAGE 2
GASIFIER
CYCLONE
RECYCLE
SOLIDS
OXYGEN
AND
STEAM
GAS
PURIFICATION
AND
METHANATION
FINAL
^PIPELINE GAS
SLAG
Figure 1. Simplified flow diagram for two-stage super-pressure gasifier.
COAL
660 «F
100 Ib, MAF -^
14.8 IbASH
1.5 Ib H2O
GAS AIR-
*n/"\ft oc
1 /UU r
435 Ib
STAGE
2
N
s
*i r
HEAT
EXCHANGER
AIR
1700 °F
228 Ib
1^
— STEAM
^ 435 Ib. DRY GAS
| 6932 scf
v
STEAM
1700 °F
105 Ib
CHAR
1115 °F
SLAG
14 Ib
2924 °F
P = 20 atm
III-4-8
Figure 2. Material balance for air-blown gasifier.
-------
99
COAL-
STEAM •
AIR
97
PRODUCT GAS
STAGE 2
THE OUTLET STAGE 2 TEMPERA-
TURE IS CHOSEN.
KINETIC MODEL: METHANE AND
CARBON OXIDE YIELDS FOR
STAGE 2 ARE CALCULATED
FROM THE STAGE 1 PARTIAL
PRESSURES. WATER-GAS SHIFT
EQUILIBRIUM IS THEN ASSUMED
AT STAGE 2 OUTLET TEMPERA-
TURE.
< 0
-u. 95
ui
93
T830°F
20120F
7400
7300
CHAR
AND ASH
CO O
< o
STAGE 1
ALL MATERIALS MIX THOROUGH-
LY AND COME TO WATER-GAS
SHIFT EQUILIBRIUM. THE MIX-
ING TEMPERATURE RESULTS
FROM THE SIMULTANEOUS SOL-
UTION OF THE ENERGY BALANCE
AND THE EQUILIBRIUM REAC-
TION.
£•57100
g«
7000
6900
ASH
210
oc
O
180
1.0
1.2
.1.4
1.6
Figure 3.
model.
General scheme of mathematical
STEAM/COAL
Figure 4. Effect of steam/coal ratio
(PV20atm).
I1I-4-9
-------
99
97
X ii
Hi 95
93
_ STEAM/COAL^
7400i
6900
210
200
190
180
5 10 15 20
GASIFIER PRESSURE, atm
Figure 5. Effect of operating pressure
(T = 1830° C).
1700 1800 1900 2000 2100
GASIFIER OUTLET TEMPERATURE, °F
Figure 6. Effect of gasifier outlet
temperature (P = 20 atm).
III-4-10
-------
COAL-12,800 BTU/lb
265 tons/hr
GASIFIER
F : 2012
F=900
= 318.5
= 589
F : 975 F : 500
1 NO. 1 LIQUID MULTI- N0- 2 LIQUID
| WILT ALII LA [ » fN ONF<5 ^^ MLTALMLAT >
1 EXCHANGER EXCHANGER
\ /
/ *
1 LIQUID
1 METAL H EAT — ,
1 EXCHANGER
F = 1374 I ,
A'R W-U72
1 *
1
GAS UQUID F -
< METAL HEAT <
EXCHANGER
i
F r 1374 3600 PSI -
P = 297
W = 749
F : 1000
P : 3600
W = 459
COMBUSTOR -|Fs3196
F : 276 F i 230
HEAT
EXCHANGER
140 ,,
1000 *F - 1000 RH -
STEAM J
1" Hg
" TURBINE I 1 «.-.,.-,.,»,«,,
///J COND.
— 6-66^)-^
753 MW
F = 900
P = 318.5
GENERATOR
COMPRESSOR
P/P - 22.0
W = 2220
F = 1453
P = 16.26
W = 2371
GAS
TURBINE
AIR IN, F = 68
W = 2061
F=717
P s 15.60
F = 500
P - 15.07
F:350
P = 14.85
MAIN
BOILER
AUXILIARY
BOILER
EVAP-
ORATOR
STACK
WATER
COOLING STEAM F = 600
P : 400
753.0 MW GAS TURBINE
->• 253.6 MW STEAM TURBINE
F = TEMPERATURE, °F
P ; PRESSURE, psia
W = FLOW, Ib/sec
1006.6 MW TOTAL
Figure 7. 1000-MW combined-cycle power generating system.
-------
5. REMOVAL OF HYDROGEN SULFIDE
FROM SIMULATED PRODUCER GAS
AT ELEVATED TEMPERATURES
AND PRESSURES
F. G. SHULTZ
U. S. Bureau of Mines
INTRODUCTION
The Morgantown Energy Research Center
of the U. S. Bureau of Mines has been investi-
gating the possibility of removing hydrogen
sulfide from simulated producer gas near the
temperature at which the gas leaves the
producer. The temperature range of interest is
1000 - 1500°F. Experimental work has been
limited to these temperatures and to pressures
between near zero and 100 psig.
Background
We have previously reported1 that hydro-
gen sulfide can be removed from this type of
hot gas by passing the gas stream through an.
absorbent bed consisting of sintered pellets.
These pellets consist of mixtures of fly ash
from a coal-burning power plant at Fort
Martin, W. Va., and either technical-grade
ferric oxide or "red mud," a residue of the
aluminum refining process. These sintered
pellets were made by wetting the mixture of
the two materials, forming the mixture into
nominal 1/4-inch spheres (pellets), drying the
pellets, and then firing them in an electric
furnace at 1900 - 2300°F for 10-15 minutes.
Temperatures required to sinter the materials
depended on the composition of the pellets and
were higher for a composition containing red
mud. Sintering was considered satisfactory
when the pellets had been fired enough that
the sharp edge of a knife blade would not cut
them, but not enough to vitrify the surface.
The absorbent pellets used were previously
reported1 as having PbS absorption capacities
of approximately 12 grams of sulfur per ,100
grams of absorbent at 1000°F and 35 weight
percent sulfur at 1500°F, when used at space
velocities of 2000, bed depths of 15 inches,
bed diameters of 1 inch, and absorption pres-
sure near atmospheric. (Space velocity is the
number of volumes of gas at standard condi-
tions passing through a unit volume of absorb-
ent per hour.)
Various other metallic oxides, including
those of zinc, molybdenum, tin, cerium,
lanthanum, and yttrium, were individually
sintered from mixtures of the oxide and fly
ash. None of these oxides proved to be an
effective hydrogen sulfide absorbent when
used in-this form.
Apparatus and Methods
These absorbents were tested by passing a
simulated producer gas through a fixed bed of
the absorbent, contained in a stainless-steel
pipe, in an electrically heated furnace. Figure
1 shows the basic test system. The test appa-
ratus consisted of compressed-gas cylinders
with individual rotameters for blending a gas
similar in composition to producer gas; the
furnace contains the absorbent and plumbing
III-5-1
-------
suitable for passing the gas through the
absorbent. Ports allowed for analytical
determinations in absorber influent and efflu-
ent gas streams. Influent concentrations
normally used were 900 - 1000 grains H2S per
100 cu ft of gas (1.5 percent); effluent con-
centrations gradually increased from zero to
100 grains per 100 cu. ft., at which concen-
tration the tests were terminated.
had condensed and been deposited down-
stream of the absorber to increase both the
flow resistance and the pressure in the
absorber. Tests made at equal pressures, with
and without tar vapor in the gas, indicated
equal absorption capacities (tests 110 and
112). These capacities were all determined at
gas-flow rates equal to a space velocity of
2000.
DISCUSSION
Effect of Pressure and Gas Velocity
One unresolved question about the previ-
ous work was the effect that tar vapors, which
would be present in an actual producer gas,
would have on the ability of the abosrbent to
remove hydrogen sulfide. A tar-vapor gen-
erator was built by partially filling an exter-
nally heated 2-inch-diameter cylindrical vessel
with a bituminous coal tar made from Pitts-
burgh-seam coal in the Morgantown Energy
Research Center's flash carbonizer. The
nitrogen from the flow system was passed
over the hot tar to pick up the vapors, then
transported through a heated line (900°F) to
the bottom of the absorber, where it was
mixed with the other producer-gas constit-
uents and then passed through the absorp-
tion bed. Gas producers yield about 130,000
cu ft of gas per ton of coal,2 and an average
tar production of about 15 gal. per ton.3 This
is approximately 0.4 ml of liquid tar as vapor
per cu ft of gas, or 12,500 ppm by weight.
Because nitrogen only was used to sweep the
tar-vapor generator, the vapor generator was
found to produce 0.8 ml of tar vapor at
600°F. The vapor generator was operated at
600 - 700°F for test work.
Table 1 shows that, at first, the effect of
tar vapor seemed to be to increase the capac-
ity of the absorbent. The reason for this in-
crease was unknown at the time. In subse-
quent tests, Table 2, it was noted that the
presence of tar vapors made little difference
in absorption capacities until sufficient tar
III-5-2
The effect of increased pressure on absorp-
tion was investigated to 10 psig, the limit of
the apparatus: Figure 2 illustrates the effect
of increasing pressure. Within this pressure
range, higher absorption pressures caused an
increase in absorption capacity.
Figure 3, relating the effect of increasing
space velocity to absorbent capacity, shows
that a space velocity of 2000 is near the maxi-
mum that would be permissible at 1000°F,
whereas at 1500°F removal was still effective
at a space velocity of 16,000. These capacities
were all determined at 3-psig pressure in the
absorption bed.
Table 1. EFFECT OF COAL-TAB VAPORS
SULFUR ABSORPTION CAPACITIES OF
SINTERED JAMAICAN RED MUD FROM
REYNOLDS METALS CO.
ON
Test
No.
99
100
101
102a
103a
104a
Tpct
temp..
°F
1000
1250
1500
1500
1250
1000
Capacity of absorbent.
g sulfur/100 g absorbent
Absorbed
from H2S
10.7
27.6
44.1
56.5
37.5
12.0
Regenerated
from S02
12.1
24.6
40.6
52.3
34.0
10.9
sulfur
balance,%
-11.6
10.9
7.9
7.4
9.3
9.2
aTest gas contained tar vapors.
-------
Table 2. RESULTS OF H2S ABSORPTION TESTS USING ALCOA MOBILE RED MUD
Test
No.
105a
106
107a
108
109
110
111
112a
113
114
Absorption
Temp.,
°F
1500
1500
1250
1500
1500
1500
1500
1500
1500
1500
Pressure
psig
0
0
0
0
0
3
0
3
6
10
Weight of
sulfur absorbed
g sulfur/1 OOg absorbent
6.8
7.4
5.1
7.0
13.3
23.1
14.1
23.6
29.1
32.8
Weight of sulfur
regenerated
g sulfur/1 OOg absorbent
4.1
8.2
4.1
5.6
11.9
22.2
14.8
23.8
27.9
28.5
Error in
sulfur
balance
39.6
9.8
19.6
20.0
10.5
3.9
4.7
0.8
4.1
13.1
aTest gas contained tar vapors.
Sintering Conditions
The effect of sintering temperatures on
capacity was thought worthy of investigation.
It was found, however, that these absorbents
have a fairly narrow temperature range in
which sintering will occur within a reasonable
time period. Hurricane Creek (H. C.) Jamai-
can red mud will sinter in 2 hours at 2200°F,
and in 10 minutes at 2300°F, but it quickly
develops a vitreous surface at 2400°F.
Materials incorporated into a fly ash mixture
generally sinter best at 1700 2000°F, but
again a temperature range of only about
200°F can cause the difference between
excessively long sintering times and the rapid
formation of a glassy surface for any specific
composition.
Durability of Absorbents
All previous work had been done using a
1-inch-diameter stainless steel pipe, type 304
or 316, as an absorption tube. This unit was
too small to permit easy unloading of absorb-
ent pellets. The inside walls of the pipe react
with the gases used in this work, causing a
scale formation that gradually (but partially)
envelops the pellets in contact with the pipe
wall. After several tests the pellets could be
removed only by redding out the pipe and
crushing the pellets.
To assess the durability of absorbents
better, a split-shell furnace with a bore suffi-
cient to hold a 2.5-inch-diameter pipe was
substituted for the original furnace. In the
1-inch-diameter unit, some absorbents disin-
tegrated into dust, some fused into a mass,
and others appeared fairly durable. The more
durable absorbents tested in the 1-inch unit
were: (1) sintered Fort Martin fly ash, (2) a
sintered mixture of 25 percent Fe203 and 75
percent Fort Martin fly ash, and (3) a sintered
mixture of 40 percent H.C. Jamaican red mud
and 60 percent Fort Martin fly ash. Although
the fly ash alone demonstrated good dura-
bility in the smaller unit, no additional work
was done with it because it had relatively low
sulfur-absorption capacity. Both the other
absorbents were used in the larger diameter
absorber; however, each was tested through
six H2$ absorption-air regeneration cycles.
Because the larger unit permitted easy
unloading, weights of absorbents before and
after testing could be determined. After test-
ing, the red-mud-containing absorbent had
lost 3 percent of its weight, probably as dust;
10 percent of the pellets had fused into two
separate lumps; and a large amount of pellet
fracture had occurred. The Fe20e-containing
absorbent did not lose weight, no fusion
III-5-3
-------
between pellets had resulted, and only a small
amount of pellet fracture was observed.
Absorption capacity was somewhat lower in
the larger diameter unit: only 16 weight per-
cent at a space velocity of 500, a temperature
of 1500°F, and a LO-inch bed depth. The
reason for this lower absorption capacity is
being investigated.
Absorption at Higher Pressures
The effect of higher pressures (up to 100
psig) was investigated in the 2.5-inch-diameter
unit with a 5-inch bed depth of the sintered
25 percent Fe2C>3-75 percent fly-ash absorb-
ent that had been used in ten previous tests.
(Table 3 shows the results.) Premixed gas in
pressurized cylinders was used for the
100-psig test. Absorption tests were made at
1500°F and pressures of 15, 30, 45, 50, and
100 psig. Space velocities used ranged from
1100 to 4400. Capacities varied from 15.3 to
16.3 weight percent, except for a45-psigtest
in which flow was erratic. This capacity was a
20 percent increase above that obtained at
1500°F and 3 psig with a 5-inch bed depth,
but no trend of capacity change as a function
of pressure was indicated in this region. The
effect of pressure on absorption capacity for
this system becomes asymptotic at 10 to 15
psig, within the region of space velocity
utilized.
The 25-percent Fe2C>3 and fly-ash absorb-
ent has been tested through 15 cycles of
absorption and regeneration (Table 3): the
only evidence of degradation has been some
fracturing of pellets and the fusion of 2 per-
cent of the pellet weight into one lump during
the series of absorptions at elevated pressure.
The maximum temperature at which this
absorbent is durable is 1500°F; at 1600°F,
fusion between pellets becomes severe.
Table 3. H2S ABSORPTION CAPACITIES OBTAINED DURING 15 ABSORPTION-
REGENERATION CYCLES WITH A 25 PERCENT Fe2O3-75 PERCENT
FLY-ASH SINTERED ABSORBENT3
Absorption
temp.,
°F
1000
1500
1250
1500
1250
1000
1000
1250
1500
1000
1500
1500
1500
1500
1500
Gas
flow,
scfh
15
15
15
15
15
15
15
15
15
15
30
45
40
66
16.5
H2S
g/1 00 ft3
1030
890
900
960
910
945
990
1080
1040
1010
1100
1050
1050
1050
393
Absorption
pressure,
psig
3
3
3
3
3
3
3
3
3
3
15
30
45
50
1000
Space
velocity
2000
2000
2000
2000
2000
2000
1000
1000
1000
1000
2000
3000
2670
4400
1100
Duration
of test,
hr
1.5
3.5
2.5
3.5
3.0
2.25
3.25
4.33
6.5
4.75
4.0
2.5
2.25
1.75
18.75
Sulfur
absorbed,
wt %
6.2
12.0
9.0
13.5
11.0
8.4
6.1
8.9
12.8
9.1
16.3
15.3
12.2
15.7
15.8
aFirst 6 tests made with 2.5-inch bed depth; last 9 tests, with 5-inch.
III-5-4
-------
Regeneration of Absorbent
Regeneration of these absorbents with air
has been found to proceed best when initiated
at about 1200°F. Sulfur dioxide is liberated
in concentrations of 5-8 volume percent until
about 60-70 percent of the absorbed sulfur is
burned off at air-input space velocities of
2000. The concentration of sulfur dioxide
then rapidly declines but never becomes zero,
even after a regeneration time of 80 hours. An
SO 2 concentration of about 50 pprn is
reached after 24 hours of regeneration, and
the absorbent is then reused for H2S absorp-
tion. During regeneration a small quantity of
elemental sulfur is also liberated.
Water vapor is formed during the absorp-
tion step and condenses in lines downstream
of the absorber. No doubt some 803 forms
during regeneration and combines with this
condensed water, because sulfuric acid has
been identified in the effluent system.
CONCLUSIONS
Conclusions deduced from work to date
follow:
1. With the type of sintered absorbent de-
scribed here, H^S can be absorbed from
a producer gas at near the exit tempera-
ture of the gas, thus preserving the sensi-
ble heat of the gas for producing work.
2. Absorbents containing more than 37
percent Fe2C>3 degrade more rapidly
than those with lesser amounts of
3. Degradation consists of either pellet
fusion or disintegration into dust.
4. The maximum temperature at which this
absorbent can be used is 1500°F.
5. Tar vapors in the producer gas do not
adversely affect the capacity of the
absorbent for H2S during the absorption
time intervals used in this work.
6. Absorptive capacity increases with pres-
sures up to 15 psig.
BIBLIOGRAPHY
l.Shultz, F. G. and J. S. Berber. JAPCA.
20(2): 93-96, February 1970.
2. Lowry, H. H. Chemistry of Coal Utiliza-
tion, Vol. II. National Research Council
Committee. New York, John Wiley & Sons,
Inc.: 1868 p., 1945.
3. Walters, J. G., C. Ortuglio, and J. Giaenzer.
BuMinesBull. 643: 91 p. 1967.
III-5-5
-------
GAS SAMPLE
H20 OUT
VENT
1
H20 IN
J-.ABSORBER DRAIN
f
TOS02
ABSORPTION TRAIN
MANOMETER
,*,* *
i^V
FURNACE
CONTROL
TUBULAR GLOBAR
3\ TYPE FURNACE
N\BSORBENT
0-30" H20
AIR, 100 psig
'ROT. 10-100 scfh
SAMPLE
sj)nxh
12345
PRESSURE IN ABSORPTION BED, psig
Figure 2. Sulfur absorption capacity
as a function of absorption-bed
pressure for Alcoa mobile red mud.
401 —
To
DRIP LEG
ROT.
,0.3-3.0
FM scfh
ROT.
1 -16
kscfh
ROT.
3.06 - 0.6
Fl } scfh f p|
I FILTER
)CO^
(CO
|HoS
26%
1.5%
17%
2000
Figure 1. Flowsheet for removal of sulfur
from hot producer gas.
epop_ 10000
SPACE VELOCITY
14000
Figure 3- Sulfur absorption capacity
as a function of space velocity for
Reynolds H. C. Jamaican red mud.
III-5-6
-------
6. GASIFICATION OF SOLID PARTICLES
CONTAINING CARBON
C. Y. WEN AND S. C. WANG
West Virginia University
INTRODUCTION
Fluidized-bed gasifiers have been shown to
possess technical and economical advantages
over conventional coal-gasifying devices in
solving air pollution problems resulting from
the use of high-sulfur coal. However, the com-
plex phenomena of chemical reaction, heat
and mass transfer on solid particles, and the
interaction of gas and solid flow patterns
within the fluidized bed have hampered
design development. Researchers in recent
years have, however, provided insight into the
rate of processes occurring on a single particle
and with the phenomena associating bubbles
in fluidized beds. It is the purpose of this
paper to present the method by which realis-
tic mathematical model describing a single-
particle phenomenon can be formulated and
how such information can be incorporated
into a fluidized-bed model that accounts for
the bubble growth and coalescence in order to
arrive at a more reliable design for fluidized-
bed gasifiers.
SINGLE-PARTICLE SOLID-GAS REAC-
TION MODELS
The reaction models for single-particle
solid-gas reaction systems can be generally
divided into three categories: (1) unreacted-
core-shrinking model, (2) homogeneous
model, and (3) zone-reaction model.
If the porosity of the unreacted solid is
very small and the solid is practically impervi-
ous to gaseous reactants, the reaction will
occur at the interface between the unreacted
solid and the porous-product layer. Under
such conditions the unreacted-core-shrinking
model is applicable. Figure 1 shows the solid
and gaseous reactant concentration profiles
for this model. This model is also applicable
when the chemical reaction rate is very rapid
in comparison with the diffusion rate of the
gaseous reactant. The zone of reaction in such
a case is narrowly confined to the interface
between the unreacted solid and the product
solid.
On the other hand, if the solid is porous
enough for the gaseous reactant to diffuse
freely into the interior of the solid, the
unreacted-core-shrinking model is no longer
applicable. In such a case the reaction may be
said to occur, in a macroscopic sense, homo-
geneously throughout the solid to produce a
gradual and uniform variation in solid-react-
ant concentration in the particle. The
homogeneous model describes this situation
closely.
In most actual cases, however, the solid-gas
reactions cannot be completely represented
by either of the first two models mentioned
above, which represent two extreme cases in
the solid-gas reaction systems. The majority
of the solid-gas reactions probably fall
between these two extremes.
The zone-reaction model1-2 is a more
general model for solid-gas reactions because
it displays an intermediate between the
unreacted-core-shrinking and homogeneous
models, as shown schematically in Figure 2.
In fact, the unreacted-core-shrinking and
homogeneous models can be derived as special
cases from the zone-reaction model.3
III-6-1
-------
In the zone-reaction model, the kinetic
behaviors depend on the solid structures.
Figure 2 also shows three types of solid struc-
tures:
1. Volumetric reaction model: the solid
particle is composed of small grains of
solid reactaiit embedded in inert ma-
terial.
2. Grain model: the solid particle is com-
posed of small grains, each of which
reacts according to the unreacted-core-
shrinking model.
3. Pore model: the reaction occurs at the
surface of the pores.
The effects of temperature on the reaction
rate and on the concentration profiles of
gaseous reactant (A) and solid reactant (S) are
summarized schematically in Figure 3. In the
lowest temperature region, V, of the zone-
reaction model, the concentration distribu-
tion of both gaseous and solid reactants are
uniform—a limiting case in which the homo-
geneous model is applicable. The reaction rate
in this temperature region is controlled by
that of individual grains or pores. In the
highest temperature region, I, on the other
hand, the gaseous reactant concentration at
the boundary between the reaction zone and
the product layer is practically zero, so that
the reaction rate is controlled by the diffusion
of gas through the product layer.
Note that in temperature regions I and III
of the zone-reaction model, the gaseous
reactant concentration profiles in the product
layer are similar to those for the unreacted-
core-shrinking model as shown in Figure 3(b).
Because the unreacted-core-shrinking
model has wide application for many investi-
gators, and because the mathematical treat-
ment is relatively simple in comparison with
the zone-reaction model, a brief discussion of
the unreacted-core-shrinking model is given
below, with emphasis on the effects of heat
and mass transfer on reaction rate in terms of
the effectiveness factor. The detailed mathe-
matical treatments can be found else-
where.3 >4'5
III-6-2
The effectiveness factor is defined as fol-
lows:
actual (overall) reaction rate
s reaction rate obtainable when the
reaction site is exposed to the gas
concentration and temperature of
the bulk-gas phase
Because the denominator is a constant, the
effectiveness factor is, in effect, a dimension-
less surface-reaction rate.
Figures 4 and 5 show plots of effectiveness
factor versus fractional solid reactant conver-
sion, and compare unsteady-state heat trans-
fer solution to that of pseudo-steady-state.
Figure 4 indicates a case in which the pseudo-
steady-state analysis could lead to an erro-
neous conclusion. The chemical-reaction-
controlling region would never have been
realized if the pseudo-steady-state analysis
had been used. The unsteady-state analysis,
on the other hand, shows that chemical reac-
tion could be rate-controlling when the initial
temperature of the particle is sufficiently low.
Figure 5 depicts the effects of the heat capac-
ity of the unreacted core and the heat of reac-
tion on thermal instability. The occurrence of
thermal instability is less likely when the heat
capacity is high.
Figure 6 illustrates experimental results
that were computed on the basis of the
unreacted-core-shrinking model. The corre-
sponding numerical solutions (solid, dotted,
and dashed lines) are also shown in the figure
for comparison. The experiment was per-
formed in a thermabalance (shown in Figure
7) by burning a single solid sphere in a stream
of heated air. The solid pellet was prepared by
mixing the desired proportion of activated
charcoal with aluminum oxide, which serves
as an inert porous medium. Sodium silicate
(waterglass), diluted with the proper amount
of water, was used as a binder in forming the
pellets. The particle was heated at 700°C in
an inert atmosphere prior to the combustion
test to remove moisture and to ensure no
weight loss from inert solid during the test.
-------
"Extinction" occurred at about 85 percent of
solid conversion. Note that the unsteady-state
heat transfer analysis (solid line) describes
more closely the experimental result than the
pseudo-steady-state analysis (dashed line) at
the initial and transition stages. The values of
the parameters used in the correlation are
independently estimated from empirical
correlations.
Figure 8 is an example of a triangular dia-
gram that shows the reaction path of an
endothermic reaction. The corresponding
effectiveness factor versus conversion plot is
shown in the Figure 8 inset. The main advan-
tage of a triangular diagram is that it shows
the magnitude of each resistance during the
history of a reaction.
In the following discussion we consider
only the case in which solid particles react
with fluidizing gas while maintaining their
original size because of the formation of inert
solid product. The gasification of coal, the
roasting-of sulphide ores, and the reduction of
iron ores are examples of this situation. The
following stoichiometric equation can be used
to represent these reactions.
aA(gas) + S(solid) ->• gaseous
and/or solid products (1)
The proposed calculation method6 assumes
that solids follow the shrinking-core model
and that the overall conversion rate is con-
trolled by a chemical reaction step; in the
unreacted-core-shrinking model, the reaction
is confined at the surface of the core, which
recedes from the outer surface toward the
interior of the particle. When diffusion
through the product layer becomes rate-
controlling, or when other single-particle reac-
tion models are used, the conversion-versus-
time expressions must be changed accordingly
for our calculations.
The reaction of a gaseous component by a
first-order irreversible reaction can be given
as:6
1
47rrc2a
dNA
dt
1
47rrc2
dNc
dt
(2)
where rc is the radius of unreacted core and
kc is the rate constant for the reaction.
When the reaction is carried out in the
bulk-phase reactant-gas concentration, CA,
the extent of conversion, Xg, of a particle
having radius R is given by:
l=lJ£=l-(l xs)i/3
where time for complete conversion, T, is:
(3)
(4)
When the resistances of the chemical reaction
step and of diffusion through the ash layer are
comparable, the rate constant, kc, is replaced
by k, defined by:
_L=1 '
k kc
eA
(5)
Next, let us consider a reactor with a con-
stant feed rate of both solids and gas, the
solids being of uniform size and complete
mixing. Because the conversion of an individ-
ual particle of solid depends on its length_pf
stay in the reactor, the mean conversion, Xs,
of the exit stream of solids is given by:
1-XS=\ (l-Xs)E(t)dt (6)
A=0
where the exit age distribution function for a
reactor of complete mixing is:
1 -tyf
E(t)=r-e (7)
t
When chemical reaction is the rate-controlling
step in a shrinking-core particle, substitution
of equations (3) and (7) into equation (6) and
subsequent integration yield:
- T
6(i)3 [l-exp(-T/t)l
(8)
III-6-3
-------
Gas and Solid Flows Based on the Bubble As-
semblage Model
Let us summarize the essentials of the
bubble assemblage model7 as follows:
A fluidized bed may be approximately
represented by "N" numbers of compart-
ments in series. The height of each compart-
ment is equal to the size of each bubble at the
corresponding bed height. Each compartment
is considered to consist of the bubble phase
and the emulsion phase.
The bubble phase is assumed to consist of
spherical clouds; the cloud diameter.can be
derived from:
2umf/emf
- umf/emf
(9)
The size of the bubble diameter along the bed
height is approximated by:
where do is the bubble diameter just above
the distributor. The rising velocity of a bubble
is given by:
% = Uo-Umf+O.TlKgdb)1/2 (11)
From an arithmetic average of bubble sizes,
the height of the i-th compartment can be
expressed as:
(2+m) J'1
(12)
where m = 1.4 ppdp (uo/umf).
The voidage distribution assumes that up to
Ljnf, the voidage, e, is uniform, while above
Lmf, e increases linearly along the bed height.
Then the number of bubbles, n, in the i-th
compartment is given by:
_ 6St
n ~
e — emf
1-emf
Therefore, the volume of bubbles, clouds, and
emulsion in the i-th compartment can be cal-
culated, respectively:
Vbi = n-(Ahj)
(14)
Vei = StAhi- Vbi- Vci
(16)
The overall interchange coefficient of the gas
between the bubble and the emulsion phase
based on a unit volume of gas bubbles may be
given by the following experimental relation:
(Kbe)b = 11/db
(17)
The solid interchange coefficient between
both phases is assumed to be given by:
iv \ - 7 (l-emfumf ub
(Kbe)bs -
In order to describe the distribution of solids,
let us define, for convenience, the following
quantities:
7C = volume of solids dispersed, in clouds
_ and wake _
volume of bubbles
7e = volume of solids in emulsion _
volume of bubbles
The upward motion of the solids, as a part of
the wake of the bubbles rising from the i-th
compartment to the (i+l)-th compartment,
sets up a circulation in the bed with down-
ward movement of solids in the emulsion
phase from the (i+l)-th compartment to the
i-th compartment. When the solids are fed to
the bottom of the bed at a constant volu-
metric flow rate, W, the total upward flow
rate, Wj,, from the i-th compartment is given
<,„
The total downward flow rate, We, from the
(i+l)-th compartment to the i-th compart-
ment is given by:
III-6-4
-------
CAS
Cgbi
> C- •- I I l "'I V—W
>t - sbi
If the solids are fed at the top of the bed and
withdrawn from the bottom, W must be re-
placed by -W in the above equations. These
relations mentioned above are shown sche-
matically in Figures 9 and 10.
NOMENCLATURE
The following list of terms defines expres-
sions used throughout this paper, in both the
text and the figures.
Term Definition
a Stoichiometric coefficient
•^ CAo^eA(To)Cpe/aCSoke> ra^° °^
mass to thermal diffusivities in ash
layer
CA Concentration of gaseous reactant A
in ash layer, C^c
^Ac Concentration of gaseous reactant A
to unreacted-core surface
C?Am Concentration of gaseous reactant A
at boundary between reaction and
diffusion zones
Concentration of gaseous reactant A
in bulk gas phase
Concentration of gaseous reactant A
at outer surface of particle,
mole/L3
Gas-phase concentration in the bubble
in the i-th compartment
Gas-phase concentration in the emul-
sion in the i-th compartment
Heat capacity of unreacted core,
H/MT
Volumetric heat capacity of ash layer,
H/L3T
Concentration of solid reactant S
Initial concentration of solid reactant
S
Cgs Concentration of solid reactant S at
outer surface of particle, mole/L3
db Bubble diameter, cm
dy db in the i-th compartment, cm
ED
Eks
Ekv
g
G
Ah
H
AH
Diameter of cloud, cm
Bubble diameter just above the dis-
tributor, cm
Particle diameter, cm
Effective diffusivity of component A
in ash layer, L2 /0
Activation energy of reaction rate
constant
Apparent activation energy, H/mole
Activation energy in temperature
range where diffusion is controlling
1/2 Ekv
Activation energy in temperature
range where reaction rate is control-
ling
Acceleration of gravity, cm/s2
pcCpcT0/aCs0(-AH), ratio of enthal-
py of unreacted core to heat of
reaction
Distance from the distributor, cm
Convective heat transfer coefficient,
H/L20T
Height of the i-th compartment, cm
Radiational heat transfer coefficient,
H/L20T4
-1
-------
M
W
NRe
'm
R
t
T
ub
ubi
ue
umf
uo
Vbi
Vci
Vei
Number of bubbles in i-th compart-
ment; also order of reaction for
gaseous reactant
N3Gramme moles of A and B
modified Nusselt number of
convective heat transfer
/^e> modified Nusselt number
for radiational heat transfer
Reynolds number
RkmA(T0)/DeA(T0)> modified Sher-
wood number
Distance from the center of sphere (or
from solid surface) to reaction
surface
Radius of unreacted core, L
r at boundary between reaction zone
and diffusion zone, L
Radius of a particle, cm
Gas constant, H/mole T
Cross sectional area of bubble phase in
the i-th compartment, cm2
Cross sectional area of the bed, cm2
Mean residence time of particles, s
Temperature
Temperature at unreacted-core surface
Initial particle temperature
Temperature at bulk gas phase
Temperature at outer surface of parti-
cle
Velocity of rising bubble, cm/s
Velocity of rising bubble in the i-th
compartment, cm/s
Velocity of bubble with respect to
emulsion ahead of it, cm/s
Velocity of emulsion phase, cm/s
Minimum fluidizing velocity, cm/s
Superficial gas velocity, cm/s
TC/TO
Volume of bubble phase in the i-th
compartment, cm3
Volume of cloud region in the i-th
compartment, cm3
Volume of emulsion phase in the i-th
compartment, cm3
W
ei
X
13
€
emf
Volumetric feed and outflow rate of
solids, cm3
Volumetric upward flow rate from the
i-th compartment, cm3/s
Volumetric downward flow rate from
the i-th compartment, cm3/s
Fractional conversion of solid reactant
S
Mean fractional conversion of solid
PC
Pp
Void fraction in a bed as a whole
Void fraction in a bed at minimum
fluidization
Effectiveness factor based on surface
reaction
Effectiveness factor based on volume
reaction
Density of unreacted core, M/L3
Density of solid, g/cm3
a^g0)CA/DeA(T0), modified
Thiele modulus
Time for complete conversion of a
single particle, s
BIBLIOGRAPHY
1. Ishida, M. and C. Y. Wen. Chem. Eng. Sci.
26:1031,1971.
2. Ishida, M., C. Y. Wen and T. Shirai. Chem.
Eng. Sci. 26:1043, 1971.
3. Wen, C. Y. Ind. Eng. Chem. 60(9): 34,
1968.
4. Wen, C. Y. and S. C. Wang. Ind. Eng.
Chem. 62(8):30, 1970.
5. Yagi, S. and D. Kunii. Fifth International
Symposium on Combustion, p. 231. Rein-
hold, New York, 1955.
6. Yoshida, K. and C. Y. Wen. Chem. Eng.
Sci. 25:1395, 1970.
7. Kato, K. and C. Y. Wen. Chem. Eng. Sci.
24:1351, 1969.
III-6-6
-------
SOLID
REACTANT
VOLUMETRIC
REACTION MODEL
Figure 1. Concentration profile for the
unreacted-eore-shrinking model.3
Figure 2. Concentration profile and
solid structure in the particle.1
1/T
- 1
•'AO
\
£c,
^
Ek -.Efc/2
s v
A°
^c
y.
AO
AO
-
-r=5
AO
.(a) ZONE-REACTION MODH.
|b) UNREACTED-CORE MODEL
Figure 3. Schematic diagram representing characteristic behavior of solid-gas reaction
systems under various temperature regions.1
IH-6-7
-------
j3= 0.005
NSh-lOO
RTn
25
s ; 0.139
(NNu'c - 10
n : 1
A=0
-EL = 0.95
To
100
60
40
20
I I I I I
HEAT TRANSFER*:^
"UNSTEADY STATE ** - .,,
PSEUDO-STEADY-STATE ""~*
0.2
0 0.2 0.4 0.6 0.8 1.0
SOLID REACTANT CONVERSION, percent
Figure 5. Effect of heat of reaction and heat capacity of unreacted core on thermal instability
in terms of G.4
III-6-8
-------
TO TRANSDUCER
ID'3
c
'i
1
O)
„
&
DC
K
2 10
^3
OC
LU
O.
I—
DC
I I I I I I I
HEAT TRANSFER:
i im^-ri- A CN\J r*-r*-Y-r-
l PSEUDO-STEADY-STATE
»•» —
•
X
) 0.2 0.4 0.6 0.8 1
SOLID CONVERSION, X
CONVERTER AND
WEIGHT RECORDER
/
/
A
bta_^
FORCE TRANSDUCER^
GOLD CHAIN
PURGING
ct
j
"('c._
Sir-
.^
j
-n
n
GAS-*-«^
CIRCULAR SCREEN
\
GAS •+
NICKEL-CHRO
WIRE
/
n ELECTRIC,
0 HEATERS'
T0 =442' C * £ =0.556
TV, = 573°C DeA =0.321 cm2/sec
Tj = 27 ° C Ke = 0.0032 cal/sec-cm- » c O
NRe=393 E = 35,700 cal/gmol THERMOCOU
^-
»
'
^^
^"fe
ME
*
<
—
\
»LE
R = 1.244 cm Ks (To) = 5.16 x 10"4 cm/sec |^
3
h
V
/-\ft<
PQ»
^ If
jfa.
?ft>
f%
r^
4 —
3* — PURGING A
GAS
n
%J WINDLASS
*1P
11i" SCH. 40
— CARBON
STEEL PIPE
53'/4
SIGHT GLASS
w^
HMi
--^-"
q
7
^- DAI 1
^^i^^*^ Or\UL-
— -/ VALVE
CHAIN CATCH
IMCI II ATION ~
1 IN OVJL^ 1 I Wl1!
~"l/ j
_ 1
7
^
i
I
r
/
7
y,
^
/
^
d
2
i
i
REACTOR
d'/i" SCH. 40
310 STAINLESS
STEEL PIPE)
1
=^*O-*-
HOT GAS
BLEED LINE ^
• SOLID SAMPLE
1
i
1
-GAS PREHEAT
ZONE
(Vi" INTALOX
SADDLES)
1
IJ 1
Figure 6. Comparison of experjmental result
and computer simulation for oxidation of
carbon in porous solid, using unreacted-
core-shrinking model.4
Q
?**""»< GAS
Figure 7. Thermobalance.
IH-6-9
-------
0.2 0.4 0.6 0.8
SOLID CONVERSION, X
db- - m+ do - 1-4Pp dp —
0 0.2
HEAT TRANSFER
CONTROL
0.4
Uo
\ SM MASS TRANSFER
CONTROL
Figure 8. Triangular diagram showing the Figure 9. Schematic diagram of i-th compart-
reaction path and the magnitude of each ment in the bubble-assemblage model for
resistance in an endothermic reaction.4 fluidized bed.7
III-6-10
-------
. . /?"7\ • • /i*"**. • /r" i^ •• /?"*A»* /T"*A* i*/5"VV * " ' '
•••.- Q - Q;: Q/.'SQ v.Q:::©.% :.-
iononnonnnnoonn
Figure 10. Main features of solid movement and gas flow as visualized.6
II1-6-11
-------
7. MANUFACTURE OF ACTIVATED CARBON
BY GASIFICATION
IN A FLUIDIZED BED
A. A. GODEL
Societe Anonyme Activit
INTRODUCTION
Today, I propose abandoning momentarily
the study of combustion problems in fluid-
ized beds in order to discuss chemically the
manufacture of activated carbon.
Activated carbon is a most powerful
absorbent. It is used to purify both air
(removing malodourous gases) and liquids
(purifying water and blanching sugar.)
The principle of the manufacture of acti-
vated carbon which concerns us here consists
of subjecting carbonaceous matter (e.g.,
charcoal, peat lignite, coal, and briquettes) to
the action of steam at about 800 - 1000°C.
In fact, this treatment represents a true
gasification process involving selective oxida-
tion of tarry matter contained in the carbo-
naceous matter. It is sometimes considered-
in a rather simplified manner—as a cleansing
operation for existing pores or, for pores
created by the operation within the internal
structure of the carbon, pores and capillaries
which thus become "absorbent."
In 1948, our factory at Vernon was the
first in France to manufacture activated
carbon by fluidization, in accordance with a
continuous-flowing thin layer.1 This process
was later licensed to the French Societe
"Carbonisation & Charbons Actifs" which
applied it in two industrial furnaces in its
factory at Parentis (Landes).
NEW TECHNIQUE
However, it's not about this continuous
activation technique which I should like to
speak today, but rather about an entirely new
technique, still in the experimental design
stage, that is not yet developed industrially.
Contrary to the preceding process, the new
process is characterized by treating carbo-
naceous matter in deep fluidized beds under
the action of pure stream at high temperature,
by discontinuous "batch process" operation.
As indicated previously, this operation
results in the gasification of a more or less
important fraction of the raw material. To
obtain certain grades of activated carbon, it is
sufficient to gasify a quarter or a third of this
raw material; for others, the process must be
carried further—to two-thirds or even three-
quarters. The manufacture of activated
carbon is thus linked with gasification. Yield,
that is the ratio between the weight of the
activated carbon obtained and the weight of
raw material charged, varies not only accord-
ing to the nature of the carbon obtained, but
also according to the processing methods
used: the goal of the new process is to con-
siderably improve technical efficiency.
III-7-1
-------
In general, the production of activated
carbon and gasification result from reactions
of steam at red heat on carbon; these are
pressure- and temperature-dependent balan-
cing reactions situated between two equa-
tions:
C + H2O = CO + H2 (1)
(28.8 K calories per mole)
C+2 H2O = CO2 + 2 H2 (2)
(14.8 K calories per mole)
Equation (1) is dominant above 1000°C
(1830°F), requiring practically no excess
steam; equation (2) is dominant below
800°C (1470°F), requiring, by comparison,
steam consumption significantly greater than
the stoichiometric ratio. It is for this reason
that I have linked here the manufacture of
activated carbon with the gasification process.
I shall limit my discussion, however, to the
manufacture of activated carbon and shall
simply mention,, in passing, possibilities for
the application of this same process to
integral fuel gasification.
In this new technique, we operate by batch
process and not by thin-layer fluidized-bed
continuous flow: we do this to avoid both
activated carbon degradation during process-
ing by bituminous gases which are given off,
and the inconvenience which occurs inevita-
bly during continuous fluidized flow due to
the mixture of crude and activated fractions.
Having adopted the high-temperature batch
process, in order to obtain the best weighted
yield and best gas production, it is obvious
that" we must operate in cyclic phases. To
obtain, during one reaction phase, the impor-
tant calorie transfer necessary to maintain the
temperature at 1000°C (1830°F), in spite of
reaction endothermicity, it was particularly
essential to operate by fluidization in order to
make maximum use of the excellent heat
transfer ratio existing by contact with the
internal heating surfaces. In the new process,
these heating surfaces will be constituted by
III-7-2
fixed heat-accumulating masses preheated to
high temperature in a preceding phase.
As you may note, this technique resembles
somewhat that of the production of water gas
in fixed layer; however, instead of using
treated carbon for the accumulation of calo-
ries, which would only lead to insignificant
phase durations (a few minutes only), we use
important fixed heat-accumulating masses
which permit extending phase durations to
more than half an hour or an hour. Under
these circumstances, fluidized-bed tempera-
ture may be made to oscillate, at will, around
the average reaction design temperature,
whether for equation (1) or (2); although
each absorbs an unequal, but considerable,
quantity of heat, the oscillations are of the
order of 1000°C maximum.
It has always been difficult to achieve such
heat transfers by traditional processes; e.g., an
admixture of flue gas with steam at high
temperature, or the use of retorts heated from
the outside by flames, thus resulting, in
practice, in too low temperatures giving low
yield.
The following figures will clarify the ques-
tion of the required heat supply—the gasifica-
tion of 12 (the atomic weight) grams of
carbon, according to reaction 1 already men-
tioned as applicable at 1000°C (1830°F),
requires the following heat supply:
1. For endothermicity, 28.8 calories.
2. For additional superheating of the
steam, between 450 and 1000°C (840
and 1830°F), 6.3 calories.
Although theoretical, these figures show
clearly that endothermicity dominates and
that it is easy to complete superheating of
steam from 450 to 1000°C (from 480 to
1830°F) by direct introduction into the
fluidized bed. The steam is produced by the
recuperating boiler and introduced into the
base of the reactor at precisely 450°C
(840°F). To work in cyclic phases, a produc-
tional unit will preferablly be provided with)
-------
two reactors, either coupled or separate, but
which operate in alternative phases: that is,
one gasification phase and one reheating
phase of heat-accumulating refractory masses.
It is, of course, possible to have production
units consisting of three or more reactors.
Each reactor is equipped with piles of
refractory material forming heat-accumulating
masses which will be placed in good contact
with the fluidized bed. These masses will be
chosen on the basis of their refractoriness,
their non-adherence of clinkers, and their
good thermal conductivity.
With regard to thermal conductivity,
certain carborundum agglomerates which are
highly refractory (1650 1700°C-3000-
3100°F) have thermal conductivity 5 or 10
times greater than that of current refractory
materials, or almost equal to that of stainless
steel.
TWO EXAMPLES
To illustrate the method of construction
and operation of the new process, I shall
describe two examples, although other combi-
nations may be envisaged to correspond to
various uses.
Twin-Reactor Furnace
In the first example, represented by Figures
1 and 2, the two reactors in the activated
carbon and gasification production unit are
juxtaposed in a single rectangular furnace
divided into two identical, but distinct, parts
(A and B); each reactor (A and B) contains a
nest of parallel, vertical, heat-accumulating
refractory slabs (1). These slabs swell out at
the base (2), leaving only a narrow slot (3)
used for introducing reaction steam; the
fluidized bed is supported by this steam injec-
tion during the entire blowing period which
begins when carbonaceous material is charged
into reactor A at the beginning of the reheat-
ing phase. The material is charged pneuma-
tically at (4). reaching a level slightly higher
than that of the slabs.
At the end of the reheating phase, the steam
is stopped, thus provoking instantly the flow
of carbon by (3) in the hopper. The activated-
carbon product is evacuated pneumatically by
(5) toward storage (not shown).
During the reaction (or gasification) phase
in reactor A, where gas is produced, air is
introduced at (6') in reactor B; this gas thus
burns in the upper chamber of reactor B
which is then in reheating phase. The combus-
tion gases reheat the slabs (!') and escape at
the base at (3'). They are finally evacuated
from the reactor by (7') to a recuperating
waste heat boiler (not shown) which produces
superheated steam at 450°C (840°F) for
activation; this superheated steam is intro-
duced in reactor A at (8). Slab thickness,
usually greater than their spacing, is calcu-
lated so as to obtain required calorific
capacity corresponding to phase and reacting
duration.
The set of reactors A and B are put into
operation quite simply: when empty, they are
reheated simultaneously to service tempera-
ture by gas or fuel oil burners (9) and (9').
When a temperature of about 1200°C
(2190°F) is reached in the reactors and when
the boiler is under pressure, the burners are
stopped and the set of reactors is placed in
cyclic reversal operation by injecting, at the
beginning of the phase, alternately in one
reactor then in the other, the carbonaceous
material to be activated. The steam must be
injected from the beginning, in the lower part,
and during the active charging and gasification
operations.
Cylindrical Double-Reactor System
In the second example, illustrated by
Figures 3 and 4, the two reactors A and B are
separate and cylindrical, permitting the use of
III-7-3
-------
cyclone dust collectors for treating gases given
off during reaction, and also operation under
a pressure of several bars.
The heat-accumulating mass placed in each
reactor (A and B) is composed of a refractory
block (1), run through by a nest of 20 cm
diameter tubes (2). These tubes narrow out at
the base so that the orifice (3) at their ends
may let pass the reaction steam injected
through (8); the steam backs up the fluidized
bed which has been previously charged into
reactor A to a level slightly higher than that
of the heat-accumulating block.
Steam blow-in through (8) must begin in
reactor A at the very beginning of the phase,
just before introducing the carbonaceous
material pneumatically through (4); it is
stopped at the end of the activation-reaction
phase when the carbon flows down immedi-
ately, to be evacuated pneumatically through
(10) toward storage (not shown).
Reaction-gas dust collecting takes place in a
cyclone (12) in which the lightest particles are
deposited; these particles, consisting mainly
of activated carbon, may be either re injected
into reactor A or collected directly (as
shown).
The gas thus treated is fed through (11)
into a combustion chamber (15) where air is
introduced through (14); the emitted flue gas
is introduced through (16') to the bottom of
reactor B and, therefore, during the next
phase through (16) to reactor A; flue gas thus
passes from the bottom to the top of reactor
B and through a bed of finely granulated
refractpry material—on top of the nest of
tubes in the accumulating block—introduced
pneumatically through (6') at the beginning
of the reheating phase. Obviously, this granu-
lated material contributes no calories of itself,
but constitutes an auxiliary fluidized bed that
increases the transmission of calories to the
walls of the nest tubes during reheating. The
granular refractory material may be a sulfur
acceptor.
III-7-4
The flue gas evacuated from reactor B passes
through cyclone (13) which retains the flue
dust eventually emitted from the refractory
material. Flue gas is finally evacuated to a
waste heat boiler, not shown.
At the end of the reactor B reheating
period, which coincides with the end of the
reactor A activation period, gasification is
stopped and so is the circulation of flue gas.
The refractory material then flows freely
down through the lower orifices of the tubes,
from where it is evacuated at the base through
(10') and conveyed pneumatically into a silo
(not shown) where it awaits transfer during
the next reactor A heating phase. When
reactor B is reheated, it is ready to receive
first the steam, and then the load of carbo-
naceous material: the cycle thus continues.
System Operation
To put reactors A and B into operation,
procedure is as indicated in the first example
(Figures 1 and 2): they are heated simulta-
neously by gas or fuel oil burners at (9) and
(9').
In both described installations, air and
steam valves are ordinary, but the valves for
evacuating activated carbon, gas, and flue gas
are water cooled.
Valves operate automatically in accordance
with programming which takes into account
heating requirements. These requirements
may vary considerably according to the
nature of the carbons treated, or rather,
according to the nature of the ash of these
carbons: if the ash is extremely fusible, it may
be evacuated in melted form, in which c'ase, it
would be advisable to adopt a high reaction-
temperature solution. If the ash is not ex-
tremely fusible, temperature levels lower than
ash melting point will be used and the ash will
be evacuated in powder form together with
the activated carbon: ash will then be
separated, using commonly used industrial
processes: e.g., dust screening and washing.
-------
Coming back to the second type of installa-
tion which I have just described (Figures 3
and 4), the following figures are provisional
estimates.
The inside dimensions of the unit, illu-
strated by Figure 4 and which corresponds to
a production capacity of 6 to 12 tons per day
(even 18 tons per day, for production varies
considerably with the quality of the activated
carbon needed) are:
1. Section of each cylindrical reactor, 6 sq
m.
2. Inside diameter, 2.80 m.
3. Height of the cylinder, 3.50 m.
4. Total height of heat-accumulating refrac-
tory material in each reactor, 1 m.
5. Total volume of heat-accumulating re fac-
tory material in each reactor, 4 cu m.
Production of superheated steam by the
waste heat boiler is superabundant; for this
reason, excess gas can be tapped through (17).
Operation under a pressure of several bars
may be desirable, either in increase the pro-
duction of activated carbon or to improve the
purification of the gas in view of its use in gas
turbines.
Another way of increasing the production
of activated carbon and gas is to operate with
a highly expanded fluidized bed.
In fact, it is conceivable that if the installa-
tion of the reactors must be adapted to
integral fuel gasification, the compressed gas
produced, perfectly desulphurized, might be
used in turbines for the production of energy
in a mixed- gas-steam cycle, without atmos-
pheric pollution.
CONCLUSION
I must state, however, that the process
which I have just described to you was specifi-
cally designed for the production of activated
carbon, with the simple aim of being able to
adapt it eventually to integral fuel gasification
for the production of synthesis gas, rich in
hydrogen and carbon monoxide, without
either nitrogen or carbon dioxide. The gas
produced, being of a highly reducing nature,
should be easily desulphurized by traditional
scrubbing methods or by absorption in dry
conditions by sulfur acceptor, according to
new processes such as Professeur Squire's,
thus permitting effective recuperation of pure
sulfur.
In conclusion, I should like to state that
activated carbon, the subject of the manu-
facturing process which I have just described,
is one of the most precious auxiliaries for the
protection against both atmospheric and
water pollution agents.
BIBLIOGRAPHY
1. Chemical Engineering, July 1948.
2. Swiss patent No. 250, 891.
3. French patents No. 942, 699 and 951, 153.
III-7-5
-------
SECTION BB
LEGEND
1. REFRACTORY SLABS
2. SLAB BASES
3- INTER-SLAB SLOTS
4. CARBONACEOUS
MATERIAL INLET
5. CARBON OUTLET
(TO STORAGE)
6- AIR INLET
7. GAS OUTLET
(TO BOILER)
8. SUPERHEATED
STEAM INLET
(FROM BOILER)
9. BURNER (GAS
OR FUEL OIL)
Figure 1. Twin-reactor furnace.
III-7-6
-------
KEY
STEAM •
GAS -
BOILER
1
REACTOR A
8]
t ^L
i &
1..- "
REACTOR B
a';; i /
Figure 2. Twin-reactor furnace, flow sheet.
Figure 3. Cylindrical double-reactor system, flow sheet.
111-7-7
-------
-ij
00
REFRACTORY BLOCK
TUBE NEST
ORIFICES
CARBONACEOUS
MATERIAL INLET
(NOT USED)
GRANULATED REFRACTORY
MATERIAL INLET
(NOT USED)
STEAM INLET
BURNER (GAS OR FUEL OIL)
CARBON OUTLET (TO STORAGE)
TREATED GAS FROM CYCLONE
CYCLONE (REACTOR A)
CYCLONE (REACTOR B)
AIR INLET
COMBUSTION CHAMBER
GAS TO REACTOR
EXCESS GAS TAP
Figure 4. Cylindrical double-reactor system.
-------
SESSION IV:
Conceptual Design and Economic Feasibility
SESSION CHAIRMAN:
Dr. D.H. Archer, Westinghouse
-------
1. CLEAN POWER SYSTEMS
USING
FLUIDIZED-BED COMBUSTION
A. M. SQUIRES
The City College of The City University of New York
INTRODUCTION
A major opportunity for clean power from
coal involves the combustion of coal at high
pressure in presence of a desulfurizing agent,
and the generation of power by a combina-
tion of gas- and steam-turbine cycles.
This paper reviews briefly three candidate
combustion technologies for use in the fore-
going combination, and considers power cycle
arrangements suited for each of the three
approaches to combustion.
HIGH-PRESSURE SULFUR CAPTURE
In a system for combustion at high pres-
sure, an agent derived from dolomite may
advantageously capture sulfur during a first
coal-processing step. For this first step there
are three cases to consider.
1. Complete combustion, using air in excess
of stoichiometric.
2. Partial combustion, using between about
a third and half of the stoichiometric air
and yielding a fuel gas containing CO
andH2.
3. A carbonization yielding low-sulfur coke
as well as fuel gas, with heat for the
carbonization supplied by a partial
combustion which consumes on the
order of 10 percent of the stoichiometric
air.
gaseous sulfur species may be absorbed by
agents derived from dolomite:
Work is underway at The City College on
the kinetics of three reactions whereby
[CaCO3+MgO] + SO2 + 0.5 O2 =
[CaSO4+MgO] + CO2
[CaCO3+MgO] + H2S = [CaS+MgO]
H2O + CO2
[CaO+MgO] +H2S= [CaS+MgO] +
H2O
(1)
(2)
(3)
Our primary interest is in the application of
the reactions in systems where the solid is
used cyclically, a step for desorption of H^S
from the solid being present in the cycle,
thus:
[CaS+MgO] + H2O + CO2 =
[CaC03+MgO] + H2S
(4)
The equilibrium for reaction (4), the reverse
of reaction (2), is a strong function of temper-
ature; formation of H2S is favored by a lower
temperature. To obtain a gas containing H2S
at a concentration suitable for economic
conversion to sulfur in a Claus system, reac-
tion (4) should be conducted at a temperature
in the general vicinity of 600°C and at ele-
vated pressure.
This paper is confined to a consideration of
each of the three aforementioned combustion
technologies in a version which employs an
appropriate reaction selected from (1), (2),
and (3) for the capture of sulfur, along with
reaction (4) for the release of sulfur from the
solid.
IV-1-1
-------
Complete Combustion Approach
Hoy and Stanton1 have described a large-
scale experiment for fluidized-bed combus-
tion of coal at 6 atmospheres and 800°C.
They propose a pressure of 15 atmospheres
for the commercial embodiment of this tech-
nique. About 70 percent of the heating value
of coal would be transferred to boiler tubing
within the bed; the remaining 30 percent
would be available as sensible heat in the
800°C effluent gas, which would be expanded
in a gas turbine. The authors hope that alkali
and dust content of the gas will be acceptable
for its direct use in the turbine, so that equip-
ment need not be provided to cool the gas,
scrub it of alkali dust, and reheat the gas to a
suitable turbine-inlet temperature.
The CC>2 partial presure in gas leaving the
complete-combustion fluidized bed is suffi-
ciently great that CaCC>3 in dolomite added
to the bed would not decompose. Half-
calcined dolomite present in the bed would
react readily with SC»2, as in reaction (1).
Kinetics for this reaction, apparently on
account of half-calcined dolomite's porosity,
are favorable at temperatures even as low as
600°C.2'3'4
The team at The City College has demon-
strated that a way exists to regenerate
[CaCOs+MgO] from [CaSC>4+MgO] while
liberating H2S for sulfur manufacture.4 The
sulfate would be reduced to sulfide at about
800°C, thus:
[CaSO4+MgO] +4H2 =
[CaS+MgO] +4H20
(5)
The sulfide would be converted to the carbo-
nate by reaction with steam and carbon
dioxide at about 600°C, as in reaction (4).
Reactions (5) and (4), as well as reaction (1),
would be conducted at elevated pressure, suita-
bly the 15 atmospheres which Hoy and
Stanton proposed.
In our recent paper,4 we offered a scheme
employing methane reforming as a source of
IV-1-2
hydrogen for reaction (5). In the context of
Hoy and Stanton's proposal, reaction (5)
might advantageously be carried out in a small
auxiliary fluidized bed operating under reduc-
ing conditions. Solid would be transferred
from the main fluidizerf-bed boiler to the
auxiliary bed and thence to a fluidized bed
for sulfur desorption by means of reaction
(4).
It is important to note that the maximum
temperature for the dolomite-derived solids
would be 800°C in this scheme. There is
evidence5 that [CaO+MgO] does not sinter
and lose surface area at this temperature. We
are examining the question of the sintering of
other dolomite-derived solids, and have
grounds to hope that none of the solids will
undergo serious sintering under this complete
combustion approach. We would expect solid
reactivity to remain excellent through a large
number of cycles. The practical importance of
this fact will, of course, depend upon one's
degree of success in separating the dolomite-
derived solid from coal ash to be discarded.
Partial Combustion Approach
If partial combustion of coal were con-
ducted in a fluidized bed of fully-calcined
dolomite, sulfur would be absorbed by reac-
tion (3). Reaction (4) could then be used to
desorb sulfur. The resulting carbonated solid,
[CaCO3+MgO], could be returned to the
partial combustion bed, to be calcined
therein.
If all of the carbon in coal is reacted with
oxygen in air to form CO, a surplus of heat is
developed. The surplus heat could be used to
raise steam in boiler tubes passing through the
partial combustion bed or to heat combustion
air. Alternatively, heat from the formation of
CO could be used to support the endothermic
reaction of steam with carbon, steam being
supplied to the bed along with air. Adding
steam to the bed would, of course, increase
the latent heat loss from the overall power-
-------
generating system. This is not quite as bad as
might appear at first glance. Steam supplied
to the partial combustion step will eventually
reappear as steam at the gas turbine, and the
total flow of gas through the turbine will be
larger than if steam had not been supplied to
the partial combustion step. This means a
larger net production of power from the gas
turbine, and the cost of the gas turbine,
including air compressor, will be significantly
less than if steam were not used. This reduc-
tion in capital cost will provide an economic
offset against the reduction in efficiency
chargeable to the latent heat loss. The higher
the gas-turbine inlet temperature and the
higher the pressure, the less advantageous it
will be to use the excess heat from partial
combustion to raise steam for a steam cycle
and the more advantageous to use this heat to
decompose steam by reaction with carbon.
Partial combustion has potential advantages
over complete combustion:
1. About half as much gas would be pro-
duced, and so the fluidized bed would be
smaller.
2. The gas,' containing potential energy as
well as sensible heat, would represent a
larger proportion of the heating value of
the coal and would permit a second
combustion step in which the gas would
be heated to a much higher temperature
at the inlet of the gas turbine.
The drive for higher performance in aircraft
engines will continue, and experience from
such engines can be expected to maintain the
historic upward trend of temperature of gases
at the inlet of industrial gas turbines. As this
temperature rises, the second advantage of
partial combustion may appear progressively
more important.6
Partial combustion has difficulties, how-
ever:
1. Sufficient carbon must be present in the
bed to support the CC>2-carbon and
H2O-carbon reactions. Carryover from
the bed will contain carbon which must
probably be burned elsewhere, perhaps
best in an auxiliary bed for complete
combustion.
2. The temperature must be significantly
higher, probably at least 900°C, to cater
to the kinetics of the CC>2-carbon reac-
tion. The higher temperature increases
the danger that alkali in the gas will
injure the gas turbine, and therefore
increases the likelihood that gas will have
to be cooled and scrubbed.
If gas must be cooled and scrubbed to deal
with alkali from partial combustion, but not
from a total combustion at 800°C as Hoy and
Stanton hope, then the advantages of partial
combustion might disappear in a total eco-
nomic balance, at least for gas-turbine inlet
temperatures currently available.
When a temperature is selected for the
partial combustion bed, attention must be
paid to the problem of sintering of the dolo-
mite-derived solid. Preliminary work at The
City College confirms Haul's finding5 that
some sintering of [CaO+MgO] occurs at
temperatures in the vicinity of 925-975°C.
There are indications, however, that the solid
stabilizes at a surface area on the order of
about 5 m2 /gram. It must be emphasized that
this is a preliminary conclusion, but it would
appear reasonable to expect that a tempera-
ture below about 975°C should be compatible
with a relatively long life for the dolomite-
derived solid in a partial combustion bed.
If the gases from such a bed must be cooled
and scrubbed, there may be an advantage of
an ash-agglomerating fluidized bed operating
at a higher temperature and not incorporating
an agent to capture sulfur. Gases from the bed
at higher temperature will be lower in H2O
and CC>2 content, and hence will be admir-
ably suited for desulfurization by reaction
with half-calcined dolomite, as in reaction (2).
The gases should be cooled several hundred
degrees before the desulfurization step. We
have been surprised to discover that reaction
(2) is considerably faster, at a given tempera-
ture and H2S partial pressure, than reaction
(3). Accordingly, a panel bed filter7 is well
IV-1-3
-------
suited for carrying out reaction (2). Perhaps
such a filter, with the help of a filter aid, can
remove alkali dust from the gas to meet the
cleanliness required for the gas-turbine inlet.
It will be important to learn whether or not
this is so. The filter could operate at a tem-
perature below 800°C, and the life of the
filter solid should be excellent.
Carbonization Approach
One of the great inventions is the proposal
by Gorin, Curran, and Batchelor8 to desul-
furize coal char by cooperative action of
hydrogen and an "acceptor" for sulfur, such
as calcined dolomite. Under the proposal, the
acceptor and coal char would differ suffi-
ciently in particle size and density so as to be
readily separable.
This proposal's originality is appreciated if
it is recalled that the scientific basis for the
idea was available as early as 1923 from the
work of Powell9 of the U.S. Bureau of Mines.
The inventors' apparent major interest was
to provide a low-sulfur char for use in making
"formcoke" for metallurgy.
Theodore1 ° elaborated the idea to produce
a low-sulfur coke for power generation while
converting volatile matter in the coal to
pipeline gas.
I have examined the idea's potential for a
system intended simply to provide low-sulfur
fuels for power.11'12-13 Figure 1 depicts
broadly a scheme we are studying. The coal
would be carbonized at elevated pressure and
the resulting coke desulfurized by cooperative
action of hydrogen and the acceptor. Heat for
the distillation of volatile matter from coal
would be supplied by partial combustion of
the volatile matter, typically consuming 11
percent of the stoichiometric air for complete
combustion of the coal. Hence, equipment
volumes would he even less than in the partial
combustion approach. The quantity of lean
fuel gas is relatively small, and equipment for
its cooling and scrubbing, if necessary, would
be correspondingly small. The lean fuel gas
would drive a gas turbine, and a satellite
steam power station would receive waste heat
from the gas turbine's exhaust as well as from
the coal desulfurization process. An installa-
tion would operate at a steady coal feed and
would provide baseload power corresponding
to about a third of the heating value of the
coal. The installation would furnish low-sulfur
coke for production of variable power, either
in nearby equipment or at a distance.
POWER PRODUCTION FROM LOW-
SULFUR COKE
I will defer a discussion of the scheme it-
self, and first take up alternatives for the
production of power from the low-sulfur
coke.
Alternative 1
The coke might be burned in existing
power stations. Grinding the coke for pulver-
ized boilers might be costly, but using the
coke in stations having cyclone combustors
might be relatively easy, if the coke were
made from coal having ash suitable for such
combustors. Some gaseous or volatile fuel
might be needed to support combustion of
the coke, but many utilities have now
switched to low-sulfur oil, and would
welcome the opportunity to replace even a
part of this expensive oil with a cheaper fuel.
A modification of Figure 1 might provide a
suitable gaseous or liquid fuel produced from
the coal itself.
Alternative 2
Elliott14 has pointed out the advantages of
the fluidized-bed boiler for stations whose
purpose is to meet peakload requirements.
The coke would be an excellent fuel for such
stations, relieving their owners of the neces-
IV-1-4
-------
sity of providing equipment to cope with
sulfur. It would appear that the cost advan-
tage might belong to a design at high pressure,
such as that proposed by Hoy and Stanton.
Alternative 3
The efficiency of Hoy and Stanton's
scheme would be improved if gas from the
fluidized-bed boiler were heated to a higher
temperature ahead of the gas turbine by a
combustion of fuel derived from volatile
matter. This idea is depicted broadly in Figure
2. The figure assumes, perhaps too pessimis-
tically, that the gas from the fluidized-bed
boiler must be cooled and scrubbed to remove
dust. If this pessimistic assumption is correct,
the availability of a gaseous fuel to reheat the
boiler offgas would be welcome.
Alternative 4
The coke might be gasified in a partial
combustion process, providing fuel gas to
drive a gas turbine. A wider range of alter-
native gasification processes might be con-
sidered for the low-sulfur coke than for raw
coal. For example, a process using an ash-
agglomerating fluidized bed might seem more
attractive for the low-sulfur coke than for
coal. If the coke has a sufficiently large
particle size, it would be a good fuel for
Secord's slagging, gravitating-bed gasi-
fier.15-16
This alternative becomes attractive if a
major part of the coke is to be burned to pro-
vide baseload power, and if advantage is to be
taken of the higher gas-turbine inlet tempera-
tures expected to become available.
Alternative 5
The coke might be burned in a fluidized-
bed boiler using total combustion and heating
steam to temperatures well beyond those now
in use in conventional power equipment.
Elliott17 has called attention to this oppor-
tunity. In light of present knowledge, devel-
opment of a device for partial combustion of
coal or coke must be viewed as more uncertain
than development of a fluidized bed for total
combustion. It must also be admitted that the
expectation that "ultra-high" gas-turbine inlet
temperatures (i.e., beyond 1100°C) will
become available is probably not quite yet a
certainty. Under these circumstances, it
would be a mistake to overlook the oppor-
tunity which the fluidized-bed boiler offers to
improve steam cycle efficiency by raising the
steam temperature.
If the historical line of development of the
steam cycle is followed, even a modest in-
crease in steam temperature will be accom-
panied by a sharp increase in steam pressure. I
have called attention to a steam cycle employ-
ing an unusually large amount of super heat
and affording an efficiency comparable to
that of the conventional steam cycle for a
given maximum steam temperature, but not
requiring such high pressures.11'18'19 De-
tailed cost examination is still needed to judge
the worth of the new cycle, which should not
be altogether overlooked when considering
higher steam temperatures.
ACCEPTOR TECHNIQUE FOR DESULFUR-
IZING COAL
Powell's historic work9, confirmed and
extended by Batchelor, Gorin, and Zielke,2 °
teaches that the acceptor technique for desul-
furizing coal char will work best if the char is
carbonized quickly and moved promptly from
the carbonization step to the desulfurizing
environment. This fact is easily understood
when it is remembered that cokes sinter exo-
thermically at temperatures above about
600°C. If the coal char is allowed to sinter,
sulfur is locked into the coke structure, the
sulfur is inaccessible to reaction with hydro-
gen, and "deep desulfurization" of the coke is
IV-1-5
-------
impossible. In thermodynamic terms, sulfur in
a coke soaked for a long time at a high
temperature is characterized by a free energy
such that the pseudo-equilibrium between the
sulfur and hydrogen is highly unfavorable for
production of H2S.
The agglomerating tendencies of bitumi-
nous coal have posed problems for designers
of processes to carbonize such coals in fluid-
ized beds. Many American coals, particularly
in the East, are so strongly caking that
attempts to carbonize the coals in a simple
one-step fluidized-bed carbonization process
have often ended in failure caused by the
formation of massive agglomerates. Processes
have been devised to avoid massive agglomer-
ates by providing several fluidized carboniza-
tion stages at progressively higher tempera-
tures. Other processes would inject raw coal,
at a relatively low rate, either into a dilute
stream of hot recycled char or into a
mechanically mixed mass of such char. Either
dodge is poor: the first because of the length
of time during which the char is soaked at
high temperature; the second because of the
danger of back reactions returning sulfur to
the recycled char.
In an acceptor desulfurization process,
there will be an advantage if the coal is
ground quite fine, so that desulfurization is
not hindered by diffusion of the gaseous
reactants, hydrogen and hydrogen sulfide.
Figure 3 is a fuel desulfurization scheme we
are studying. Coal is supplied in a finely
divided condition, but coke particles are
allowed to grow large in a coke-self-agglom-
erating fluidized bed. Because the agglom-
erated coke is almost completely nonporous,
each fresh patch of coke (produced by
capture of a tiny sticky particle of coal) must
be desulfurized promptly after it forms upon
a coke pellet. To accomplish this, there is a
rapid exchange of coke between the agglom-
erating zone and a desulfurizing zone super-
posed upon the agglomerating zone. The
desulfurizing zone is a fluidized bed of finely
divided calcined dolomite. The fluidization
IV-1-6
velocity in the agglomerating zone is too high
to permit the dolomite acceptor particles to
sink deep into this zone, and the fluidization
velocity in the desulfurizing zone is too low
for coke pellets ejected into this zone to
remain there permanently.
Because of the small size of the acceptor
particles, they may advantageously be
calcined in the "recirculating fluid bed" de-
scribed recently by Reh.21 Such a bed is
indicated schematically for the calcination
zone of Figure 3.
Partial combustion of volatile matter
supplies heat to calcine the spent acceptor.
Partial combustion not only provides a lean
fuel gas to drive a gas turbine but is also
advantageous because the temperature in the
acceptor calcination zone may be lower, for a
given total operating pressure, than the
temperature which must be specified if com-
plete combustion of a fuel is used to provide
heat to this zone. The maximum temperature
necessarily reached by the solid in the calcina-
tion zone of Figure 3, for operation at 21
atmospheres, would be in the vicinity of
950°C. Acceptor reactivity and life should be
good.
Back reactions can restore sulfur only
superficially to the desulfurized coke product,
for coke pellets produced in a coke-self-
agglomerating bed will be almost completely
nonporous and hence only superficially in
contact with gases present in the carboniza-
tion step.
Only a pilot plant can establish with
certainty the operability of the coke-self-
agglomerating zone of Figure 3. We are
conducting fluidization tests at atmospheric
conditions which are beginning to elucidate
the principles governing behavior of a fluid-
ized bed of large particles coated with a sticky
substance. The work indicates that such a bed
can carry an astonishingly large burden of
sticky matter. This result, viewed together
with a number of successes elsewhere in the
operation of agglomerating beds, makes us
-------
confident of the operability of the coke-self-
agglomerating zone.
CONCLUSION
The reader may now object that this paper
has presented' too many proposals, and he
may wonder what I consider the best to
develop. I would reply that paper research is
cheap and the importance of the subject justi-
fies detailed engineering study and cost evalu-
ation of almost any plausible idea which
offers itself. The needs of the power industry
are so varied that a number of systems may
prove commercially viable.
Without such engineering studies, a choice
among the systems must be largely subjective,
and I can make it only by asking myself, on
which systems would I be happy to devote
substantial amounts of time and energy? I
would, today, pick three systems.
1. The acceptor desulfurization scheme of
Figure 3; it offers the overwhelming
advantage of providing low-sulfur fuel to
existing plant in a form which can be
stored and shipped.
2. A bed for complete combustion, either
of coal in presence of half-calcined dolo-
mite or of low-sulfur coke from the
acceptor desulfurization process.
3. An ash-agglomerating partial combustion
bed followed by a panel bed filter at
lower temperatures, with the bed con-
suming either coal or low-sulfur coke; if
coal is used, the panel would be charged
with half-calcined dolomite.
ACKNOWLEDGEMENTS
Work at The City College on reactions (2)
and (3) is supported by Research Grant No.
AP-00945 from the National Air Pollution
Control Administration, Consumer Protection
and Environmental Health Service. Work on
the panel bed filter is supported by Research
Grant AP-00692 from the same sponsor.
Contributions toward matter reported
herein have been made by Professors Robert
A. Graff and Robert Pfeffer and Messrs.
Melvyn Pell, Lawrence A. Ruth, Leon
Paretsky, S. Narayanan, Ralph Levy, F.
Farouk, Basil Lewris, Barry Hertz, Steven
Weinstein, and Gary Weil. Messrs. John
Bodnaruk and George Dilorio helped greatly
with experimental arrangements.
BIBLIOGRAPHY
1. Hoy, H. R. and J. E. Stanton. A.C.S. Div.
Fuel Chemistry Preprints 14 (2): 59, May
1970.
2. Coke, J. R. Ph.D. thesis (mentor: M. W.
Thring), University of Sheffield, England,
Dept. of Fuel Tech., May 1960.
3. Bertrand, R. R., A. C. Frost, and A.
Skopp. Fluid Bed Studies of the Lime-
stone Based Flue Gas Desulfurization
Process, report from Esso Research and
Engineering Co. to NAPCA, October
1968.
4. Squires, A. M. and R. A. Graff. Panel Bed
Filters for Simultaneous Removal of Fly
Ash and Sulfur Dioxide: III. Reaction of
Sulfur Dioxide with Half-Calcined Dolo-
mite. Paper presented at meeting of Air
Pollution Control Association, St. Louis,
Missouri, June 1970.
5. Haul, R. A. W. Z. Anorg. Allgem. Chem.
281: 199,1955.
6. Robeson, F. L. and A. J. Giramonti.
A.C.S. Div. Fuel Chemistry Preprints 14
(2): 79, May 1970.
7. Squires, A. M. and R. Pfeffer. J. Air Poll.
Control Assoc. 20 (8): in press, August
1970.
IV-1-7
-------
8. Gorin, E., E. P. Curran, and J. D. Batch-
elor. U.S. Patent 2,824,047, February 18,
1958.
9. Powell, A. R. J. Am. Chem. Soc. 45: 1,
1923.
10. Theodore, F. W. Low Sulfur Boiler Fuel
Using the Consol CC>2 Acceptor Process:
A Feasibility Study. Report from Con-
solidation Coal Company to Office of
Coal Research, November 1967.
11. Squires, A. M. A Role for Fluidized Com-
bustion in Clean Power Systems. Paper
presented at First International Confer-
ence on Fluidized-Bed Combustion,
sponsored by NAPCA, Hueston Woods,
Ohio, November 18-22, 1968.
12. Squires, A. M., R. A. Graff, and M. Pell.
Desulfurization of Fuels with Calcined
Dolomite: I. Introduction and First
Kinetic Results. Paper presented at
Washington, D. C. meeting of A.I.Ch.E.,
November 1969, to be published in
A.I.Ch.E. Symposium Series.
13. Squires, A. M. U. S. Patent 3,481,834,
December 2, 1969.
14. Elliott, D. E. Can Coal Compete? The
Struggle for Power. Inaugural lecture,
The University of Aston in Birmingham,
England, November 20, 1969.
15. Secord, C. H. U.S. Patent 3,253,906, May
31, 1966.
16. Hoy, H. R., A. G. Roberts, and D. M.
Wilkins. Inst. Gas Engrs. J. 5: 444, 1965.
17. Elliott, D. E. J. Inst. Fuel. 43: 258, 1970.
18. Squires, A. M. Clean Fuel Power Cycles,
ASME paper 67-WA/PWR-3, 1967.
19. Squires, A. M. U.S. Patent 3,436,909,
April 8, 1969.
20. Batchelor, J. D., E. Gorin, and C. W.
Zielke. Ind. Eng. Chem. 52: 161, 1960.
21. Reh, L. Highly Expanded Fluid Beds and
Their Applications. Paper presented at
Puerto Rico meeting of A.I.Ch.E., May
1970, and to appear in Chem. Eng. Progr.
IV-1-8
-------
- AIR A|
W ^ COMPR
i
AIR AT
R GAS
ESSOR TURB|NE »- BASELOAD
I
HIGH PRESSURE
I
COAL DIS1
PARTIAL CC
OF VOLAT
DESULFURIZA1
. , " I
UUMliUoTIQN '
A STEAM POWER
LEAN STATION
ILLATION
5MBUSTION
LE MATTER
riON OF COKE
^ COAL f
*
HIGH-PRESSURE STEAM
LOW-SULFUR COKE TO POWER STATIONS AT A DISTANCE
(VARIABLE POWER)
Figure 1. Scheme for providing low-sulfur fuels for power.
COMBUSTION
OF FUEL GAS
COAL DESULFURIZATION
PROCESS
COAL
AIR
Figure 2. Lean fuel gas from coal desulfurization used to raise the temperature of
combustion products from a coke-fed pressurized fluid-bed boiler.
IV-1-9
-------
LEAN FUEL GAS
IN SULFUR)
AIR
DESULFURIZING
ZONE
CaO
CaS
H2S
CaC03
CaS
CaCOo
SULFUR
DESORPTION
CLAUS
SYSTEM
STEAM AND CO2-
AIR
COAL HYDROCARBONIZING
AND COKE SELF-
AGGLOMERATING ZONE
ELEMENTAL
SULFUR I
TO MARKET W
RECYCLE OF PORTION
OF LEAN FUEL GAS
(CONTAINING HYDROGEN)
LOW-SULFUR
COKE PELLETS
Figure 3. Acceptor process converting coal into low-sulfur fuels.
IV-1-10
-------
2. FLUIDISED-BED COMBUSTION
AND THE
DESIGN OF BOILERS
S. J. WRIGHT
National Coal Board, England
INTRODUCTION
Conventional electricity-generation boilers,
whether fired by pulverised coal or oil, are
characterised by very large combustion cham-
bers. The size is dictated: first, by the need to
allow sufficient time for combustion despite
temperatures in the range 1200-1600°C; and
secondly, by the need to allow molten and
potentially corrosive ash constitutents to
resolidify before coming into contact with the
banks of convective superheater tubing. The
result is a supporting structure more than 200
ft high, for a 660-MW boiler, costing some 14
percent of the total boiler cost with, in addi-
tion, extensive civil works to accommodate
the substantial point loads.
Fluidised beds have two properties which
allow a radical rethinking of boiler design
concepts.
1. The rapid movement of particles
within the fluidised bed gives rise to relatively
high rates of heat and mass transfer between
the solids and the fluidising gas. The result is
that coal can be burned in a fluidised bed of
some 2-3 ft deep at volumetric heat release
rates in excess of anything achievable by
conventional methods and at mean bed
temperatures between 800 and 850°C. At
these relatively low temperatures: ash consti-
tuents do not melt; almost all the alkali
metals, known corrosion promoters, are
retained in the bed; and, by the addition of
limestone to the bed, a substantial proportion
of the sulphur content of the fuel can be
FV-2-1
retained as CaSO4- Thus the need for a large
combustion volume is eliminated.
2. The rapid movement of the particles
within the fluidised bed gives rise to relatively
high heat-transfer coefficients between the
bed and surfaces immersed in it, up to 5 times
greater than those achievable in conventional
gas-tube heat exchangers. The environment of
a fluidised-combustion bed allows steam-
raising tubes to be placed in the bed. Thus,
substantial savings can be made in the length
of tubing required to extract a given amount
of heat, and possibly, the less corrosive
environment will allow less expensive alloys
to be used in the superheater and reheater
sections.
DESIGN OF FLUIDISED-BED BOILERS
The most important decision in the design
of a fluidised-bed boiler operating at atmos-
pheric pressure is the choice of fluidising
velocity. The fluidising velocity fixes the
oxygen availability, and hence heat-release
rate, per unit cross-sectional area of bed
which, in turn, fixes the total cross-sectional
area required for a given steam duty. There-
fore, considered in isolation, the capital cost
of the boiler falls with increasing fluidising
velocity because of savings in containment
costs.
However, two factors oppose the incentive
to increase fluidising velocity.
-------
1. As the fluidising velocity is increased, a
progressively coarser size spectrum of bed
particles must be used in order to limit elutri-
ation losses. As the mean particle size of a
fluidised bed increases, the heat-transfer
coefficient within it decreases: e.g., as the
mean particle size increases from 0.4 mm
(0.016 in.) to 1.0 mm (0.039 in.), the coeffi-
cient falls about 35 percent. The result is a
tendency for the bed depth, and hence pres-
sure drop and running costs, to increase in
order to accommodate the increased area of
tubing required in the bed; this is in addition
to the increased capital cost of tubing.
2. As the fluidising velocity increases,
elutriation rates, and hence carbon loss from
the bed, tend to increase due to the increased
agitation and degradation within the bed, and
residence time in the combustion zone tends
to decrease. Therefore, larger and more elabo-
rate recycle systems are required to achieve
any particular combustion efficiency with a
consequent increase in both capital and
running costs.
Thus the choice of fluidising velocity
becomes an optimisation of conflicting eco-
nomic incentives.
If, on the other hand, the boiler is pressur-
ised, an additional degree of freedom is intro-
duced in that, for a given velocity and a given
cross-sectional area, the heat release may be
increased by increasing the pressure, and
hence oxygen availability. Therefore, for a
given steam output the containment cost can
be reduced by increasing the pressure. There
is no heat transfer penalty because velocity,
and hence bed size spectrum, remains con-
stant; however, bed height, and hence pres-
sure drop, increases because more heat must
be extracted from a bed of fixed cross-
sectional area, and there is a limit to the
density of tube packing. The apparent insen-
sitivity towards operating under pressure is,
however, offset by the increased plant com-
plexity imposed by the pressure and also by
the availability of industrial gas turbines
designed for a suitable pressure ratio and
capable of being modified without incurring
substantial development charges. Although
the remainder of this discussion will consider
only atmospheric pressure boilers, much of it
will be relevant to the case of operation under
pressure.
Having chosen a fluidising velocity, the
cross-sectional area of the unit is fixed for
a given steam output. Because the bed
temperature will be limited to between
800 and 900°C by ignition and ash fusion
criteria, respectively (at least for English
coals), the required distribution of heat-
transfer surface between the fluidised bed and
conventional surfaces above the bed may be
calculated. Typically some 60 percent of the
heat release must be extracted directly from
the bed if a temperature of about 850°C is to
be maintained.
There is a cost incentive to place those
parts of the steam/water circuit where the
more exotic materials have to be used in the
fluidised bed. If carbon steel has a cost index
of 1.0 (under English conditions), 2 percent
Cr has 2.4, 9 percent Cr has 5.4, and
Austenitic 316 has 13.4. Therefore, the
incentive is to put the superheater and re-
heater in the fluidised bed and make up any
deficiency with evaporator surface leaving
the water walls and any gas-space surface to
the remaining vaporisation and the econo-
miser. However, under part-load conditions,
the downsteam ends of in-bed reheaters and
superheaters can get relatively hot, requiring
the use of austenitic steels where ferritic steels
are adequate at the continuous-rated output.
Thus, particularly at relatively low fluidising
velocities with consequently high heat-trans-
fer coefficients, there may be an economic
incentive to place the downstream sections of
the reheater and superheater in the gas-space
above the bed and avoid the use of expensive
steels.
Finally, there is an incentive (at least under
English conditions) to maximise factory
IV-2-2
-------
construction, thereby minimising site work
and the consequent scope for construction
errors—all of which reduce the time of con-
struction and hence the interest charges
incurred. Interest during construction can be
as much as 26 percent of the capital cost of a
2000-MW power station. Fluidised-bed com-
bustion presents, for the first time, the possi-
bility of designing a large boiler as a number
of semi-independent units which can be
factory-assembled, leaving the minimum of
connecting-up on site. For English road trans-
port conditions, the maximum transportable
size is about 14 ft wide x 40 ft long x 16 ft
high; however, for a coastal power station,
sizes as large as 25 x 44 x 22.5 ft can be
brought in by sea. As an example, using road
transport criteria, a 120-MW boiler operating
at a fluidising velocity of 8 ft/s can be de-
signed as five factory-assembled beds (two
evaporator, two reheat, and a superheat) with
a secondary superheater and an ecomoniser in
the gas-space above the bed. This arrangement
is estimated to save 9 months in construction
time. With high interest charges, the incentive
to utilise factory assembly can be strong
enough to decide a fluidising velocity at the
expense of say a slightly lower combustion
efficiency.
CONCLUSION
The foregoing shows that many of the
potential advantages of fluidised-bed combus-
tion, for reducing the capital cost of boiler
plants, are self defeating unless the design is
optimised in respect to each variable.
The design should:
1. Obtain maximum heat-transfer coeffi-
cients in the bed and the bulk of the
boiler surface operating at metal
temperatures in excess of 400°C
immersed in the bed.
2. Obtain the maximum combustion effi-
ciency by maximising the residence time
of gas and particles in the bed and the
combustion space above it.
3. Obtain maximum factory assembly in
packaged units.
4. Reduce structural and civil work to a
minimum.
5. Reduce to a minimum the lengths of
high-pressure pipework connections.
Design aims 1 and 2 argue for low fluidising
velocities; aims 3, 4, and 5 argue for high
fluidising velocities. Thus, if full advantage is
to be taken of the potential of fluidised-bed
combustion for reducing the cost of power
station plant, an economic optimum must be
sought whereby the fullest advantage is taken
of all possible savings.
ACKNOWLEDGMENTS
The work described in this paper was
carried out as part of the research programme
of the Research and Development Depart-
ment of the National Coal Board. Views
expressed are those of the author, not neces-
sarily of the board.
IV-2-3
-------
3. IGNIFLUID BOILERS
FOR
AN ELECTRIC UTILITY
R. H. DEMMY
UGI Corporation
INTRODUCTION*
The UGI Corporation operates an electric
utility (Hunlock) in Northeastern Pennsyl-
vania, serving an area of 440 square miles near
Wilkes-Barre, along the Susquehanna River.
UGI's Hunlock Generating Plant (88,700-kw
net) burns anthracite from the Northern and
Middle Pennsylvania anthracite fields on
Barley stokers and in pulverized-fired boilers.
This plant's latest addition, completed in
1959, is a Foster Wheeler 400,000 Ib/hr
pulverized boiler utilizing anthracite bank
material with a quality of approximately
9000 Btu/lb. Stack emission is controlled by
electrostatic and mechanical precipitators
installed by Research-Cottrell.
UGI's Planning Department indicated that
new generation was required in the early
1970's, both to meet new sales and to replace
old facilities. To satisfy this projected need
for new generation, we investigated the possi-
bility of burning fuels adjacent to our service
territory. We found that Babcock-Atlantique
of France had a boiler utilizing the Ignifluid
combustion process which would be capable
of burning the anthracite culm banks of our
area (there are 100,000,000 tons within 40
miles of our projected plant site).
Small samples of the culm bank material
were sent to France: first, four 5-lb bags; later
40 tons in 55-lb bags. When preliminary
The technical assistance of P.A.Mulcey, fuel consultant of
Dallas, Pennsylvania, is acknowledged in the development of
this paper.
laboratory tests indicated that a full-scale
combustion test should be attempted,
approximately 800 tons of 40-percent ash,
7500 8000 Btu/lb culm bank fuel was
shipped to an area in the French Alps near
Grenoble, where the Cartonneries de la
Rochette operates a cardboard factory.
OBSERVATIONS OF PRODUCTION TESTS
Today, 1 would like to discuss with you our
observations in both Casablanca, Morocco,
where the largest Ignifluid boiler is located,
and in La Rochette, France, where the com-
bustion tests were performed.
In Casablanca
Arriving in Casablanca late in the after-
noon, we were concerned that we would be
unable to visit the plant that day. We pro-
ceeded immediately to the power plant,
approximately 1 mile from the center of Casa-
blanca.
As we approached, we were shocked to see
that the stack indicated no power generation.
Because we were approaching an anthracite
plant we had anticipated what we have experi-
enced in the United States in the way of
anthracite stacks; seeing the stack completely
without any visible emission, we were quite
upset that we had travelled such a distance to
find the plant not operating.
IV-3-1
-------
When we arrived in the control room, the
Moroccan operators told us that the plant was
at full load, a fact which was difficult for us
to accept. But the recording charts indicated
that the plant, with a maximum capacity of
60 mw, was operating at its normal capacity,
approximately 55 mw, and that the boiler had
been operating at approximately the same
output most of the day.
The stack emission control for this boiler is
a Lurgi electrostatic precipitator preceded by
a mechanical precipitator. The precipitator
was to have been a three-cell unit but the
Moroccans, without sufficient money to buy
all three cells, purchased two internals and the
shell for the third. Results were so successful
that they have not added the third electro-
static section.
The dust removal from these units is similar
to what we are used to in the States; a hydro-
vac or similar system is used. The one differ-
ence is that all of the dust which is collected
by the precipitators is reinjected into the
furnace. There is no fly ash pond. All ash
from the Ignifluid boiler is removed as
bottom ash, and sluiced into the Atlantic
Ocean which, at the site of the power plant, is
400-ft deep.
The burnout of the carbon is quite com-
plete: 4 to 6 percent carbon is left in the ash.
Electrostatic precipitation is successful
because the fly ash of an Ignifluid boiler is
high in carbon and therefore low in resistivity
which is the opposite of the pulverized boilers
operated in the States. Therefore, a clean
stack is not a desire but a fact. For an
operator of an anthracite-fired plant or a
bituminous coal low-sulfur plant which would
again have high resistivity, it was indeed hard
for us to believe our own eyes when we saw a
boiler at full load with an optically clear
stack.
In La Rochette
Visiting the site of our production tests in
the French Alps at La Rochette, we found
that the Ignifluid boiler there has no electro-
static preci- pitators, but two mechanical
precipitators. The stack emission at that site
was similar to the best emission we have at
Hunlock Plant, utilizing electrostatic precipi-
tators installed last year.
La Rochette's coal is transported by rail
from Antwerp, Belgium, and dumped directly
into the silos at La Rochette.
The developer of the Ignifluid process
(described more fully at the end of this paper)
is Ms. Albert Godel, of Socie'te' Anonyme
Activit, who spoke at the First International
Conference in 1968, and is still very active
and interested in this process. Babcock-
Atlantique has purchased the patents from
him and are directing the development pro-
gram.
A steam flow of 100,000 Ib/hr from a grate
area approximately 3 ft wide by 25 ft long
(compared to 100,000 Ib/hr from a standard
anthracite spreader stoker 20 ft wide by 28 ft
long) indicates the improvement in the state-
of-the-art of burning anthracite.
The tests were under the supervision of
CERCHAR, the French equivalent of the U.S.
Bureau of Mines and Bureau of Standards,
combined.
The continuous removal of the ash from
the Ignifluid boiler is accomplished by an
inclined grate rising at an angle, of 10-12
degrees. The level is maintained in Ifte fire bed
by having variable air pressure supplied to the
various zones of the fire by separate air ducts.
In this boiler there are six zones; the
highest pressure is where the raw fuel is
supplied to the boiler by a Redler feeder. The
pressure decreases as the fuel progresses
toward the front of the furnace; the area
IV-3-2
-------
where the ash rises out of the fluidized bed
requires additional air for final combustion of
the carbon in the ash.
During our test, the combustion was
stopped due to a plugging of the coal feeder
and we were able to witness a reignition of
boiler. The coal is supplied to the small grate
in a manner similar to the way a spreader
stoker would receive its coal; the fire is
started in the same manner, with excess air
supplied until the fire starts effervescing.
When this effervescent condition is more or
less uniform over the entire grate, which is
not moving at the time, the bed is then
fluidized by the introduction of sufficient air
to cause the particles to rise from the grate
and be balanced by the upward flow of the
primary air and the action of gravity on the
particles.
Air distribution in the boiler does not seem
to be critical. In this boiler, ambient air
(approximately 50 percent primary air and 50
percent secondary air) is introduced just
above the fluidized bed in the same zone as is
the reinjected fly ash. When the bed is fluid-
ized, the furnace looks like a gas furnace with
an incandescent white heat.
An inspection of the internals of this boiler
indicated the tubes to have no slag accumu-
lation. The 2-year-old grate looked brand
new, undoubtedly because of the introduc-
tion of primary air to keep the grate from
reaching high temperatures. The ash carbon
burnout for our tests, as was predicted in
model tests in the laboratory, was not as high
as those in Casablanca (4-6 percent) and La
Rochette (6-8 percent): the carbon in the ash
of the Pennsylvania culm was 17-20 percent.
In the final design of the proposed boiler,
burnout to a level of less than 10 percent will
be accomplished by a post-combustion fur-
nace similar to the type utilized for expanded
ash for lightweight aggregate purposes.
PROPOSED NANTICOKE STATION
After evaluating the combustion tests in
France and completing the overall engineering
evaluation, it was determined that a
300,000-kw station could be built utilizing
two Ignifluid boilers and one turbine-
generator. A unit size of 300,000 kw was
selected due to the availability of units of this
size already operating. To further enhance the
availability, two boilers are being used. Since
the major cause of loss of generation is boiler
failure, the use of two boilers will allow
operation at half load while repairing the
failed boiler. Steam conditions at the throttle
of the turbine will be 1800 psig
1000°F/1000°F. The turbine will also be
supplied with the option for extraction:
adjacent to the power plant site, the
Commonwealth of Pennsylvania is planning to
install a pilot plant for acid mine-water
demineralization which will require approxi-
mately 300,000 Ib/hr of low-pressure steam.
The fuel for this plant is above ground and
available. No deep mining of coal will be re-
quired for the proposed 30-year life of the
plant. The fuel will come from existing
anthracite culm banks (refuse from the
preparation of anthracite), 100,000,000 tons
of which are located within 40 miles of the
plant site. This fuel will be upgraded to
approximately 8000 Btu/lb, a heating value
which the boiler has proven itself capable of
handling. Material not taken to the power
plant will be left at the bank sites and
regraded into usable contours so that all of
the present areas which are not usable as real
estate today will be available for real estate
development. The high ash content (40
percent) of the fuel would normally cause a
considerable problem in handling; however,
tliis is not anticipated as the ash has proven
itself to be excellent for utilization in cinder
blocks and has potential use as a lightweight
building aggregate.
We believe that this plant will meet the
ecological requirements of today's society.
IV-3-3
-------
1. We do not anticipate participate emis-
sion problems.
2. Since the fuel has less than 0.5% sulfur,
there will be no SCH emission.
3. NOX emission will be reduced. Tests in
France showed that the NOX emission of
an Ignifluid boiler was approximately
half that of a pulverized boiler firing the
same coal.
4. Thermal discharge to the Susquehanna
River will be controlled by the utiliza-
tion of cooling towers.
THE IGNIFLUID PROCESS APPLIED TO
UTILITY BOILERS*
the stoker, as is the case with conventional
stokers or spreaders.
An Ignifluid-fired boiler burns any type of
small fuel: sized from 1/8 in.-O to 3/4 in.-O,
from anthracite to bituminous coal, coking or
not. and regardless of the ash content. The
efficiency is as high as with pulverized firing.
Mechanically, the Ignifluid unit is similar to
the chain or traveling-grate stoker, including
as it does the same drive and return sprockets,
traveling grate or chain, and wind-box
arrangement consisting of six or more
compartments.
The Ignifluid process is a fluidized-bed
method of combustion of small coal which is
maintained in stable aerodynamic suspension
under the effect of an ascending air current
which gives the layer the appearance of an
incadescent boiling liquid, hence the trade-
mark Ignifluid.
The process utilizes a standardized fluid-
ized combustion burner at the base of the
boiler. Consisting of a very narrow inclined
stoker of less than one-tenth of the total
section of the boiler, it supports the fuel bed
which is fluidized by a controlled primary air
current; it also extracts the clinker from the
bed in its ascending rearward motion.
Secondary air, blown over the top of the
fluidized bed, completes combustion. All the
grit carried over is cycled back on the surface
of the fluidized bed, with the result that all
the ash from the coal is extracted as clinker
by the chain-grate.
A high rate of combustion is obtained (over
300 Ib of coal per sq ft) because combustion
takes place in the whole 2-ft-deep mass of the
turbulent bed and not only on the surface of
*The assistance of F.W. Kuehn, engineering consultant of
AUentown, Pennsylvania, is acknowledged for this portion
of the paper.
Essential Differences
The essential differences are:
1. The grate width is only about 10 percent
that of the traditional anthracite stoker.
2. The grate is tilted 8-12 degrees from the
horizontal, in the direction of travel of
the grate, as contrasted to the level
stoker.
3. There is a different under grate air-
pressure pattern from front to back-
substantially a complete reversal—with
the maximum (main duct) pressures
about 50 percent higher.
4. The coal feed is different: instead of
being carried by the stoker from a feed
hopper into the furnace under a leveling
gate, coal is dropped by a Redler con-
veyor into a steep-angled pipe which
intersects the furnace front wall a few
feet above the grate. Air injected
through the bottom of the feed pipe
tends to spread the coal outward and
forward as it falls toward rJie fluidized
bed which, in turn, has an inherent
spreading proclivity due to the boiling
action.
On a stoker, the wind-box pressures usually
range drom 1/2-in. water in the first compart-
ment, to a maximum of 4-5 in. in the fourth,
and grading off to 3 in. in the sixth. In the
IV-3-4
-------
Ignifluid unit, pressure is highest (about 8 in.
of water) in the first compartment; it grades
down uniformly to 1 in. in the sixth, where
the coal bed is fluidized in normal operation,
as distinguished from start up (which will be
explained).
As might be surmised, with a chain width
only 10 percent that of the stoker, the
burning or reaction is much more intense.
Startup
To understand the action, let us run
through a startup. A small amount of coal is
fed into the unit so it lies on the grate to a
depth of 6-12 in. at the front. The induced
draft fan is started. Next, oil or gas torches are
lighted and directed down onto the surface of
the coal until it is uniformly ignited. The
forced draft fan is started and undergrate
(called primary air) air pressure is slowly built
up in the compartments as indicated, with a
straight-line sloping pattern from maximum
pressure in the first (or front) compartment
to a minimum in the last.
As the flow and pressures are increased, a
point is reached where the coal lifts off the
grate and suspension firing begins. At this
point, the fluidity action is rather unstable,
and the air flow must be increased further at
constant pressures to somewhere under the
point in velocity where most of the particles
would have left the fluidized bed (become gas
borne). Fortunately, because there is a wide
band of velocities possible between the low or
unstable point at which the bed tends to
collapse, and the high point where the bed
would be carried off with the gas, the opera-
tion is inherently very stable.
Under normal operation then (from full
load down to about 15 percent of rating), the
top surface of the fluidized bed is level, look-
ing like boiling lava. Depth above grate ranges
from a maximum of about 20 in. at the front
(over the first air compartment) to nothing
but the clinker thickness over the fifth air
compartment of a six-compartment unit.
When the coal particles are all consumed,
the remaining bits of fine ash become sticky
and agglomerate with others forming dense
clinker which falls on to the grate and is
carried over the rear sprocket in a ribbon 1-3
in. deep.
Blankets of dead coal cover and protect the
sloping refractory walls along the sides and
front of furnace, below the water-cooled
walls. The dead coal layers are maintained
because there is no air flow in these areas to
promote and support combustion.
Vertical water jackets, extending from the
top grate line down to the refractory at base
of the crickets outside the wind boxes, form a
barrier between the moving hot clinker on the
grate and the dead coal; they also support the
latter.
An air seal, at the interface between the
bottom of structural bridge across the furnace
and the top of the moving chain grate just
ahead of the first undergrate air compart-
ment, prevents any loss of siftings in this area,
especially during startup.
Very little loss of fuel in the form of sift-
ings is experienced through the grate, it is
claimed, because of the higher undergrate air
pressures used (as compared to conventional
stokers) and to the small air openings. Individ-
ual stoker chain links are notched in the
middle to facilitate separation of the clinker
at the rear sprockets.
Secondary air, which in larger units is
usually preheated to higher temperatures than
the primary air, is introduced into the furnace
through side walls a few feet above the
fluidized bed to complete combustion of the
fuel rich gases.
An inclined air nozzle under the rear arch is
used to blast any fine coal particles which
IV-3-5
-------
drop onto the surface of the clinker bed back
into the active combustion zone.
Control
Control is essentially conventional, and
typical of later automatic stoker control in
the States except for a unique fuel-bed thick-
ness probe and controller. Primary air is
controlled by deviation of a master steam-
pressure controller from a set point, steam
pressure being sensitive to load changes. As
primary air flow increases or decreases under
direction of the master controller, it changes
the rate of coal combustion and, therefore,
the depth of the fluidized bed. A water-
cooled fuel-bed thickness probe, inserted in
the bed just above the grate near the front of
the furnace, monitors the thickness of the
bed, and passes its signal continuously to a
fuel-bed thickness controller. The controller,
set to hold the fluidized bed at a constant
thickness, passes on the deviation-from-
normal signal to the fuel-bed controller.
Secondary air flow is trimmed by desired-
percentage CC>2 in furnace exit gases. Judging
from typical air duct sizing, secondary air
appears to be about 50 percent of the total
air. Induced-draft fan dampers, in turn, hold a
constant furnace draft of minus 5 mm.
Operating Advantages
Operating advantages claimed for the
Ignifluid unit over the stoker, which taken
together account for its ability to efficiently
burn low-grade coals containing up to 50
percent ash, include:
1. Automatic spreading of coal inherent.
2. Clinkers sink through fluidized bed to
grate.
3. Intimate and intense reaction of air and
fuel.
4. Fluidized bed completely pervious to air.
(With a bed in a fluid state, it is obvious
that the cleanliness of the coal is not
important; in fact, the coal can be
unwashed.)
5. Loss of sittings through air holes in grate
chain link is not as great as in conven-
tional stoker firing because of higher air
pressures required by the process which
in turn means smaller air holes through
the grate. (Claimed riddling or sifting
losses are as low as 0.5 percent.)
Unusually high thermal efficiencies (up-
wards of 88 percent) are claimed, even with-
out air preheating when using a medium
volatile (6 percent), reasonably well prepared
(16.4 percent ash), fairly dry (6.6 percent
moisture) anthracite of proper sizing (zeruto
5 mm). Boiler efficiency of 81.74 percent was
guaranteed for the UGI units.
As the coal becomes poorer (i.e., higher in
ash), a significant higher loss to be reckoned
with is that in the sensible and latent heat loss
in the clinker, part of which will be reclaimed
with a post-combustion furnace.
Forced outage rate for the 55-metric ton
evaporation per hour Ignifluid burning unit at
La Rochette (given to Socie'te Babcock &
Wilcox by the owner-operator paper
company, Cartonneries de La Rochette) for
the year August, 1966 to August, 1967 was
2.14 percent: 0.3 percent was due to the
stoker alone; 1.84 percent was for such things
as repairs to ash-reinjection system, coal-
handling system breakdowns, and replace-
ment of belts on the forced draft fan.
CONCLUSION
I
The ecological demands of the United
States today are forcing electric utilities not
to consider coal-fired plants due to the severe
problems of electrostatic precipitation of high
resistivity ash which results from either
pulverized-anthracite or low-sulfur-bitumi-
nous firing. With insufficient natural gas
available, the utilities are leaning toward low
sulfur oil and nuclear material as the future
IV-3-6
-------
sources of energy for electrical generation. If
the coal industry is to participate in the
electric utility energy market, a modification
in the method of burning fuel is indicated. We
believe that the fluidized bed as proposed by
Babcock-Atlantique is the only method of
fuel combustion available today in sizes which
can be upgraded to the large steam produc-
tion requirements of the electric utility
industry of the United States. Although this
method is available in commercial size today,
we fear that the delays inherent in developing
commercial pressurized fluidized beds will
prohibit their commercial development in the
United States. However, if the Ignifluid
process is introduced into the United States,
we will have fuel technologists knowledgable
in fluidized combustion capable of guiding
the development of the next step forward in
this technology; namely, pressurized fluidized
beds.
IV-3-7
-------
4. THE MODULAR
FLUIDIZED-BED BOILER
CONCEPT
J. W. BISHOP
Pope, Evans and Robbing
ABSTRACT
The development of the Pope, Evans and
Robbins modular fluidized-bed boiler concept
is described. The early and unsuccessful
attempts to design a fluidized-bed boiler along
more "conventional" lines are also described
and the lessons drawn from these failures
presented. A preliminary conceptual design
and cost estimate is presented for an electric
utility boiler having a capacity of 300,000
pounds of steam per hour.
INTRODUCTION
The purpose of this paper is to describe the
experimental and design history leading to the
presently proposed modular design concept of
a fluidized-bed boiler, along with some details
of the most recent concept of a 300,000 Ib/hr
small utility unit.
EARLY HISTORY
In 1965, as part of our program to extend
the capacity of packaged coal-fired boilers,
the late LeRoy F. Deming visited various
European combustion and boiler developers,
including BCURA. It was obvious that
fluidized-bed firing and direct transfer of heat
from the bed to the heat transfer surface
represented the most promising system for
providing a high-capacity coal-burning boiler.
Imbedded-Tube Concept
The first effort was to develop a boiler
concept as a guide for the ensuing experi-
mental work. Using BCURA's initial project
of 1,000,000 BTU/hr/ft2 of bed area, the
boiler concept of Figure 1 was developed.
This concept consisted of a continuous
imbedded-tube bank unit with a longitudinal
steam drum and three lower headers. Coal was
to have been screw-fed at two locations in the
hope that the fluidizing air issuing from the
grid would lift and distribute the coal.
The atmospheric steam boiler of Figure 2
was then built, originally with a longitudinal
screw-feeding mechanism and a perforated
plate grid. Of course, the air did not distribute
properly and the coal merely caked above the
screw.
Water-Cooled Column
We then digressed to a water-cooled
column wherein the air distribution problem
was solved: first, with a "rock sandwich" grid
(i.e., a layer of rock between two perforated
plates); later, with the "button grid" cur-
rently in use. This system directs air down
onto the grid plate, precludes dead spots,
prevents sifting into the plenum, and permits
operation at over 2000°F, a condition which
has not been approached with other grid plate
designs. Relying on BCURA's experience, we
IV-4-1
-------
screw-fed the coal into the column with air
jets blowing the coal off the screw, but we
were unable to duplicate the BCURA opera-
tion. We then developed a system for
dropping coal into a pneumatic transport line.
Coal-Distribution Problems
The boiler of Figure 2 was placed in opera-
tion, but problems of coal distribution were
encountered in the presence of the imbedded
tube bank. Any agglomeration of particles
between tubes rapidly formed clinkers which
stuck between the tubes. Although we could
remove these clinkers, by lifting the boiler off
the grid plate plenum assembly, it was
obvious that this could not be done with an
actual boiler. Erosion was heavy on the lower
bank of tubes. The lessons learned here were:
1. Avoid use of imbedded tube banks if
possible.
2. If imbedded tube banks are required,
consider the effect on fuel distribution,
access for cleaning, and the minimum
height above the distributor to preclude
erosion. (Design for heat removal is only
one of a number of considerations.)
The imbedded tube bank system was
abandoned.
The design of the atmospheric-pressure
boiler had been predicated on a heat transfer
coefficient between the bed and the im-
bedded tubes of 50 Btu/hr/ft2/°F. It was
found, however, that this coefficient was not
useful in reproducing either the furnace exit
gas temperature or the bed temperature. It
was necessary to almost double the value of
the measured coefficient to reproduce the
performance with a given imbedded surface
area. A sizable fraction of the released energy
was removed from the bed by mechanisms
other than direct contact heat transfer. It was
necessary (as in the case of the later BCURA
shell-and-tube boiler experience) to cut and
remove tubes so as to reduce the heat removal
rate so that, in turn, a deeper bed could be
IV-4-2
operated in which good combustion could be
achieved. The lessons learned were:
l.A measured heat transfer coefficient is
useful in predicting the heat flux into a
single tube.
2. Bed temperature and exit gas tempera-
ture cannot be predicated on the basis of
wetted surface area and a heat transfer
coefficient in a water-cooled enclosure.
An empirical correlation has to be
derived as a prediction tool.
Open Modular Concept
Because so little "wetted" surface was
needed to control bed temperature if
"viewed" surface were available, the design
concept of Figure 1 was seen to be unwork-
able, and the problems inherent in that design
unnecessary. The simplest way to avoid coal
distribution and maintenance problems was to
go to the open modular concept of Figure 3.
By selecting the distance between tube
walls, the required heat exchange surface
(wetted and viewed) could be provided, still
retaining a relatively large maintainable open
combustion space. If fuel and air to each
module were controllable, it would be possi-
ble to achieve a larger turndown.
The first checkout of this system was per-
formed on the boiler of Figure 2 by halving
the width (to 18 in.) and removing the tubes
on one side. Pending construction of an actual
single-module unit, the modified boiler of
Figure 2 was operated to develop coal feeders
and light-off procedures.
Coal Feed and Flue Gas Travel
The system which provided the most even
coal distribution was an up-through-the-grid
method. However, with single-screened coal,
any moisture resulted in plugging of the feed
tube. Also, erosion at the turns was excessive.
-------
Therefore, a straight-in coal feed system was
optimized and adopted for use on the single-
module boiler (FBM). Excessive carryover of
bed material was noted in all FBX test runs.
This was attributed to the gases taking a
horizontal path above the bed, entraining the
material without the allowance for particle
recovery provided by a straight vertical
express gas path. This assumption was verified
in later FBM tests: a 6-percent ash coal held
the bed level; i.e., made up for attrition elutri-
ation losses. The lessons learned included:
1. Do not specially prepare the coal used in
the development work; use coal identical
to that to be used in practice.
2. Do not consistently feed as-received coal
pneumatically if an upward directed turn
is required, at least not where forced
draft air is the motivating force. (Com-
pressed air has not been used because of
economics.)
3. Avoid horizontal movement of the flue
gas until it has passed through a screen
such as a convection bank, superheater,
or tubular air heater.
4. Bed particles will carry over if gases take
a horizontal turn even at heights well
above the theoretical disengaging height.
5. A vertical express route for flue gas pro-
vides for simple trouble-free design.
Initial FBM Test Results
The FBM design (see Figure 4), based on
the lessons learned in the FBX, utilized a
direct coal injection system and incorporated
an integral system for fly ash reinjection.
Both coal and fly ash were fed to the same
boiler bed.
Figure 5 graphically presents a heat balance
during typical operation. Note the high loss
from unburned carbon in the fly ash.. Utiliz-
ing, in a separate operation, a smaller
fluidized-bed column (the FBC), insulated so
as to reduce heat loss to the walls, collected
fly ash was refired, and the carbon loss was
reduced to an acceptable level. See Figure 6.
The result of this experiment was the
Carbon-Burnup Cell—a separate combustion
zone or segregated region for which the
primary fuel is fly ash.
During the initial test phase, the present
system of light-off was developed. By
"puddling" a flame into one section of the
bed surface, ignition is accomplished,
followed by propagation of the fire to all
portions of the module through openings in
the dividing tube wall.
It is realized that acceptable carbon burnup
has also been achieved in a single fly ash rein-
jected combustion space by utilizing extensive
uncooled refractory surfaces above the bed
and imbedded heat exchange surface area,
utilizing the same principle as the refractory
arch in a conventional anthracite-fired boiler.
In other words, if the elutriating unburned
carbon leaving the bed can be held long
enough in a hot non-radiant heat-removing
zone, the fly ash carbon will burn out. A
column employing 80 percent of its total
height and volume as a refractory-lined, non-
heat-transferring, elutriated-carbon burnup
section has demonstrated acceptable carbon
loss. The combustion of carbon in this exten-
sive freeboard area results in a temperature
rise in that area, further enhancing the carbon
combustion potential. However, such a
system can only be applied to boiler design
employing a large open "furnace" above the
bed, thus defeating the objective of compact
combustion and lowered cost.
In addition to the above size and cost con-
sideration, the integral (one combustion zone)
concept requires multiple recycle of fly ash
with a consequent increase in dust loading to
the collectors. For this reason, present
particulate emission standards are currently
forcing existing pulverized coal and spreader
stoker fired units to restrict recycle of fly ash.
The once-through separate carbon-burnup cell
precludes the "multiplication" of dust
loadings.
IV-4-3
-------
Early Modular Boiler Concepts
In addition to the initial concept of Figure
3, other preliminary designs based on initial
experimental work were prepared, and per-
formance predicted for efficiencies ranging
from 85 to 93 percent. These early concepts
have included:
l.A 250,000-pound-per hour packaged
boiler designed for the production of
250-psig saturated steam.
2. A 250,000 pound-per-hour packaged
boiler designed for the production of
600-psig, 750°F steam
3. A 2,000,000 pound-per-hour boiler
designed for the production of
2000-psig. 1000°F steam (and for reheat
to 1000°F).
4. An 8,000,000 pound-per-hour boiler
designed for the production of
2400-psig. 1000°F steam (and for reheat
to 1000°F).
Further Development in the FBM
One of the important design features that
required checking was the effect of bed
communication between adjacent modules,
particularly between a boiler cell operating at
a relatively low temperature and a carbon-
burnup cell (CBC) operating at a high tem-
perature.
The FBM was modified and a small insu-
lated region was added. (See Figure 7.) While
this small column is not a realistic CBC, which
would be much larger, it was the right size for
the quantity of fly ash generated by the FBM.
To ensure that good communication and
equal bed height were obtained, an open area
of 90 square inches was initially provided be-
tween the FBM and the CBC.
Initial operation of the FMB/CBC was
poor. With the FBM at 1600°F, the CBC
could not reach 2000°F, the temperature
IV-4-4
required for adequate burnup. The major
cause was the rapid interchange between beds
through the 90-square-inch opening. When
there is too much intercommunication be-
tween the two zones, the desired temperature
difference cannot be maintained. Reduction
of the opening of 2 square inches was found
to be adequate for light-off and bed-level
equalization. This small opening permitted
the desired temperature to be achieved.
A second energy loss resulted from the
carryover of hot bed material into the hori-
zontal duct. It was also impossible to deter-
mine burnup performance since the collected
ash was heavily contaminated with bed
material.
This carryover problem was solved by the
addition of a BCURA-conceived baffle screen.
However, the screen acts as a remarkably
effective heat sink. A high (approximately 40
Btu/ft2hr °F) heat transfer coefficient has
been computed for this device.
A vertical coal feeder has been tested. The
uniformity of coal distribution can be deter-
mined by measuring oxygen content at
various points above the bed. Oxygen in flue
gas may be held to within ±0.2 percent across
the width and length of the bed with coal
feeders as developed in this work. The tech-
nique lends itself to a simple coal delivery
system. Figure 8 shows the present FBM
boiler setup in the Alexandria, Virginia,
laboratory.
The lessons learned from this work include:
1. Very small (but properly located) open-
ings between regions permit eq{ial bed
levels and interbed light-off. ; '
2. Tubular baffle screens, properly posi-
tioned, serve as effective heat ex-
changers, almost as good as "wetted"
surface, and may also be used in lieu of a
high freeboard.
3. Downward-directed vertical pneumatic
coal feeders provide for an even fuel
distribution over a relatively large area
-------
and greatly simplify the problem of
boiler coal supply.
4. Cold air, supplied by a forced draft fan,
can provide all the motive force required
to inject coal into a fluidized bed.
Carbon loss from the FBM/CBC, when
employing a water-cooled baffle screen
directly above the bed, is as high as that
found in a typical spreader stoker operation
(see Figure 9). The present CBC is simply too
short* and the baffle screen is too effective as
a heat exchanger. By replacing the water-
circulated baffle with uncooled alloy steel
rods, satisfactory combustible loss (0.8 per-
cent, see Figure 10) has been attained. Laying
out and equipping the unit with total freedom
could have provided a design for a thermal
efficiency in the 90-percent range. Although
the present FBM/CBC will not achieve an
overall thermal efficiency much greater than
80 percent, much has been learned from its
operation; contemplated revisions do not
include a departure from the basic modular
concept.
PRESENT DESIGN CONCEPTS
Present design concepts are:
1. An open fluidized-bed combustion space
should be utilized so as to minimize coal
distribution problems and to provide
adequate access for boiler maintenance.
2. Coal injectors should be served with
motive air supplied directly from a
forced-draft far.
3. Heat transfer surface should be propor-
tioned betveen direct-contact, viewed
surface, and baffle surface so as to trans-
fer directly from the bed 50-60 percent
of the energy released in burning the
fuel.
* The CBC, added 2 years after the FBM was installed, had to
be located beneath the steam drum. The distance between
grid and horizontal discharge is only 40 in.
4. Flue gas should follow a long vertical
express route, including convection and
economizer banks.
5. The design of the boiler should permit
all boiler cells to communicate with the
CBC, especially if a regenerative SC>2-
capture method is to be used.
Based on these concepts, a 300,000-lb/hr
shippable unit has been conceived for an
eastern utility. (See Figures 11 and 12.) The
modules (cells) run parallel to the steam drum
and connect to a single CBC. At least until
additional experimental work can be done on
coal-feed techniques, a 4.5-ft spacing for coal
feed points seems advisable. The design shows
eight fuel-receiving points, each serving two
adjacent locations. Coal and limestone feed
(where used for SC>2 abatement) are com-
bined. Primary superheaters are shown in the
bed, but may also be arranged to serve as
baffle screens above the bed. When a module
is shut down, superheating stops, providing an
automatic attemperation mechanism. Some
external attemperation will still be required.
Fly ash from the boiler cells is collected in the
primary cyclone and fed to the CBC via
parallel circuits to four mushroom feeders.
There is also provision for auxiliary coal addi-
tion. No point in the CBC bed is farther than
2 feet from a feed point. Control of the CBC
is based on maintaining constant bed tempera-
ture. The CBC is cooled by the secondary
superheater, taking advantage of the higher
available temperature differentials in the
section. Preliminary predicted data for this
unit is outlined in Table 1.
Referring to Figure 11, it should be noted
that two of the vertical bed-immersed tubes
extend into the bed, completely surrounded
by the bed particles. As suggested by Seibel of
Erie City, a few rows of tubes (say 2 or 3)
could extend from the walls, forming a re-
stricted tube bank, enhancing heat transfer
but not interfering with the dispersion of
coal feed or the maintenance accessibility of
the boiler. This would permit widening the
cells (reducing the required number of cells
IV-4-5
-------
Table 1. PRELIMINARY DESIGN DATA FOR 300,000-POUND-PER-HOUR
FLUIDIZED-BED UTILITY BOILER (1500-PSIG DESIGN,
1270-PSIG OPERATING, 925°F)
Service
Final superheater in CBC
(In-bed)
(Above-bed)
Primary superheater
in boiler cells
Evaporation
(In-bed)
(Particle recovery
screen)
(Convection)
Economizer
Regenerative air
heater
Totals
Duty,
MBtu
11.2
5.4
64.0
98.4
74.9
25.1
48.0
35.4
362.4
AT diagram,
°F
2000*^2000
860 -»• 925
2000 -M 270
860*- 834
1600*+1600
575 -*• 834
1600*+1600
575*+575
1600**1600
575** 575
1 600 -M 270
575** 575
1270-> 709
525*- 385
709 -> 274
70-* 550
Tubing data
Surface
ft2
223
203
1,640
1,930
1,820
2,980
9,480
18,276
Length, ft
(O.D., in.)
430
(2)
390
(2)
3,130
(3)
2,460
(3)
1,460
(3)
1,280
(2)
5,700
(2)
18,100
(2)
29,030 (2)
3,920 (3)
Bed data
Grid area, ft2
Air rate, Ib/hr ft2:
superficial velocity, fps
Boiler section
364
CBC section
Fan data
Draft loss,
in. w.g.
Grid
Bed
16
794:11.7
Convection
& economizer Air heater
45
730:12.8
Total
31
Fan power @ 76,560 SCFM, 31 in. w.g. = 498 HP
Heat balance data, MBtu: Fuel input = 358.6; stack loss 16.6;
unburned combustible = 7.2; radiation and misc. = 7.4; boiler efficiency = 91.8%.
IV-4-6
-------
for a given capacity) and/or permit the reduc-
tion of bed level (and draft loss). Arrange-
ments incorporating this concept are under
study and will be incorporated in the design.
A preliminary cost estimate has been pre-
pared based on post-development conditions
but using 1970 costs. This estimate, sum-
marized in Table 2, shows a basic boiler cost
Table 2. PRELIMINARY COST ESTIMATE FOR 300,000-LB/HR
FLUIDIZED-BED UTILITY BOILER
1. Basic boiler
a) Steam drum 106,500 Ibs
b) Drum headers 6,400 "
c) Drum internals 4,100 "
d) Tubes 51,600 "
e) Lower headers 28,800 "
f) Casing 3,000 "
f) Contingency 59,600 "
Total weight 260,000 Ibs
Estimated cost @ $1.32/lb $343,200
Add 50% for special construction 171,600
Total basic boiler
2. Boiler appurtenances
a) 409 ft2 of air distributor $ 40,000
b) Economizer (30,000 Ib) 30,000
c) Air heater (120,000 Ib) 55,000
d) Boiler trim 25,000
e) Coal/ash/additive supply 200,000
f) Ducts/plenum/breeching 40,000
g) Insulation (30,000 Ib) 60,000
h) F.D. fan and drive 40,000
i) Collectors (2) 30,000
j) Electrostatic precipitator 290,000
k) Controls and instrumentation 120,000
Total boiler appurtenances
3. Normal design; installation and
contingency
a) Installation (includes miscel-
laneous piping and electrical 350,000
b) Contingency 200,000
c) Normal design 120,000
Total design, installation, and
contingency
Total normal installation at existing plant
Add 25% for first prototype unit
Estimated prototype facility cost
$514,800
930,000
670,000
2,114,800
528,700
$2,643,500
IV-4-7
-------
of $514,800 and a total plant cost of
$2,114,800. For the first installation, we have
applied the multiplier of 1.25 to the esti-
mated cost for post-development conditions.
The cost for the first unit is, therefore, esti-
mated at $2,643,500.
CONCLUSIONS
During the past several years, a number of
design arrangements have been proposed for
fluidized-bed boilers. Three true boilers have
actually been built: two by Pope, Evans and
Robbins; followed by BCURA's sheU-and-
tube unit. PER's "FBM" is the only unit
currently in service. In two cases it was found
necessary to remove some or all of the
imbedded tubes, since at the time these two
designs were completed, the removal of
energy from the bed by mechanisms other
than direct contact heat transfer was not
anticipated. All three of these actual boilers
have exhibited unacceptable carbon losses.
The use of a segregated fly ash carbon-
recovery section operating at high tempera-
ture (2000° F) has been demonstrated as a
means of recovering the unburned carbon
without resorting to low velocities and exces-
sively large bed areas (pressurized fluidized-
beds excepted).
The use of extensive closely spaced steam
generating tube banks immersed in the bed
(particularly horizontal banks requiring
forced circulation) are felt to be an unneces-
sary complication for most fluidized-bed
boilers. The arrangement of tubes in single-
row banks so as to form walls has been
demonstrated in the FBM to be a useful
design concept. The modular approach is the
only concept that has thus far been exten-
sively tested under actual operating con-
ditions.
In summary, it is felt that the design con-
cepts proposed here offer the following
advantages:
1. Maximum flexibility.
2. Simplest coal and sulfur acceptor feed
capability.
3. Maximum accessibility for mainte-
nance.
4. Good turndown capability.
5. No requirement for an induced draft
fan.
6. The use of natural, rather than forced,
circulation.
7. Compact construction by the use of
high-mass gas flows and heat release
rates.
8. Low bed particle carryover; ability to
operate with low-ash coal.
9. Segregated fly ash combustion region;
capable of operation at optimum
burnup conditions, without com-
pounding the particulate emission
problem. (This feature alsp obviates
the need for large furnace-like free-
boards for the consumption of un-
burned carbon.)
10. Capability of using a "built-in" lime-
stone regeneration process.
11. Low capital costs.
Regardless of the direction taken by
fluidized-bed boiler development, it is felt
that experience should be effectively dissemi-
nated and considered. There can be no good
reason for reintroducing design features
already shown to be deficient—unless a new
"twist" has been conceived to obviate the
previously demonstrated deficiency. Boiler
design features must be based .upon data
derived from large-scale experimental equip-
ment duplicating, or very closely simulating,
the proposed arrangement.
IV-4-8
-------
Figure 1. Plastic model, packaged fluidized-bed boiler.
-------
STEAM OUTLET
Dot
COMBUSTION ZONE 0
3ft. " 0
OVERALL LENGTH 7 ft.
BOILING WATER JACKET
FLUIDIZED BED
THERMOCOUPLE
AND
MANOMETER CONNECTIONS
GRID
COAL FEED SCREW
Figure 2. Imbedded-tube atmospheric-pressure boiler with 3 x 6 ft combustion zone.
IV-4-10
-------
TO DUST
COLLECTOR
FEEDWATER
INLET TO
ECONOMIZER
FEEDWATER
INLET TO
DRUM
FLUEGAS
COLLECTION
BREECHING
REINJECTED
FLY ASH AND
ADDITIVES
CONVECTION
BANK
HIGH. LOW
COAL LEVEL
SWITCHES
LIGHT-OFF
BURNER
PNEUMATIC
COAL FEEDER
TUBES
CROSS HEADER
CROSS HEADER
COAL FEEDER
TUBES
AIR SUPPLY
PLENUM
AIR DISTRIBUTION
GRID
OXIDIZING
FLUID BED
(SHOWN IN
ONE CELL ONLY!
Figure 3- Industrial modular boiler (original concept).
-------
Figure 4. Fluidized-bed boiler module (FBM).
BOILER ABSORPTION
4.450 KBtu (54.8%)
AIR PREHEATER CYCLE
540 KBtu (6.6%)
— FUEL INPUT —
8,130 KBtu (100%)
LOSSES
RADIATION AND
UNACCOUNTED
140 KBtu (1.7%)
CARBON LOSS
1.240 KBtu (15.2%)
/
FLUE GAS
2,300 KBtu (28.2%)
Figure 5. FBM heat balance without carbon
recovery (boiler convection and economizer
heat transfer deleted).
BOILER ABSORPTION
4,450 KBtu (59.7%)
AIR PREHEATER CYCLE
500 KBtu (6.7%)
LOSSES
RADIATION AND
UNACCOUNTED
480 KBtu (6.4%)
CARBON LOSS
70 KBtu (0.9%)
/
/
FLUE GAS
2.460 KBtu (33.0%)
FUEL INPUT
7,460 KBtu (100%)
Figure 6. FBM heat balance with carbon
recovery (boiler convection and economizer
heat transfer deleted).
IV-4-12
-------
COOLING WATER INLET
STOP-SLOP SCREEN
COOLING WATER OUTLET
TWO 1 x 2 m.
INTERCOMMUNICATION
SLOTS (NOT SHOWN)
ACCESS DOOR
MUSHROOM FEEDER
(FLY ASH)
AIR DISTRIBUTION
GRID
CBC PLENUM
0' T 2' 3' 4' 5'
GRAPHIC SCALE
Figure 7. Present boiler arrangement.
Figure 8- Test boiler, Pope, Evans and Robbins Alexandria, Virginia, laboratory.
IV-4-13
-------
BOILER ABSORPTION
6,186 KBtu (82.~8%)~
LOSSES
RADIATION
100 KBtu (1.2%)
COMBUSTIBLES
265 KBtu (3.4%)
ASH AND B.D.
12 KBtu (0.2%)
AIR PREHEATER CYCLE
512 KBtu (6.9%)
GAS
907 KBtu (12.2%)
FUEL INPUT
7,470 KBtu (100%)
Figure 9. FBM test B-5 heat balance (with
integral CBC water-cooled stop-slop screen).
BOILER ABSORPTION
4,926 KBtu (73%)
LOSSES
RADIATION
179 KBtu (2.7%)
ASH AND B.D.
60 KBtu (0.9%)
COMBUSTIBLES
56 KBtu (0.8%)
AIR PREHEATER CYCLE
300 KBtu (4.5%)
7_
•FLU EG AS
1,529 KBtu (22.6%)
FUEL INPUT
6,750 KBtu (100%)
Figure 10. FBM test B-14 heat balance (with
integral CBC, uncooled stop-slop screen).
COMBUSTION AIR
FROM F.D. FAN
EXIT GAS
—I (-*• TO PRECIPITATOR
f | OR SCRUBBER
MAIN DUST
COLLECTOR
TO CBC
REINJECTION
PREHEATED
AIR DUCT
RUNAROUND REDLER CONVEYOR
STEAM DRUM
SATURATED SfEAM
FEEDERS TO SUPERHEATER
PRIMARY SUPERHEATER
INLET HEADER
COAL/ADDITIVE
DROP LINES
CBC DUST
COLLECTOR
SUPERHEATER
OUTLET HEADER
PLENUM
CBC
ECONOMIZER
SURFACE
INLET FROM
PRIMARY SUPER-
HEATER OUTLET
SECONDARY
«--*"* SUPERHEATER
FINAL
SUPERHEATER
OUTLET
FLY ASH FROM MAIN
DUST COLLECTOR
Figure 11. Factory-assembled, coal-fired, fluidized-bed utility boiler (300,000 Ib/hr,
1270 psig, 925 °FTT) side view.
IV-4-14
-------
REDLER CONVEYOR
FORCED-DRAFT AIR HEADER
COAL FEEDER
SIGHT PORT
ECONOMIZER
COAL SUPPLY TUBE
CONVECTION BANK
SATURATED
STEAM FEEDER
TO SUPERHEATER
r PRIMARY
SUPERHEATER
INLET HEADER
PRIMARY
SUPERHEATER
PRIMARY
SUPERHEATER
OUTLET HEADER
LOWER HEADER
(ONE OF THREE)
Figure 12. Factory-assembled, coal-fired, fluidized-bed utility boiler (300,000 Ib/hr,
1270 psig, 925 °FTT) front view.
IV-4-15
-------
5. FLUIDIZED-BED BOILERS-
CONCEPTS AND
COMPARISONS
D. L. KEAIRNS AND D. H. ARCHER
Westinghouse Reserach Laboratories
INTRODUCTION
Fluidized-bed boilers offer economic and
social benefits of compact, cheap, and effi-
cient power generation systems; clean air; and
effective use of natural resources and raw
materials for fuel. Westinghouse, as part of a
contract with the National Air Pollution
Control Administration, is preparing and
evaluating boiler designs in order to provide a
technical and economic measure of the poten-
tial of fluidized-bed combustion boilers for air
pollution control. These concepts must then
be compared between themselves and with
conventional boiler or power systems.
Both new and existing concepts are being
considered. Fluidized-bed boiler concepts
have been generated by considering various
fluidized-bed configurations and operating
conditions. Operating conditions are selected
based on the evaluation of fluidized-bed com-
bustion data, fluidized-bed technology as
applied to other processes, and boiler and
power system requirements. The critical
fluidized-bed combustor parameters are cal-
culated from heat and material balances, heat
transfer relations, kinetics of combustion and
desulfurization, and fluid flow relations given
the configuration, operating conditions, pollu-
tion control limitations, and customer specifi-
cations. These parameters include cross-
sectional area of the bed, heat transfer surface
requirements, bed height, bed composition,
pressure drop through the system, solids and
gas distribution, and auxiliary equipment
specifications.
The projected investment and operating
cost and pollution control capabilities of the
fluidized-bed boiler power systems considered
can be compared with each other and with
other power systems. The best power system
is the one which meets customer specifica-
tions and pollution control limitations at the
minimum cost.
A pressurized boiler power system is used
to illustrate a fluidized-bed concept and
comparison with other power systems. The
high-pressure fluidized-bed boiler power
system utilizes the potential advantages of
fluidized-bed combustion over conventional
combustion systems and has several potential
advantages over an atmospheric fluidized-bed
boiler system:
1. Boiler size is reduced. The potential for
shop fabrication of utility boilers holds
promise for large cost savings—both
materials costs and field erection costs.
2. Power cycle efficiency can be increased.
3. Problems of solids handling and good
distribution of fuel and air are reduced
since the fluid bed area is small.
4. Scale-up problems are minimized.
5. Heat transfer surface is reduced by using
a combined gas steam-turbine cycle.
Preliminary design calculations for a high-
pressure fluidized-bed boiler power system are
presented to illustrate the development of the
concept and the identification of critical
fluidized-bed combustor parameters. The
design and evaluation of this concept are
IV-5-1
-------
proceeding based on these parameters. Tech-
nical and economic evaluations may result in
parameter changes in order to achieve an
optimum design. Power plant efficiency and
cost for the pressurized system have been
calculated based on projected boiler effi-
ciency and cost. The projected high-pressure
system performance is compared with a con-
ventional coal-fired power plant.
Fluidized-bed boiler power systems have
the potential for burning a broad range of low
grade fuels, including solid wastes.
This provides a means for the effective use
of natural resources and raw materials.
Thermal discharge to cooling water may also
be reduced if a pressurized combined gas
steam cycle proves economic and operable.
DESIGN BASIS
The development and evaluation of flu-
idized-bed boiler concepts must consider
pollution control projections; market projec-
tions; power cost analyses; fluidized-bed
combustor operating and design parameters;
and power system operation, maintenance,
and reliability.
Market Projections
The impact of fluidized-bed combustion
with sulfur absorption in the bed on air pollu-
tion abatement and on the economics of
steam and power generation has been
studied.1 Results of the market studies have
been used to establish fluidized-bed boiler
specifications. These results are summarized
elsewhere.1
Pollution Control Projections
Projected uncontrolled emissions from
power plants and projected air-quality goals
for SC>2, NOX, and particulates have been
considered.1 Power plant design requirements
for the year 2000, based on these projects,
might be:
Emission
SO2
NOX
Particulates
Control Level
Leaving Stack
lOOppm
100 ppm
O.OlOgr/scf
Control targets, for small .particles (< 5 micro-
meters) have not been projected. However,
regulations are anticipated which will require
significant reductions from current emission
levels. Information indicates that fluidized-
bed combustors, operating with a coarse coal
feed, may reduce small particle production by
'v 60 wt percent.
Power Cost Analyses
Investment and operating costs projected
for fluidized-bed boiler power systems must
be compared between themselves and with
conventional and proposed power systems.
The fluidized-bed boiler power system must
have an energy cost which is competitive with
those of nuclear plants and fossil fuel plants
with air pollution control. Present power
plant investment costs are $186/kW for a
600-MW coal plant without pollution control
and $269/kW for a 1100-MW nuclear plant.
Operating and Design Parameters
Operating and design variables for a
fluidized-bed boiler system must be selected.
Table 1 is a list of some of the important
variables which must be considered. A given
concept may not require consideration of all
the parameters listed. However, careful con-
sideration of these parameters, based on the
analysis of available data and technology, is
important if an optimum design is to be
achieved.
IV-5-2
-------
Table 1. OPERATING AND DESIGN PARAMETERS FOR DESIGN OF
FLUIDIZED-BED BOILERS
INDEPENDENT VARIABLES
Operating variables
Design variables
Fuel
Sulfur absorbent
NOX control
Bed temperature
Pressure
Gas velocity
Excess air
Particle size
Particle flow
Bed height
Reactor configuration
Arrangement of boiler functions (e.g.
economizer, superheater,
regenerator, solids removal,
carbon burnup cell)
Heat transfer surface (tube size,
orientation, arrangement)
Air distribution
Solids handling
Water steam flow
Sulfur absorbent regenerator
DEPENDENT VARIABLES
Operating effectiveness
Pollution control
Combustion
Sulfur sorbent utilization
Fluidization
Heat transfer
Elutriation
Attrition
Corrosion
Erosion
Agglomeration
Economics
Tubing and fabrication
Structures
Controls
Auxiliaries
Water circulation
Maintenance
Steam headers
Construction time
Solids preparation and distribution
Air preparation and distribution
Control
Turndown
System efficiency
Solids distribution
Air distribution
Particle diffusivities
Reliability
Pressure drop
Heat release
Operation, Maintenance, and Reliability
Any new boiler design must consider the
maintenance, operability, and reliability
requirements of the system. The importance
of these factors is clear in light of the present
day brownouts and blackouts. The ability to
achieve part-load operation with rapid re-
sponse time is an important consideration for
intermediate load generation which must vary
in power production throughout the day.
1V-5-3
-------
A Basis for Evaluation
These constraints and guidelines provide a
basis for the development, analysis, and evalu-
ation of fluidized-bed boiler concepts.
However, they must be viewed with flexi-
bility. An example is the projected pollution
control standards. Present data on the
removal of sulfur dioxide in a fluidized-bed
combustor desulfurizer is not adequate to
permit the design of units to remove greater
than approximately 90-95 percent of the
sulfur. Thus, for a coal containing 4.3 percent
S, it is not possible to design for control levels
below M80 ppm with reliability until further
data is available.
DESIGN CONCEPTS
Several boiler systems have been built,
operated, or proposed which incorporate
fluidized-bed combustion. Concepts go back
to 1928 when Stratton developed a spouting
fluidized-bed boiler.2 The early boiler con-
cepts incorporating fluidized-bed combustion
were generally developed for burning low-
grade fuels. These designs did not recover heat
or consider sulfur removal in the fluidized-bed
combustor. In the last decade, designs which
incorporate heat recovery and sulfur removal
in the fluidized-bed combustor have been
conceived in England and in the United
States.
In order to utilize the potential benefits of
fluidized-bed combustion in minimizing air
pollution while minimizing energy cost, new
and existing concepts have been considered.
Atmospheric and pressurized systems have
been considered which incorporate various
fluidized-bed configurations (single beds,
stacked beds, unmixed beds, and segmented
beds) and system component arrangements
(solids feeding and removal, final carbon
burnup, sorbent regeneration, heat transfer
surface, and boiler functions).
IV-5-4
One concept is the high-pressure fluidized-
bed boiler combined gas steam cycle system
which offers both large potential savings in
investment and operating costs and effective
pollution control. The concept illustrates the
development and analysis of a fluidized-bed
boiler system. Preliminary design calculations
have been completed for the pressurized
utility boiler system shown in Figure 1.
Power system performance studies and
boiler cost and performance analyses for the
system presented in Figure 1 show that the
system should operate at 10 arm, using sub-
critical steam conditions. Preliminary design
calculations have been made for a 600-MW
pressurized utility boiler system operating at
10 atm with once-through subcritical steam.
The operating conditions are summarized in
Table 2.
Figure 2 is a process flow diagram for the
pressurized utility boiler. The system consists
of a primary fluidized bed combustor desul-
furizer, cyclones for particulate removal, a
carbon burnup cell, CBC, to achieve high
combustion efficiency, and a limestone re-
generator. The sulfur removal system being
considered would feed regenerated stone from
the regenerator to the primary fluidized bed.
The regenerator is based on the system
developed by Esso.4 The high efficiency
primary multicyclone removes 92.5 percent
of the solids from the flue gas leaving the
boiler. The collected solids are fed to the
CBC, the combustion efficiency of which is
90 percent. The flue gas from the CBC is com-
bined with the effluent from the primary
cyclone and sent through the secondary
multicyclone, which removes 97 percent of
the solids, to the gas turbines. The combus-
tion efficiency is approximately 98.5 percent;
1 percent of the carbon fed is assumed to be
carried out in the flue gas and the CO concen-
tration in the flue gas is M).2 percent. Table 3
summarizes the material balance for the boiler
and limestone regenerator. Figures 3 and 4
summarize the energy balance for the boiler
and limestone regenerator.
-------
Table 2. PRESSURIZED FLUID-BED BOILER POWER SYSTEM
OPERATING CONDITIONS
Fluidized-bed combustor/desulfurizer
Pressure
Temperature
Fuel
H.H.V
Excess air
S02 removal
agent
10atm
1750°F
Pittsburgh
No. 8
seam coal
1 3,000 Btu/lb
10%
Limestone
Estimated limestone
feed to achieve
95% removal
Primary bed
material
Solids elutriation
Ash
CaO/CaSC>4
8 times
stoichiometric
CaO to react
with S02
S02 removal
agent
100%
M).5% of bed
weight/hr
Carbon burnup cell
Pressure
10atm
Limestone regenerator
Pressure 10atm
Temperature 2060° F
Steam conditions
Steam flow
Final feed-
water temp
Superheater
outlet press.
Superheater
outlet temp
Reheat steam
3,500,000 Ib/hr
578° F
2500 psi
1000°F
3,200,000 Ib/hr
Gas turbine system
Air temp
leaving
compressor 636 F
Flue gas temp
to gas
turbine(s) 1600°F
Flue gas temp
to upper stack
gas cooler 831"F
Temperature
Regeneration
of CaSC>4
Reheater
inlet temp
Reheater
outlet temp
Reheater
inlet press.
Reheater
outlet press.
Flue gas temp
to lower stack
gas cooler
Flue gas temp
leaving system
2000 F
~90%
650 F
1000°F
600 psi
580 psi
525° F
275° F
IV-5-5
-------
Table 3. MATERIAL. BALANCE FOR PRESSURIZED UTILITY BOILER
Stream
No.a
1-1
2-1
2-2
2-2A
2-3
2-3A
2-4
Description
Ohio seam,
Pittsburgh #8 coal
Air
Air to FBCD
Air to CBC
Flue gas from FBCD
Flue gas from CBC
Solids to CBC from
primary cyclone(s)
Flow rate
Solids
tons Air
207
29.4
19.9
27.2
Gas
moles/hr
,
160,000
151,400
8,600
158,900
8,600
Composition
Solids weight %
C 71.2
H 5.4
0 9.3
N 1.3
S 4.3
Ash 3.5
100.0
Ash 59.9
C 29.9
Spent
limestone 10.2
100.0
Ash 81.9
C 4.0
Spent
limestone 14.1
100.0
Same as 203
Gas mole %
N2 77.5
O2 20.4
H2O 2,1
100.0
Same as 2-1
Same as 2-1
Temperature
°F
636
636
636
M750
-V2000
M 700
-------
Table 3 (continued). MATERIAL BALANCE FOR PRESSURIZED UTILITY BOILER
Stream
No.a
2-5
2-6
2-7
2-8
3-1
3-2
3-3
3-4
Description
Flue gas from
primary cyclone(s)
Combined flue gas
to secondary
cyclone(s)
Flue gas to gas
turbine(s)
Waste solids
Regenerated lime-
stone to FBCD
Spent limestone
from FBCD
Regenerated stone
from regenerator
Waste stone
Flow rate
Solids
tons/hr
2.2
22.1
0.66
(0.15gr/SCF)
21.4
139.3
145.6
123.3
12.3
Gas
moles/hr
158,900
167,500
167,500
Composition
Sol ids weight %
Same as 2-3
Ash 79.7
C 6.6
Spent
limestone 13.7
100.0
Same as 2-6
Same as 2-6
CaO 77.4
CaC03 20.4
CaSO4 2.2
100.0
CaO 73.5
CaSO4 26.5
100.0
CaO 97.2
CaS04 2.8
100.0
Same as 3-3
Gas mole %
N2 74.2
C02b 15.0
CO 0.2
H2O 8.7
02 1.8
NO "v/400 ppm
S02 M70 ppm
100.0
Same as 2-6
Same as 2-6
Temperature
°F
M700
'vieoo
1600
1600
M750
^2000
-------
00
Table 3 (continued). MATERIAL BALANCE FOR PRESSURIZED UTILITY BOILER
Stream
No.a
3-5
4-1
5-1
5-2
Description
Makeup limestone
Coke
Air
S02 rich flue gas
Flow rate
Solids
tons/hr
28.4
9.2
2.2
Gas
moles/hr
6,070
6,800
Composition
Solids weight %
CaC03 97.0
C 98.0
Ash 2.0
100.0
CaO 72.7
CaS04 18.2
Ash 9.1
100.0
Gas mole %
Same as 2-1
N2 70.1
CO2 22.4
CO negligible
S02 7.5
100.0
Temperature
°F
^2000
aSee Figure 2.
''Includes CO2 produced from calcining makeup limestone.
-------
Figure 5, a temperature-enthalpy diagram
for the boiler, graphically represents bed
temperatures, water and steam temperatures,
estimated boiler tube-wall temperatures, and
the energy breakdown of the system. Table 4
summarizes the heat transfer surface require-
the preliminary design since reliable data is
not available to design for sulfur removal
greater than 90-95 percent. Techniques for
reducing nitrogen oxides have not been con-
sidered in the initial design. Design specifica-
tion for the particulars removal system after
Table 4. PRESSURIZED FLUID-BED BOILER OPERATING CONDITIONS
AND DESIGN PARAMETERS
Function
Heat transferred,
Btu/hr x 10*
Overall heat
transfer coeff ,
Btu/hr-ft2-°F
Surface
requirements, ft2
Inlet tube wall
temperature
estimates, °F
Outlet tube wall
temperature
estimates,0 F
Pre-evaporator
FBCD
6.3
47
12,000
660
730
CBC
1.2
47
1.900
750
750
Evaporator
8.8
47
17,200
730
730
Superheater
14.1
43
36,400
820
1110
Reheater
6.1
40
16,700
870
1150
Total
36.5
—
84,200
-
-
ments for the boiler and estimates of the
tube-wall temperatures. The maximum tube
wall temperature estimate is 1150°F. These
calculations are based on a bed-to-tube heat
transfer coefficient of 50 Btu/hr-ft2-°F,
which is considered pessimistic.
Table 5 summarizes the performance of the
pressurized-bed boiler power system. The net
power production is 635 MW, with the steam
cycle producing MJ5 percent of the power.
The boiler efficiency, 88.6 percent, is based
on conventional procedures for calculating
boiler performance. A term has been added to
account for the loss from sensible heat of the
spent limestone leaving the bed. Pollution
control requirements for the preliminary
design concentrated on sulfur removal. A
basis of 95 percent removal was selected for
the fluidized-bed boiler is based on particulate
loading requirements for the gas turbine.
Results from the BCURA high-pressure boiler
erosion tests have been used as the basis.4
Their results indicate that particulate loadings
of 0.15 gr/scf may be permitted in the gas to
the turbine. Additional tests and analyses are
required to determine the effect of coal ash
and spent limestone or other SC>2 sorbent on
turbine blade life. Final dust removal would
be added after the gas is cooled to achieve the
0.01 gr/scf stack emission level.
Details of the fluidized-bed combustor
desulfurizer design have been considered
based on the operating conditions and design
parameters. A shop-fabricated modular design
is being considered for the high-pressure
1V-5-9
-------
boiler in order to achieve reduced boiler
erection time and added savings in construc-
tion costs. Rail shipment requirements are
approximately 12 x 16 x 40 ft. With a vertical
unit, the active fluidized-bed cross-sectional
area can be 50-60 ft2. The number of
modules and the bed depths will be deter-
mined by the heat transfer surface require-
ments, heat transfer surface configurations,
operating and structural considerations, and
cost. Table 6 shows that the effect of tube
bundle design on the volume requirement is
Table 5. PRESSURIZED FLUID-BED BOILER
POWER SYSTEM PERFORMANCE
Power
Steam turbine
Gas turbine
Plant requirement
Net power
Efficiency
Plant heat rate
Plant efficiency
Boiler system losses3
Dry gas (based on stack
temperature)
Hydrogen and moisture
in fuel
Moisture in air
Unburned combustible
Radiation
Sensible heat of solids
Unaccounted for losses
Total losses
Boiler efficiency
Pollution control
S(>2 in stack gas
Paniculate loading
to gas turbine(s)
538.4 MW
113.1 MW
16.1 MW
635.4 MW
8975 Btu/kWhr
38%
3.88%
4.14%
0.08%
1.51%
0.15%
0.11%
1.50%
11.37%
88.6%
160ppmb
O.l5gr/scf
aBased on heats of reaction for coal-02 and
M 3300 Btu/lb coal.
^Emission is less than that from boiler due to air
added to gas turbine system for cooling.
important for large utility boilers. If a single
vertical deep-bed concept were adopted for
each module and if the tube bundle design
consisted of horizontal 1-in. tubes on a 2-in.
triangular pitch, the boiler would require five
shop-fabricated vessels with 20-30 ft deep
beds, each serving a separate function. If a
stacked bed design were adopted with each
module containing each of the respective
boiler functions, six vessels would be requir-
ed; the extra vessel would be required because
of the increased space requirements for gas
distributors and freeboard. The number of
vessels required for each concept is doubled
by going to 2-in. tubes with a 4-in. pitch. The
boiler would then require 10-12 vessels. With
this design, shop fabrication must be weighed
with field erection of a larger unit(s). Con-
struction costs, external piping, solids
handling, and control problems must be
studied in order to determine the best system.
The optimum heat transfer arrangement will
be selected based on steam-side pressure drop,
tube costs, header design, and fluidization, in
addition to the effect of boiler tube configu-
ration on bed volume.
Variation in pressure drop through the
boiler system has little effect on plant per-
formance for the pressurized power system. A
1-percent increase in the pressure drop
through the boiler system results in a
0.16-percent decrease in power and a
0.09-percent increase in heat rate. This
permits the use of deep beds, and high pres-
sure drop distribution plates and dust collec-
tion equipment without significant loss in
efficiency.
Under contract to Westinghouse, Foster
Wheeler is providing (for NAPCA) proposal
drawings of both a pressurized and. an atmos-
pheric utility boiler. Foster Wheeler will
describe their work in a companion paper.
DESIGN COMPARISON
Fluidized-bed boiler designs and power
systems must be compared between them-
IV-5-10
-------
selves and with conventional boiler and power
systems. The pressurized boiler power system
concept is used to illustrate the types of
comparisons which must be considered.
Comparison of Fluidized-Bed Boiler Designs
Fluidized-bed boiler concepts are compared
on the basis of pollution control effectiveness,
cost, efficiency, and operating characteristics.
The variable space which must be considered
in developing a fluidized-bed boiler design was
discussed and summarized in Table 1. To
achieve an optimum design, the alternatives
available for each variable will require analysis
and comparison.
Selection of a fluidized-bed configuration
for the pressurized boiler is an example of a
critical design task which is being evaluated.
Several configurations offer promise: vertical
units with single deep beds, vertical units with
stacked beds (each fulfilling a separate boiler
function), open horizontal beds, and seg-
mented horizontal beds. Each concept can be
considered for shop fabrication or field erec-
tion. A comparison of these configurations
must be made based on operating character-
istics (turndown, solids handling, air distribu-
tion, erosion, etc), heat transfer surface
design, sulfur removal and sorbent regenerator
design, freeboard requirements, fluidization,
combustion efficiency, and cost. For
example, single deep (on the order of 30 ft)
beds offer several potential advantages: solids
handling problems can be reduced (feeding
and removal), the number of air distributor
plates is reduced, available space is used
effectively, heat transfer surface out of the
bed is minimized, higher combustion effi-
ciency may be achieved as the result of longer
residence time*, and boiler size and cost can
be reduced. Problem areas can also be visu-
alized: coal distribution through the bed,
temperature profile through the bed, vibra-
tion, startup and turndown capability, and
tube maintenance. Stacked vertical beds offer
a method of providing the four boiler func-
tions in each module which may simplify
startup and turndown of the system. The
horizontal configuration simplifies structural
* Deep beds may be able to achieve the high combustion
efficiencies with coarse particles that have been projected
for fine particles in shallow beds.
Table 6. FLUIDIZED-BED VOLUME REQUIREMENT FOR
HEAT TRANSFER SURFACE
Tube
arrangement
Pre-evaporator
volume, ft3
Evaporator
volume, ft3
Superheater
volume, ft3
Reheater
volume, ft3
Total
volume, ft3
Triangular configuration
1-in.diam,
2-in. pitch
1,280
1,580
3,340
1,530
7,730
1-in. diam.
3-in. pitch
2,880
3,560
7,530
3,450
17,420
2-in. diam,
4-in. pitch
2,560
3,160
6,700
3,070
15,490
Rotated-square
configuration
2-in. diam.
4-in. pitch
2,950
3,650
7,720
3,540
17,860
IV-5-11
-------
design problems and offers alternate methods
for achieving turndown and recycling solids.
This configuration does not use the volume as
efficiently as does the vertical unit and may
also limit the flexibility in heat transfer
surface design. Fuel distribution may be more
difficult over the larger area in a horizontal
design. If lower gas velocities are used, carbon
carryover may be reduced; although deeper
beds with vertical units may provide the same
result. These comments illustrate the types of
comparative analyses which must be con-
sidered before selecting a configuration.
The achievement of high combustion effi-
ciency is another major concern. A vertical
deep-bed design may permit solids to be
recycled back to the bed as a means for
achieving high combustion efficiencies. With
shallow beds and gas velocities greater than
2-4 fps, the CBC concept conceived by Pope,
Evans and Robbins5 has marked advantages.
With a deep bed, the possibility exists for
achieving high efficiencies at high velocities
by recycling material. The tradeoffs which
must be considered are solids handling prob-
lems, operating cost, and control. It does not
appear practical to separate the CBC from the
pressurized boiler unit since this would in-
volve additional transport of solids in and out
of high-pressure vessels. Several CBC concepts
have been conceived to accomplish this goal.
Since the CBC represents a small fraction of
the total bed volume (see Table 4), it is not
practical to allow a single bed in each module
for the CBC. A single CBC could be designed
for one module and solids transferred from
the other modules; however, this increases the
solids handling problem and is not recom-
mended. Although a single high-pressure field-
erected unit might reduce the problems with a
single CBC, it is not recommended as a solu-
tion. Thus, design configurations have been
considered which incorporate the CBC in a
section of one of the beds in a stacked bed
design; e.g., in a quadrant of the bottom bed.
Solids are then transferred to the CBC from
the upper beds through an internal cyelone(s).
IV-5-12
The design of the sulfur removal system
illustrates another area where critical analysis
is required. Processes for high-temperature/
high-pressure regeneration of the sulfur
removal agent must be considered. Given a
regeneration process, the location must be
considered. If the regenerator is external to
the fluidized bed, solids must be transported
at high temperature and pressure to the re-
generator and the regenerated solids must be
returned to the boiler. Incorporating the
regenerator in the high-pressure fluidized-bed
boiler may greatly reduce the solids handling
problem, minimize heat losses, and reduce
space requirements. An additional process
scheme which may offer advantages is to
combine the carbon burnup cell (CBC) and
the regenerator into a single unit. Both
systems operate near 2000° F and the energy
requirements for the regenerator can be
supplied by the solids to the CBC. (See
Figures 3 and 4.) If the CBC is operated with
30 percent excess air, the excess air may
reduce the SC<2 concentration in the flue gas
from the regenerator-CBC to a level which
cannot be handled by a sulfur recovery plant.
The carbon monoxide requirements for
efficient operation of the regenerator must be
studied as well as the potential problems of
matching the operation of the CBC, regen-
erator, and the fluidized-bed combustor/
desulfurizer.
Solids feeding (tangential, bottom, side,
and top), solids removal, and arrangements of
boiler functions and heat transfer surface
configurations (vertical, horizontal, spiral, and
serpentine) are being studied and evaluated. A
range of each operating condition must be
considered and the tradeoffs equated: e.g.,
the effect of bed temperature on corrosion,
deposits, SO2 absorption, NOX emissions,
combustion efficiency, heat transfer surface
requirements, and ash fusion; and the effect
of particle size on kinetics, combustion effi-
ciency, heat transfer, elutriation, and prepara-
tion costs. Higher plant efficiencies may also
be possible since steam temperature may not
be limited by corrosion and deposits.
-------
These examples are presented to illustrate
the types of analyses and comparisons which
must be made between fluidized-bed boiler
designs. They are by no means complete.
Comparison with Conventional Systems
The high-pressure boiler requires the anal-
ysis of a total power system. Thus, compari-
son with conventional systems must ulti-
mately be based on overall power plant per-
formance. The high-pressure FBB power
system is compared with a conventional boiler
and power system on the basis of common
heat transfer surface, efficiency, cost, operat-
ing characteristics, and pollution control.
Heat Transfer Surface. The estimated heat
transfer surface for the pressurized boiler is
compared with the surface requirements for
the Hammond Unit No. 4 which has essen-
tially identical steam conditions. The surface
requirements for the pressurized boiler are
approximately 30 percent of those for a con-
ventional coal-fired boiler. Table 7 presents
the comparison.
Efficiency. The projected pressurized FBB
efficiency of 88.6 percent compares favorably
with the 89 percent of the Hammond Unit
No. 4. The pressurized FBB power efficiency
is 38 percent, 1 percent greater than that of
the Hammond unit. Higher efficiencies could
be achieved with the pressurized boiler system
by increasing the gas temperature to the gas
turbine. This could be achieved by modifying
the system to produce some fuel gas which
can be combusted at high temperature and
mixed with the low-temperature fluidized-bed
combustor gas. If the combustor were
operated at a higher temperature, the sulfur
removal process would not be effective.
Cost. The projected plant cost (on an
installed basis for plant operation in 1975) for
the pressurized power plant is $158/kW. This
is based on a boiler efficiency of 89 percent, a
fluidized-bed boiler/desulfurizer system cost
of $40/kW of steam turbine power, and a 15-
Table 7. COMPARISON OF HEAT TRANSFER SURFACE AREA
Function
Pre-evaporator
Evaporator
Superheater
Reheater
Heat transferred, Btu/hr x 1 08
FW Hammond
unit No. 4
(3,600,000
Ib steam/hr)
7.8b
9.1
14.5
6.3
37.7
Pressurized
fluid-bed
boiler
(3,500,000
Ib steam/hr)
7.5
8.8
14.1
6.1
36.5
Heat transfer Surface, ft2
FW Hammond unit No. 4
Reported
)
> 25,900
)
81,600°
1 13,000
220,500
Total surface
estimated3
(
81,000 {
\
118,000
109,000
308,000
Pressurized
fluid-bed
boiler
13,900
17,200
36,400
16,700
84,200
Surface
reduction
with pressurized
boiler
%
)
62
)
69
85
73
aConverts surface reported as projected surface to actual tube surface.
^Based on temperature leaving economizer of 593° F.
cSurface includes both projected and actual.
IV-S-13
-------
percent reduction in indirect costs. A boiler
system cost of $40/kW (steam) does not
appear unreasonable for a modularized design,
which also permits indirect cost savings on
construction and interest during construction.
A conventional 600-MW coal-fired power
plant cost in 1970 dollars is estimated to be
$186/kW, without pollution control equip-
ment for sulfur removal.
Turndown. The modular design of a pres-
surized boiler power system is projected to
have the capability of achieving turndown
ratios of 4:1 or greater. This compares favor-
ably with turndown ratios of conventional
coal-fired plants.
Pollution Control. The pressurized-boiler
power system offers advantages for control-
ling sulfur and particulate emissions. Al-
though control of nitrogen oxide emissions
has not been considered in the present design,
techniques are being studied to lower them.
The pressurized-boiler was designed for 95
percent sulfur removal. Data is limited which
permits design calculations to be made at the
95-percent level. Thus, refinement of the
designs to achieve high SC>2 removal may be
required. Particulate removal from the pres-
surized system has not been considered
beyond the gas turbines. Particle size distribu-
tions of emissions from fluidized-bed combus-
tors indicate that the emission of particles <5
micrometers may be significantly less than
those from conventional units. Conventional
coal-fired power plants using a similar coal
have SO2, particulate, and NOX emissions of
approximately 3200 ppm, 0.2 gr/SCF, and
500 ppm, respectively.
Table 8 summarizes a comparison of the
preliminary design of a pressurized-boiler
power system with a conventional coal-fired
power system. Based on the preliminary
design and analysis of a pressurized fluid-bed
boiler power plant, such a system has the
potential for providing not only compact,
cheap, and efficient power, but also clean air
and effective use of our natural resources.
IV-5-14
CONCLUSIONS
Fluidized-bed boiler concepts are being
generated and evaluated for cheap and effi-
cient power generation with air pollution
control. A broad range of fluidized-bed com-
bustor operating and design parameters are
being studied in order to achieve an optimum
design. The concepts are developed and com-
pared with other fluidized-bed boiler systems
and with conventional power systems, based
on: projections for emission regulations of
SC»2, NOX, and particulates; the power gen-
eration market; boiler and power plant costs;
and boiler and power system operation,
maintenance, and reliability.
A high-pressure fluidized-bed boiler power
system illustrates the development and evalua-
tion of a concept. Power system operating
conditions have been specified, preliminary
engineering calculations have been performed,
critical fluidized-bed combustor parameters
have been identified, and proposal designs and
evaluations are underway. Preliminary com-
parisons have been made with conventional
power systems. Data indicates that sulfur
emissions can be reduced 95 percent. The
fluidized-bed boiler also offers advantages in
controlling particulates and may reduce NOX
emissions. The amount of heat transfer sur-
face required by a pressurized boiler operating
in conjunction with a combined gas/steam
turbine power cycle is estimated to be •V'SO
percent of that required by a conventional
boiler for the same steam conditions and
flow. The preliminary cost estimate for the
fluidized-bed boiler power system is
$158/kW. The cost estimate for a conven-
tional coal-fired power plant, allowing
$10/kW for SO2 removal, is $106/kW. The
projected efficiency of the fluidized-bed
boiler is 88.6 percent, which compares favor-
ably with a conventional plant of the same
capacity. Thus, the high-pressure fluidized-
bed boiler power systems shows potential
technical and economic advantages over a
conventional coal-fired power plant.
-------
Table 8. COMPARISON OF PRESSURIZED-
BOILER POWER PLANT WITH
CONVENTIONAL POWER PLANT
Pressurized
Conventional fluid-bed boiler
Function coal-fired plant power plant
Boiler tube surface, ft2 308,000
Efficiency
Boiler, %
Plant, %
Cost, $/kW
Turndown
Pollution control
SO2, ppm
Particulate
Total, gr/SCF
NOX, ppm
89.0
37.0
186a
4:1
<\,3200
-V0.2
15
>500
84,200
88.6
38.0
158
4:1
160
<0.15b
<250
a600-MW plant without pollution control equipment.
bParticulate removal beyond the gas turbine require-
ments has not been considered in the initial design.
BIBLIOGRAPHY
1. Archer, D. H., D. L. Keairns, and W. C.
Yang. Marketable Designs for Fluidized
Combustion Boilers. Paper presented at the
Second International Conference on
Fluidized-bed Combustion, October 1970.
2. Stratton, J.F. O. Power. 68: 486, Septem-
ber 1928.
3. Skopp, A., J. T. Sears, and R. R. Bertrand.
Fluid Bed Studies of the Limestone Based
Flue Gas Desulfurization Process, Final
Report, GR-9-FGS-69, Prepared for
NAPCA. 1969.
4. Hoy, H. R. and J. E. Stantan. Amer. Chem.
Soc. Div. Fuel Chem. Prepr. 14, 2: 59, May
1970.
5. Bishop, J.W., E. B. Robison, S. Ehrlich, A.
K. Jain, and P. M. Chen. Papers presented
at the annual meeting of the ASME,
December 1968.
ACKNOWLEDGMENT
The work discussed in this paper was
carried out under the sponsorship of the
National Air Pollution Control Administra-
tion, Department of Health, Education, and
Welfare. Mr. P. P. Turner has monitored the
work for NAPCA.
The authors wish to thank: Mr. N. E.
Weeks for providing fluidized-bed boiler
power system performance and economic
analyses; Mr. J. R. Hamm for this analysis of
fluidized-bed boiler power system concepts;
Foster Wheeler Corporation (Mr. R. J.
Zoschak, Mr. R. W. Bryers, and Mr. J. D.
Shenker) for providing, under contract to
Westinghouse, information on the fluidized-
bed boiler design; and the NAPCA contractors
working on fluidized-bed combustion for
their cooperation and support.
IV-5-15
-------
DUMP
DUMP
AIR
POWEh
TURBINE
COMPRESSOR
TURBINE
I
TO
CONDENSER
BOILER -* j
1. DIAGRAM IS FOR 3500-
PSIA SUPERCRITICAL
BOILER. FOR 2400-
PSIA BOILER, #8
HEATER IS REMOVED.
2. FOR NON-INTERCOOLED
GAS TURBINE,
3. STATES (jj) to gg IN
STEAM CYCLE REPRESENT
LP END LOSS.
4. STATE @ IN STEAM
CYCLE REPRESENTS
FINAL FW TEMPERATURE
TO BOILER.
1
n
___%id
I -S- !
STACK
Figure 1. High-pressure fluid-bed boiler power system (block diagram, combined cycle).
IV-5-16
-------
NOTE: TABLE 3 DEFINES CIRCLED NUMBERS.
FLUE GAS
TO SULFUR RECOVERY PLANT
AIR.
SPENT LIMESTONE
LIMESTONE
REGENERATOR
FUEL
REGENERATED STONE
FLUE GAS TO GAS
TURBINE (S)
WASTE MAKEUP COAL
STONE LIMESTONE
AIR
Figure 2. Process flow diagram for pressurized utility boiler.
IV-5-17
-------
FLUE GAS TO GAS TURBINE (s)
SENSIBLE HEAT: 21.2 x 108 Btu/hr
COMBUSTIBLES: 0.41 x 108 Btu/hr
FLUE GAS FROM FBCD:
SENSIBLE HEAT 20.1 x 108 Btu/hr
COMBUSTIBLE GAS~0.4 x 108 Btu/hr
CARBON ~2.57 x 108 Btu/hr
LOSSES FROM FBCD & CBC
3.3 x 108 Btu/hr
SPENT CaO/LIMESTONE
WASTE SOLIDS
COMBUSTIBLE: 0.42 x 108 Btu/hr
COMBUSTIBLE SOLIDS
2.37 * 108 Btu/hr
FLUE GAS
SENSIBLE HEAT
1.1 x 108 Btu/hr
CARBON
0.23 x 108 Btu/hr
CBC
HEAT OF COMBUSTION
2.1 x 108 Btu/hr
FBCD
HEAT FROM COMBUSTION
51.2 x 108-Btu/hr
HEAT GENERATED FROM
CaO-S02 REACTION
1.4 x 108 Btu/hr
HEAT TRANSFERRED
TO STEAM
35.3 x 108 Btu/hr
REGENERATED CaO/LIMESTONE
COAL
53.8 x 108 Btu/hr
HEAT TRANSFERRED TO STEAM
1.2 x 108 Btu/hr
AIR
0.3 x 108 Btu/hr
AIR
AIR 6.3 x 108 Btu/hr
BASIS: 77?F
,6.6 x 108 Btu/hr
Figure 3. Energy balance for pressurized utility boiler.
IV-5-18
-------
LOSSES
0.26 x 108 Btu/hr
AIR.
FUEL
T
SO RICH FLUE GAS f
-------
SESSION V:
Conceptual Design and Economic Feasibility (Continued)
SESSION CHAIRMAN:
Mr. T. C. L. Nicole, National Coal Board, England
-------
1. DEVELOPMENT OF FLUIDISED-BED
COMBUSTION FOR FIRING
UTILITY STEAM BOILERS
D. H. BROADBENT
National Coal Board, England
The national Coal Board's interest in
fluidised-bed combustion stems from a desire
to maintain its coal sales to its largest outlet,
electricity generation, against growing compe-
tition from other sources of energy.
Figure 1 is a projection of the United
Kingdom energy market up to the end of the
century. It reflects the most optimistic fore-
casts available for the part to be played by
nuclear energy and natural gas, yet shows an
increasing use of fossil fuels from 1975 on.
Hence the Board's justification for research
and development of fluidised-bed combustion
as a possible means of reducing the capital
and running costs of fossil-fuel-fired boilers.
We have compared our energy forecasts
with those for the U.S.1 which show a similar
trend up to 2000 A.D.; but in the U.S., there
is the added incentive of pollution control
which makes fluidised-bed combustion even
more attractive.
Work at both National Coal Board labora-
tories (B.C.U.R.A. and C.R.E.) has proceeded
for 7 years and costs are currently running at
£1/2 ($1-1/4) million per annum. We have 94
scientists working on the project, backed by
an engineering workshop and other support-
ing staff.
There are 11 major research tools: 7 hot
and 4 cold; 10 at atmospheric pressure and 1
pressurised to 6 atmospheres. Results from
these rigs have been used to constnict a
mathematical model from which the param-
eters of a variety of boiler designs have been
established.
In addition, some £200,000 has been spent
over the last 2 years in obtaining designs of
20-, 120-, and 660-MW boilers from manu-
facturers of international standing. In all
cases, at least two designs have been obtained
for each boiler size from different contractors
to ensure unbiased designs. N.C.B. owns these
contractors' designs exclusively.
The programme of development shown in
Figure 2 covers both pressurised and atmos-
pheric units and is planned at a rate which
will enable building of a 660-MW unit in as
short a time as possible without risking exces-
sive scaleup factors. The programme covers a
19-year period, 7 of which have gone.
The next major capital investment is on the
Grimethorpe boiler, planned for one of the
Board's power stations in Yorkshire. It is 20
MW with a 12- x 36-ft bed and will take 12-18
months to build and commission. The boiler,
designed in detail, was sent out to tender to
six firms: five offers were received, each gave
a design to specification and also their own
version. Results of bids were remarkably
close, and the Board is now in position to
place a firm contract. Eighteen months of
V-l-1
-------
experimentation is planned for the Grime-
thorpe boiler, by which time we would have a
120-MW design out for tenders. A 120-MW
boiler takes some 3 years to build and com-
mission and would then require 2 years of
testing. During this 2-year period, a 660-MW
design would be put out for tenders, and
would take 5 years for construction and com-
missioning.
Total cost of this programme is some £45
million (at 1969 prices) over the 19-year
period. Our market surveys indicate that
there would be considerable offsetting of
these costs before the end of the 19-year
period by virtue of specific experimental
work we will carry out for other people,
commercial exploitation of boilers in the
20-MW range possible from 1973, and similar
exploitation of boilers in the 120-MW range
from 1978.
If pressurised fluidised-bed combustion is
proceeded with alone, it would follow a
similar pattern; but the time scale would also
be influenced by time requirements for suita-
ble industrial gas turbine developments.
We would aim to run the pressurised and
atmospheric programmes concurrently in
which case the information crossflow will
ensure some considerable cost sharing.
The question now is when does the N.C.B.
stop going it alone in the U.K. and start to
attract financial support from others. We have
decided to seek partners for the Grimethorpe
stage.
BIBLIOGRAPHY
1. Spaite, P.W. and R.P. Hangebrauck. Sulfur
Oxide Pollution: An Environmental Qual-
ity Problem Requiring Responsible
Resource Management. Paper to the 19th
Canadian Chemical Engineering Con-
ference. Edmonton, October 1969.
V-l-2
-------
700
600
U
§ 500
to
"5
8 400
in
c
o
c
o
E 300
t
cr
1
ft 200
100
1966
COAL
OIL
NUCLEAR
NATURAL GAS
TOTAL
MTCE
174.7
111.7
10.2
1.1
297.7
306
„ . ENERGY GAP._
TO BE FILLED BY FOSSIL FUEL
(BY DIFFERENCE!:
49
290
NUCLEAR ENERGY
1965
1970
1975
1980 1985
YEAR
1990
1995
2000
Figure 1. Projections showing the total fossil fuel "energy gap" up to the year 2000.
i
i
O
i
<
6.0
b.O
4.0
3.0
2.0
1.0
! "I I i ' '••
^^_
^—
1 1
- C02N0SMWUCT CONSTRUCT
PROTOTYPE 1 20"MW PROTOTYPE
BOILER COMMISSION BOUJH' COMMISSION
\ AND DEVELOP nc^?no
\ 20-MW / DEVELOP
_ LOAD FACTOR \ PROTOTYPE roOTOTYPE
*"" RASIC RESEARCH AND
1 1 1 1 1 1 1 1 1 '
', *
1 1 1 1
^^_
DESIGN AND CONSTRUCT
660-MW POWER STATION
LOAD FACTOR
40% 40%
X
ANALYTICAL SUPPORT FOR PROTOTYPES
1 ~\- \~ \
1 1
63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81
^^
1
82 8
CALENDAR YEAR ENDING
Figure 2. Cost and time scale of developing fluidised-bed combustion for central power
station boilers (present day prices).
V-l-3
-------
2. MARKETABLE DESIGNS
FOR
FLUIDIZED - COMBUSTION BOILERS
D. H. ARCHER, D. L. KEAIRNS, AND W. C. YANG
Westinghouse Research Laboratories
INTRODUCTION
Westinghouse, under contract to NAPCA, is
providing market studies, hardware designs,
performance evaluations, and economic data
for fluidized-bed boilers suited to industrial
and to utility applications. Recommendations
are also being furnished regarding the develop-
ment program and pilot plant operation
required to realize practical boiler systems
which are effective in air pollution control
and economic in operation.
MARKET STUDIES
The industrial and utility boiler market
surveys, which are nearing completion, have
two purposes:
l.To determine the possible impact of
fluidized-bed combustion with sulfur
absorption in the bed both on air pollu-
tion abatement and on the economics of
steam and power generation.
2. To establish functional specifications for
the fluidized-bed boiler designs that are
responsive to customer needs.
Industrial Boiler Installations
Table 1 shows historic data for the indus-
trial water-tube boiler market; statistics are
being projected so that boiler sales can be
predicted through 1980. Certain trends are
already clear:
V-2-1
1. The total capacity of industrial boilers
installed each year is increasing.
2. The number of new coal-fired installa-
tions are sharply decreasing and gas-fired
installations are increasing.
3. The average capacity of new boilers has
increased from about 60,000 Ib/hr to
75,000 Ib/hr in the past 7 years. (Coal-
fired boilers are 30-60 percent larger
than the average.)
The operating life of an industrial boiler is
about 30 years. A conservatively low estimate
of the total industrial boiler capacity now
extant is 1.2 x 109 Ib/hr; total annual fuel
costs for these might be about $2.5 billion.
The possible pollution control and eco-
nomic benefits of developing a successful
fluidized-bed combustion industrial boiler are
difficult to assess. Purchasers of industrial
boilers currently are avoiding 862 and partic-
ulate (but not NOX) emission problems by
using natural gas or low sulfur oil. But the
supplies of such fuels are limited and their
costs can be expected to rise. (These points
will be expanded later.) In certain areas of the
U.S. it has been announced that available
natural gas is inadequate for new industrial
customers. There may well be a growing need
for low-cost packaged industrial boiler in sizes
up to 250,000 Ib/hr (perhaps even larger) that
will burn coal in such a way as to minimize
SO2, NOX, and particulate emissions. Our
preliminary design work indicates that the
fluidized-bed boiler is indeed more compact
than the conventional coal-burning boiler and
-------
Table 1. ANNUAL SALES OF INDUSTRIAL WATER-TUBE BOILERS
IN THE U.S.
1963 1965 1968 1969
Total capacity sold, Ib/hr x 1CT6
Coal-fired
Oil-fired
Gas-fired
Other3
50.5
10.0
12.3
20.2
8.0
76.9
12.1
17.4
34.6
22.8
67.0
4.1
15.6
39.6
7.7
78.0
2.1
13.3
47.8
14.8
Total number sold
Coal
Oil
Gas
Other3
FOB costs of a packaged industrial boiler
Coal (up to 60,000 Ib/hr)
(above 60,000 Ib/hr
field erection)
Oil (up to 250,000 Ib/hr)
Gas (up to 250,000 Ib/hr)
879 1055
125 104
263 296
409 561
82 94
908
32
194
617
65
$1.25 per Ib/hr
$3 to $4 per Ib/hr
$0.80 to $1.00 per Ib/hr
$0.80 to $1.00 per Ib/hr
Includes bagasse, black liquor, bark, and waste.
thus can be packaged in sizes up to 250,000
Ib/hr. Emissions of SO2 can be reduced by a
solid absorbent—such as CaO or dolomite—in
the bed, or it may well be cheaper to use a
desulfurized char fuel. Use of such a fuel
would eliminate (or minimize) the use of a
solid absorbent and the problem of disposing
of it or regenerating it. Fluidized-bed combus-
tion may also be effective in reducing NOX
because of its lower operating temperature.
Particulates may also be more readily re-
moved from stack gases; they are, in general,
larger than particles from pulverized fuel
combustion; because they are not sintered,
wear of cyclones and mechanical collectors
should be minimized; and their high carbon
content may improve the operation of elec-
trostatic precipitators.
Fluidized-bed industrial boilers may well be
effective and economic in air pollution con-
trol if natural gas and low sulfur oil rise in
V-2-2
cost and if supplies of a desulfurized char
become readily available. The costs of operat-
ing an industrial boiler with a sulfur absorbent
in the bed, regenerating the absorbent, and
recovering (or disposing of) the sulfur will be
assessed to determine the economic feasibility
of using a high sulfur coal (or oil).
Utility Boiler Installations
Table 2 predicts the total generating capac-
ity and new installations of generating capac-
ity in the U.S. The new installations are
broken down into:
1. Base load plants—with a load factor
greater than 80 percent and employing
fossil fuel and a boiler.
2. Intermediate load plants—with a load
factor around 45 percent and employing
fossil fuel and a boiler.
-------
Table 2. ANNUAL SALES OF UTILITY STEAM
GENERATORS IN THE U. S. AND ELECTRIC
UTILITY POWER GENERATION
1970 1975 1985
Total installed capacity, gigawatts
New installations, gigawatts/yr
Base load—coal, oil, gas
Intermediate load—all fossil
Peaking load-all fossil
Electric utility power generation
Coal, %
Gas, %
Oil, %
Nuclear, %
340 530 1000
30
10.1
6.2
5.8
40
12.7
13.3
2.3
1960 1968
66.3
26.0
7.6
0.1
61.9
27.6
9.4
1.1
Installed costs of a coal-fired utility steam generator
Generator and support, per kW $35
Generator and support (including
draft feedwater, control, coal
and ash handling, dust collection,
piping, and other auxiliaries),
per kW $60
5.2
23.4
7.2
3. Peaking load plants-with a load factor
of less than 20 percent and operating
with gas turbines.
The most obvious market for fluidized-bed
combustion boilers are those now predicted
for the base and intermediate load fossil-fired
plants; this market is expected to increase by
70 percent-from 16 gigawatts in 1970 to 28
gigawatts in 1985. If the fluidized-bed boilers
are effective in air pollution control and if
their capital and operating costs are suffi-
ciently low, they may capture an additional
share of the generation market now ceded to
nuclear-fired base-load plants and to oil-fired
intermediate-load plants. Higher plant effi-
ciencies might be possible with increased
steam temperature and with pressurized
boilers with combined gas/steam turbine
cycle.
A gigawatt of power generating capacity is
roughly equivalent to 10 x 106 Ib/hr of steam
generating capacity and a coal consumption
rate of 400 tons/hr. The present utility boiler
market is therefore larger than the industrial
boiler market by a factor of 2 in terms of
capacity and by a factor of 7 in terms of
money. Utility boilers also present a more
serious problem in air pollution abatement.
Purchasers of the smaller industrial boilers are
turning to natural gas and low sulfur oil fuels.
Because of coal's low cost and availability
(despite recent problems of mining it in suffi-
cient quantities), coal remains a prime fuel for
the utilities. Most of the coal available to
power plants in the eastern United States has
quantities of sulfur beyond air pollution
limits. A 1000-megawatt (1-gigawatt) plant
burning 3.0-percent sulfur fuel products 24
tons/hr of SO2- A fluidized-bed utility boiler.
V-2-3
-------
which can economically absorb this pollutant
(permitting the recovery of sulfur) while mini-
mizing the production of NOX and partic-
ulates, can be of great benefit in reducing air
pollution since about half of the SC"2 emis-
sions and a quarter of the NOX emissions are
attributed to electric power generation by the
utilities.
Functional Boiler Specifications
In addition to examining the past and
future demand for industrial and utility
boilers, the market survey has also considered
customer requirements for the boilers they
purchase. Figures 1 through 5 show capacity
and steam conditions for future industrial and
utility boiler markets.
Figure 1 shows that the largest portion of
the industrial boiler market will be supplied
by units in the capacity range of
150,000-250,000 Ib/hr, a size that, not coin-
cidently, is the largest that can now be fac-
tory assembled and shipped as a package. If a
fluidized-bed boiler can be more compact
than a packaged gas- or oil-fired boiler of
conventional design, larger industrial boilers
may become more popular.
Figures 2 and 3 show that the steam con-
ditions of pressure and temperature in general
exhibit a broad range of preference: about
two-thirds of the boilers have steam pressures
below 600 psi and temperatures below 750°F.
Figure 4 shows that the largest share of the
fossil-fired utility boiler market will be sup-
plied by units in the 400-600 megawatt range;
a number of waste-heat recovery boilers in the
100-200 megawatt range may also be pur-
chased to operate in conjunction with gas
turbines in combined cycle power plants.
Figure 5 shows that super-critical boiler units
operating at around 3600 psi and sub-critical
units operating at around 2400 psi steam pres-
sure will continue to be popular with cus-
tomers.
V-2-4
Table 3 summarizes the information on
functional requirements used in establishing
specifications for the industrial and utility
fluidized-bed boiler designs. The final designa-
tion of specifications is discussed in papers
dealing with the detailed boiler designs.
Fossil Fuels for Boilers
Table 4 shows the proved recoverable fuel
reserves available for possible use in the U.S.
boilers. Actual economically recoverable
reserves of these fossil fuels may be as much
as 2-3 times the proved values; but still the
message appears clear. Coal is the most plenti-
ful source of fossil energy. Domestic oil and
gas reserves are in relatively short supply.
Their prices can be expected to increase
especially if foreign developments hinder the
free flow of oil to this country. Slightly over
half of the coal energy of the U.S. is in de-
posits east of the Mississippi River: most of
this coal has a sulfur content of 2.5-4.0 per-
cent by weight. Ultimately industrial and
utility boiler operators will probably be re-
quired to utilize coal or a coal-derived fuel.
Table 5 and Figure 6 summarize predicted
costs of fossil fuels. Major factors—including
production costs, alternate market opportu-
nities, and government regulation—influencing
the market price of each are different.
Political and environmental constraints, al-
ready shaping the price, supply, and demand
of each fuel, will continue to affect fuel price
trends over the next 15 years.
Table 6 gives transportation cpsts for fuels.
Costs indicate that coal and gas cannot be
used economically in large utility boilers far
from the source of supply.
Table 7, based on a careful consideration of
all the technical, economic, and political
factors involved in the choice of fuel supply,
indicates that utilities will use fossil fuels for
generating over two-thirds of the electric
-------
Table 3. FUNCTIONAL SPECIFICATIONS FOR FLUIDIZED-BED COMBUSTION
BOILERS DEVELOPED BY THE BOILER MARKET SURVEY
Characteristic
Steam conditions
Industrial boiler
Utility boiler
Capacity
Maximum
Most frequent
Minimum
350-500 x 103 Ib/hr
150-250 x 103 Ib/hr
25 x 103 Ib/hr
1300MW
500-700 MW
100 MW
Size, MW <300
Pressure, psig
Temperature, ° F
Performance
Efficiency, %
Special requirements
Turndown
Overpressure & flow %
Dynamics, %/minute
Startup, shutdown
Pressurized operation
TV viewing
Packaged construction
Multifuel capability
150-600
(sat-775)
86
1/3
-
—
-
No
No
Preferred
Not usually demanded
1800 2400
950-950 1000-1000
89-92
1/4 (1/2)
5 (on each)
5
Automatic
Yes
Yes
Unfeasible at present.
Desired
3600
1050-1050
but desired
power in 1980. A similar prediction for
industrial boilers is much more difficult at
this time. In any case, the development of an
adequate solution to the air pollution abate-
ment problem is a very important factor in
determining the choice of fuels and in the
economics of steam and power generation.
Table 4. PROVED RECOVERABLE RESERVES OF
FOSSIL FUELS IN 1968
Fuel
Coal, U.S., tons x
Gas, U.S., MCF x
109
109
Proven
reserves
265
287
Annual
consumption
0.5
19.9
Oil, U.S.,bblsx109
U.S. 31
Balance free world 381
Communist block 59
Alaskan field (est) 20-30
4.78
7.57
2.08
Alternate Means of SC>2 Pollution
Control in Boilers
A number of stack-gas processing systems
have been proposed for SO2 removal. These
systems have been considered for possible use
with conventional coal-fired boilers in electric
power plants. Some of these systems are
ready for commercial application or nearly so;
some require additional development and
testing. Table 8 estimates their effectiveness
and cost.
In general, the capital costs of the systems
are from a third to half the cost of the utility
steam generator itself. Monsanto's catalytic
oxidation costs equal the $35 per kilowatt
cost of the boiler; TVA's dry lime process is
about a quarter of this amount. While most of
the processes claim the 90-percent SC>2
removal required by today's air pollution
abatement goals, it is not clear how many of
V-2-5
-------
Table 5. PRICE TRENDS OF FOSSIL FUELS FOR POWER GENERATION
Fuel
Coal (mine mouth)
Steam coal
Gas
Residual oil (sea terminal)
Low-S (0.3%)
High-S (2-1/2%)
1968
Price,
d/106 Btu
17
25
32
32
Price Indices (1968=100)
1970
218
108
230
144
1975
205
287
267
141
1980
243
347
291
148
1985
291
386
295
156
Table 6. COMPARATIVE TRANSPORTATION
COSTS FOSSIL FUELS FOR ELECTRIC UTILITIES
Fuel
Cost3
Coal (unit train)
Under 300 miles
Over 600 miles
Gas (48-in. pipeline)
Residual oil
Large tankers
Large river barges
3.34
2.92
1.50
0.40
0.60
aCost in ^/million Btu/100 miles.
them can economically accomplish the 95-98
percent removal which may ultimately be
required as the use of coal in power genera-
tion increases. These stack-gas cleanup proc-
esses are unlikely to be economic for indus-
trial boilers. Scaled down to a size corre-
sponding with 250,000 Ib/hr steam boiler,
their cost would be in the range of
$1.50-$3.00 per Ib/hr, greater than the cost of
the boiler itself.
If fluidized-bed combustion boilers are
more effective in air pollution control and
lower in cost than conventional boilers plus
stack-gas cleanup equipment, they may
capture a sizeable portion of the predicted
market.
Pollution Control Targets
Projections of power generation by fossil
fuels have been used to estimate annual emis-
sions of SC>2,NOX and particulates to the year
2000. These estimates are based on the fol-
lowing assumptions:
1. Sulfur content of coal increases from the
current 2.7 percent to 3.5 percent in the
year 2000.
2. Coal and oil generation are combined.
3. Nitrogen oxides are produced at a rate of
800 Ib/billion Btu for coal and oil.
4. Heat rates are 10,300 Btu/kWh for coal
and oil and 12,500 Btu/lb for coal.
5. Flue-gas/coal ratio by weight is 12.5.
Tables 9, 10, and 11 show the degree of
pollutant reduction or removal required to
achieve acceptable air standards in 1970 and
to maintain the same total emissions (despite
increasing quantities of fossil fuels used in
power generation) in 1985 and 2000. Al-
though 90-98 percent reductions of SC«2,
NOX, and particulates may be beyond the
state of present art, such reductions are the
targets of fluidized-bed boilers. As such, they
provide a basis for design and a direction for
future development.
V-2-6
-------
Table 7. RELATIVE SHARE OF ELECTRIC UTILITY STEAM GENERATION
Fuel
Fossil
Coal
Gas
Oil
Nuclear
Total
1969
kWh x 109
1283.0
793.1
354.6
135.3
19
1302
%
98.5
60.9
27.2
10.4
1.5
100
1970
kWh x 109
1366.0
858.8
371.5
135.7
63
1429
%
95.6
60.1
26.0
9.5
4.4
100
1975
kWh x 109
1741.0
1101.7
461.7
177.6
461
2202
%
79.1
50.0
21.0
8.1
20.9
200
1980
kWh x 109
2028.0
1252.5
536.0
239.5
933
2961
%
68.5
42.3
18.1
8.1
31.5
100
1985
kWh x 109
2416
1396.1
625.4
394.5
1648
4064
%
59.5
34.4
15.4
9.7
40.5
100
-------
Table 8. ASSESSMENT OF FLUE GAS DESULFURIZATION PROCESSES
Process
Plant
size,
MW
Process cost
Capital
$/kW
Operating
rf/106 Btu
Years to
full-scale
plant
%SO2
removal
Chance
of
success
Ready for commercial application
C.E. dry lime-wet scrub 500 11 4.2 3 sold 85 Good
Monsanto cat-ox 500 36 0.2a ready 90 Good
Nearly ready for application
NAPCA-TVA dry lime 500 8 4.3 ready 40 Fair
Wellman-Lord 500 20b 3.9a-b 1 year 90 Fair
In development
S & W-lonics electrolysis
Esso-B&W dry adsorb
A.I. molten carbonate
Consol Coal potassium formate
1200
800
800
1300
19b
17"
15b
17b
0.6a-b
0 a-b
4 a,b
5 a.b
5 years
6 years
7 years
7 years
90
90
Fair
Unpredict
Unpredict
Unpredict
aAssumes byproduct credit but no credit for lower stack, eliminated precipitator, etc.
'•'Optimistically low cost, estimated by process developers.
Table 9. PROJECTED S02 EMISSION LEVELS
FROM POWER GENERATION PLANTS
Table 10. PROJECTED NITROGEN OXIDES EMIS-
SIONS FROM POWER GENERATION PLANTS
Year
1970
1985
2000
Control
level3
I
II
III
I
II
III
I
II
III
Average S02
leaving stack, ppm
2000
2000
220
2220
1230
140
2600
500
55
S02 reduc-
tion, %
0
0
89
0
45
94
0
81
98
aControl levels are: I—uncontrolled; II—equivalent to
maintaining total SO2 emissions from power plants
at the 1970 level; and Ill-equivalent to S02 emis-
sions if all the 1970 generating capacity was based
on using M>.3 percent sulfur fuel.
Year
1970
1985
2000
Control
level3
1
II
III
1
II
III
1
II
III
Average NOX
leaving stack, ppm
520
520
260
520
320
160
520
130
65
NOX reduc-
tion, %
0
0
50
0
39
69
0
75
88
aControl levels are: I-uncontrolled; II—equivalent to
maintaining total emissions to the atmosphere from
power plants at the 1970 level; and III—equivalent
to reducing the total NOX emissions to 50 percent of
the 1970 level.
V-2-8
-------
BOILER DESIGN
The results of the boiler market studies and
the air pollution abatement considerations
have been used in establishing functional
specifications for one industrial and two util-
ity boilers. A number of boiler concepts have
been reviewed and evaluated to determine
their compatibility with these specifications
and the general and economic requirements of
the market.
Functional Specifications for a
Fluidized-Bed Industrial Boiler
Results of these considerations, with regard
to the utility boiler, are presented in a com-
panion paper. This paper deals with the indus-
trial boiler. Table 12 gives its functional speci-
fications. A 250,000-lb/hr capacity was
chosen because of the market for this size and
Table 11. PROJECTED PARTICIPATE EMISSIONS
FROM POWER GENERATION PLANTS
Year
1970
1985
2000
Control
level3
I
II
III
I
II
III
I
II
III
Emission,
gr/SCF
3.7
0.5
0.037
3.7
0.30
0.023
3.7
0.12
0.009
Reduction, %
0
87
99
0
92
99.4
0
97
99.8
aControl levels are: (-uncontrolled; 11-equivalent to
maintaining total emissions rfom power plants at
1970 control level of 'V 87-percent removal;b and
Ill-equivalent to maintaining total emissions from
power plants at a level corresponding to 99-percent
removal in 1970.
bNAPCA. Control Techniques for Particulate Air
Pollutants. Publication No. AP-51. Washington,
D. C., January 1969.
because it may be the largest fluidized-bed
boiler shippable as a single module. A steam
pressure of 600 psi with superheat from the
saturation temperature of 489°F to 750°F
has been selected; these values have been
chosen because two-thirds of the industrial
boilers operate at or below these conditions.
The fuel for this boiler is a crushed Pittsburgh
coal, undried, with 4.3 percent sulfur. (Pulver-
ized coal has been ruled out because of the
expense and operating problems involved with
coal mills.)
Reduction of SC>2 emissions by 90 percent
is to be accomplished by lime (CaO, CaCC>3,
or dolomite) additions to the bed. Alternate
Table 12. FUNCTIONAL SPECIFICATIONS FOR A
FLUIDIZED-BED INDUSTRIAL BOILER
Characteristic
Specification
Steam capacity, Ib/hr 250,000
Steam conditions
Pressure, psig 600
Saturation temperature, °F 489
Superheated to, °F 750
Water to economizer at,u F 250
Water leaving economizer at, °F 350
Fuel
Coal Ohio Pittsburgh
No. 8 seam
Crushed to -1/4 in. xO
Air pollution control targets
S02 reduction by limestone
absorption, % 90
NOX
Partciulates reduction, % 95a
Boiler efficiency, % >85
Packaged construction
Factory fabrication Yes
Shippable Yes
Dimensional limits of
primary modules, ft 12 x 16 x 40
aAII ash and 3 percent of lime assumed to be elutri-
ated.
V-2-9
-------
absorbents are not considered because their
cost is in general higher and their effectiveness
is not yet demonstrated. A regeneration
system is specified for the lime absorbent
because large quantities of this material will
probably be required to achieve a 90-percent
reduction in SC>2 emissions. It does not seem
economically or technically feasible to dis-
pose of large amounts (up to half the coal
tonnage) of spent absorbent. Regeneration
with sulfur recovery does not appear easy or
cheap but it remains to be evaluated. A more
economic solution for industrial applications
may be to use a desulfurized coal char in a
fluidized-bed boiler. Such a boiler is probably
the only economic means for using char while
minimizing NOX and particulate emissions.
Reduction of the particulate ash and lime-
stone emissions in the stack gases is to be
achieved by a combination of cyclones and, if
required, an electrostatic precipitator.
Finally, an overall boiler efficiency greater
than 85 percent has been specified in order to
obtain reasonable fuel costs.
Fluidized-Bed Industrial Boiler,
Operating Conditions
Figure 7 shows that the fluidized-bed boiler
comprises three main sections—the combustor
and SC>2 absorber, the carbon burnup cell
(CBC), and the lime regenerator. The func-
tions of the last two sections might be com-
bined if the amount of carbon elutriated from
the combustor can be minimized. When large
amounts of carbon carryover must be con-
sumed, combining the CBC and the regen-
erator produces a gas mixture dilute in SO2-
This bed can be operated intermittently, vary-
ing temperature and flue gas composition, to
produce periodic surges of SC»2 from the
combined CBC/regenerator. But in either
case, capture or recovery of the sulfur is
complicated.
About 85 percent (11 ton/hr) of the total
coal required is fed into the primary fluidized-
V-2-10
bed combustor (FBC) with coal combustion
efficiency of 87 percent. The remaining 13
percent unburned coal is elutriated, along with
the ash and 3 percent of the spent limestone.
Ninety-five percent of the elutriated solids
from the FBC is recycled through the primary
cyclone back to the CBC for further combus-
tion. In addition to the recycled coal, fresh
coal (about 15 percent of total coal required)
is also fed to the CBC. Consequently, about
75 percent of the total coal is burned in the
FBC; the other 25 percent is burned in the
CBC. The ash and spent limestone recycled
back to the CBC are again elutriated along
with additional ash and spent limestone
produced in the CBC and eventually collected
in a secondary cyclone and electrostatic pre-
cipitator.
The spent limestone from the combustor is
regenerated in the limestone regenerator. The
total SC«2 removed in the regenerator is
assumed to be 90 percent of that picked up in
the combustor. The average composition of
the recycled limestone is assumed to have 5
mole percent sulfate. To maintain reactivity,
10 percent of the recycled limestone is re-
jected and replaced by fresh calcined lime.
Table 13 details operating conditions for
the boiler. Bed temperatures chosen are based
on the experience of all those in the NAPCA
program: they are high enough to obtain good
combustion and heat transfer, yet low enough
to obtain good performance from the lime
absorbent. High gas velocities (15 ft/sec) in
the beds have been specified to achieve high
boiler compaction. A large top size of the
crushed coal (1/4 in.) has also been specified
to minimize elutriation and to ease coal distri-
bution problems: fine particles( tend to
devolatilize and burn closer to the point of
their introduction. Coarser particles are ex-
pected to burn more uniformly throughout
the bed.
Minimizing elutriation has also been an
important consideration in choosing a particle
size distribution for the lime absorbent. A
-------
Table 13. OPERATING CONDITIONS FOR A FLUIDIZED-BED INDUSTRIAL BOILER
Condition
Temperature, "F
Coal combustion efficiency, %
Air excess, %
Superficial fluidizing velocity, ft/sec
Fluidized-bed
combustor
1650
87
10
15
Carbon
burnup cell
1900
90
30
15
Limestone
regenerator
2000 (bed)
-
-
_
Boiler
—
—
-
SC>2 removal, % _
Regenerating gas _
Particulate removal efficiency, % -
Stack gas temperature, °F -
Air preheat temperature, °F —
Boiler efficiency, % -
Heat losses
Dry flue gas, %
Evaporation of water formed by burning hydrogen, %
Evaporation of water in fuel, %
Heating water in air, %
CO and unburned C, %
Radiation, %
90
C02/C0 = 2
95
350
— None
86.2
5.65
4.37
0.30
0.14
3.00
0.36
Limestone feed
Stone—BCR 1359 as the primary bed material with characteristics
Particle size-1000-5000 microns with average diameter ^2500 microns
Feed requirements—^90% of 802removal w'tn 6 times stoichiometric feed ratio
SOo/CaO reaction heat generation-3 x 10* Btu/ton CaS04 produced and the stone enters
boiler at CaO at 1900° F after regeneration
Recycle rate-3% elutriated with the fly ash and 10% rejection in the regeneration
lime flow rate has been specified to yield a
molar Ca/S ratio of 6. Data from NAPCA
contractors indicates that this ratio should be
adequate to obtain a 90 percent reduction in
S02.
Excess air amounting to 10 percent is fed
to the combustor; and 30 percent to the CBC.
The regenerator is operated with an air defi-
ciency so that the molar CO2/CO hi the SC>2
containing flue gases is 2.
The temperature of the stack gases leaving
the boiler has been specified as 350°F; water
temperatures entering and leaving the econ-
omizer, as 250 and 350°F. These temperature
designations are based on conventional boiler
practice.
V-2-11
-------
Preliminary Design Computations
and Considerations
Material and heat balances, together with
heat transfer computations, have been carried
out for the fluidized-bed industrial boiler;
some of the results of this work are presented
in Table 14 and Figures 7 through 1 1. Figure
7 shows the large quantities of lime absorbent
and particulate solids which must be handled
in the system; it seems important to closely
integrate both the lime regenerator and CBC
with the fluidized-bed combustor and sulfur
absorber. This integration can minimize the
distances over which solids must be trans-
ported.
Figures 8 through 11 are heat balances, in
the form of temperature-enthalpy diagrams.
for various fluidized-bed boiler functional
arrangements. These diagrams show tempera-
tures throughout the boiler and the corre-
sponding heat quantities transferred from the
burning solids or combustion gases to the
water and/or steam expressed in terms of Btu
per Ib of fuel burned. About 20 percent of
the heat released increases the water tempera-
ture from 250°F to the saturation value,
Table 14. HEAT TRANSFER COMPUTATIONS FOR FLUIDIZED-BED BOILER DESIGNS
Characteristic
Heat load, Btu/hr x 106
Submerged in bed
Above bed
Superheater
Convection pass
Total
Log mean temperature difference, °F
Submerged in bed
Above bed
Superheater
Convection pass
Heat transfer coefficients, Btu/ft2/hr/°F
Submerged in bed
Above bed
Superheater
Convention pass
Heat transfer surface, ft2
Submerged in bed
Above bed
Superheater
Convection pass
Total
Shop-assembled
conventional
boiler
78.0
26.1
143.4
247.5
1910
1410
796
29.2
30.4
20.9
1400
609
8630
10639
FBC boiler
Vertical tubes in bed
extending into
freeboard
136.8
63.4
43.6
20.8
264.6
1161
742
1280
291
50
30
30
20.4
2350
2840
1140
3510
9840
Horizontal tubes
in bed
136.8
10.8
43.6
73.3
264.5
1161
1075
1280
495
50
30
30
30
2350
335
1135
4950
8770
V-2-12
-------
489°F; 50 percent vaporizes the water: and
15 percent superheats the steam to 750°F.
Figure 8 corresponds to a fluidized-bed
boiler with vertical evaporator tubes extend-
ing through the bed and freeboard. (These
tubes might be distributed more or less uni-
formly or clustered in banks or platens.) The
tube surface above the bed cools the gases
from the 1650°F bed temperature to 925°F.
A convection pass further cools the gases to
672°F; and an economizer recovers the
balance of the heat from the stack gases. The
superheater is located in the CBC: sufficient
coal is fired with the char to provide the
necessary heat. Dotted lines show a possible
alternate of bringing water to saturation in
the bed and using the convection pass as an
evaporator.
Figure 9 corresponds to a boiler with hori-
zontal tubes in the fluidized beds. In this
particular situation, the CBC does not super-
heat all of the steam; a portion (24 percent)
of the heat released by combustion gases as
they are cooled from 1650 to 1530°F is trans-
ferred to the steam in a convection super-
heater.
Figures 10 and 11 show two of many alter-
nate arrangements of different functional
units of an industrial boiler. In Figure 10 an
air preheater is used which, in this example,
transfers the same amount of energy that ,the
economizer does and preheats the air from 80
to 430° F For each Btu of heat transferred to
the combustion air in a preheater, an addi-
tional Btu must be transferred to the steam in
the fluidized-bed combustor if the bed tem-
perature is to remain constant. The econ-
omizer is put into the fluidized bed, and the
convection pass surface is packed above the
fluidized bed to take advantage of higher heat
transfer coefficients. The economizer is elimi-
nated in Figure 11 and the heat transfer sur-
face above the bed preheats the water. The
superheater is put in the CBC with a small
portion in the fluidized bed of the main
carbon combustor; in another alternative, the
superheater is in the fluidized-bed combustor
and the CBC serves as an evaporator. Numer-
ous other arrangements can also be proposed:
the optimum design will depend not only on
the economic consideration but also on steam
quality, velocity in the tubes, and the turn-
down characteristics of the boiler plant.
Consideration of all the possible alternatives is
a way to an optimum design.
Preliminary heat transfer computations
have been carried out on the vertical and hori-
zontal tube designs whose heat balances are
represented in Figures 8 and 9. Table 14
compares the results of these computations
with similar results for a conventional oil-fired
packaged boiler. The fluidized-bed boilers
transfer 7 percent more heat and have 10 per-
cent less transfer surface. Heat transfer coeffi-
cients above the bed and in the superheater
have been chosen conservatively low (30
Btu/ft2/hr/°F): they may well be40Btu/ft2/
hr/°F or higher. In this instance, a fluidized
coal-fired boiler will have 30 percent less heat
transfer surface than a comparable convection
gas- or oil-fired boiler.
For NAPCA, and under subcontract to
Westinghouse, Erie City is providing proposal
designs of an industrial boiler whose func-
tional specifications, operating conditions.
and material and heat balances have been dis-
cussed. Erie City will describe its work in a
companion paper.
PROBLEMS IN BOILER DESIGN
AND DEVELOPMENT
Proposal designs of one industrial and two
fluidized-bed boilers are now being prepared.
The primary purpose of these designs is to
evaluate the effectiveness and economics of
such boilers in generating steam and power
without air pollution. Their secondary pur-
pose is to reveal areas in which additional
knowledge or more development is needed to
produce a commercial fluidized-bed boiler.
V-2-13
-------
Although designs are still far from com-
plete, some problem areas have already been
encountered. A companion paper discusses
those involved with a pressurized utility
boiler; this paper is concerned with the prob-
lems involved with the design of an atmos-
pheric industrial boiler. These problems fall
into two areas—those involved primarily with
operation and those with design.
Operational Problems
The most important choice in determining
the mode of operation of a fluidized-bed
boiler is that of gas velocity (and of the corre-
sponding bed particle size). A high gas veloc-
ity, together with a coarse coal and limestone,
has been chosen for the industrial boiler to
minimize the cross-sectional bed area. Hope-
fully the boiler will then be most compact.
Although coarse coal may also minimize coal
distribution problems, coarse particles result
in lower heat transfer coefficients. High air
velocities result in: greater elutriation of
carbon from the bed; higher attrition in the
limestone absorbent; and, perhaps, erosion of
the heat transfer surfaces. The problem is to
determine an optimum gas velocity-particle
size choice; present performance and eco-
nomic data do not appear adequate to make
this choice.
A second operational problem is that of
boiler turndown; several procedures are possi-
ble:
1 "Slumping" the bed; i.e., merely decreas-
ing the air flow and depending on bed
cooling, bed contraction, or defluidiza-
tion to reduce heat transfer and thus
steam generation.
2. Decreasing air flow and simultaneously
draining a portion of the solids from the
bed to reduce heat transfer area.
3. Shutting off the air flow to whole sec-
tions of the boiler.
Selecting the procedure that will provide
the greatest flexibility in turndown, yet
V-2-14
minimize operating problems and boiler costs,
is a problem.
Design Problems
The most important design problem is
designating the heat transfer surface arrange-
ment. Vertical tubes permit natural circula-
tion, but horizontal tubes with forced circu-
lation may be more compact. The choice of
tube orientation affects the air distributor,
coal feeding, and tube manifolding systems.
Also the choice of tube diameter and pitch
has considerable effect on the design and
operation of the boiler. Small tube diameters
and close spacing produces a compact
boiler; however, they may also result in poor
fluidization and uneven fuel distribution.
Tube erosion, clinkering, and excessive carry-
over of carbon may also occur.
A second design problem is the choice and
location of the various boiler functions—air
preheater, economizer, evaporator, and super-
heater—in the boiler. It is not clear than an
economizer alone is the most effective and
economic heat trap; it may be better to com-
bine an economizer and an air preheater.
Other questions involve the location of
evaporators and superheaters in the beds and
in the convection passes: Which should be
placed where? And should the superheater be
distributed throughout the boiler to facilitate
turndown by sections?
A third design problem concerns using hori-
zontal baffle tubes at the surface of the
fluidized bed to minimize the splashing and
spouting of particles. Baffle tubes are effec-
tive in minimizing the carryover of particles.
And heat transfer coefficients for these tubes
are apparently almost as high as for sub-
merged tubes. But the reduction of particles
above the bed may reduce heat transfer rates
in the freeboard above the bed surface. Too
severe cooling of the bed gases may inhibit
combustion processes in the freeboard and
result in increased carbon monoxide and
hydrocarbon losses. Data is lacking for the
-------
design of an optimum baffle tube and free-
board section of a fluidized-bed boiler.
Other Problems
Two other general problems encountered in
the design of a fluidized-bed industrial boiler
have already been mentioned.
1. Designation of an optimum sulfur clean-
up system. Lime has been designated as
the absorbent, but should it be a once-
through or a regenerative process? How
should byproduct sulfur be recovered?
Might it be more realistic in the long run
to use a desulfurized char fuel?
2. Designation of an optimum design and
operating conditions for NOX and partic-
ulate emissions control. Data on simul-
taneous control of SC>2, NOX, and
particulates is inadequate to choose an
optimum method. Might a two-stage
combustion process be used to minimize
NOX formation? Can an electrostatic
precipitator be used effectively to
remove fly ash from fluidized beds since
carbon may provide some electrical con-
ductivity of the carryover?
CONCLUSIONS
Market studies for industrial and utility
boilers, together with studies of fuel availa-
bility and of stack gas cleaning processes,
indicate that a coal-burning, sulfur-absorbing
fluidized-bed boiler can have a significant
effect on both air pollution abatement and on
the economic generation of steam and power.
Functional specifications, including target
values for SC>2, NOX, and particulate reduc-
tions, have been drawn up for an industrial
boiler. Operating conditions have been sug-
gested, preliminary engineering computations
have been performed, and proposal designs
and evaluations are underway. Some problem
areas require further data and development to
confirm the practicality and effectiveness of
fluidized-bed boilers.
ACKNOWLEDGEMENT
The work discussed in this paper was
carried out under the sponsorship of the
National Air Pollution Control Administra-
tion, Department of Health, Education, and
Welfare. Mr. P. P. Turner monitored the work
for NAPCA.
The authors wish to thank for their
cooperation and support: Messrs. H. L. Smith
and S. J. Jack for their work on the analysis
and projections of the utility boiler market;
Erie City Energy Division of Zurn Industries
(Messrs. R. V. Seibel and W. D. Schwinden)
for their work, under contract to Westing-
house, on industrial boiler market analyses
and industrial fluidized-bed boiler design; and
the NAPCA contractors working on fluidized-
bed combustion.
V-2-15
-------
O
Q.
O
cn
O
CD
D
Z
O
Z
O
(J
<
a:
35
30
25
15
10
PsH
—
__
1
L*c
o
o
o
u
r
"5
•4
O
1
1
&
1
1
1
i
•*•!
^
in o
cj> cn
o
o
o
8
o
o
o
CO
CN
sn
^=
|
1
1
PC
^
i
'»
,_^
o
o
o
r~
o
o
0
in
p
1
X
i
s
1
o
o
in
o
o
o
1
!
';.•}
1
3
1
1
•'•:
'S
|
^.
1
i
i
*
*
\\
i
K
1
—
..,
—
U
p
;'i
i
3
rb i
0 0
o o cr
o q LU
o o >
in m o
CN C*l
O O o
§ § °
"~ *~ LO
m LO co
45
40
Q
S 30
v
i—
U
| 25
(J
i
D
* 20
>- 15
U-
o
Z
o
u 10
<
cr
BOILER CAPACITY RANGE Ib steam hr
Figure 1. Projected distribution of
U.S. industrial boiler sales - steam
generating capacity.
—
1
1
!i
ai
•~
o
-
§
&
J>
^
«
1
|jl
!."*
'O
i
i
I
•*
cn
CN
8
1
cncn
n
i
jj
!••
1
3Pi
1
1
j
!
*
01
m
S
CN
S3
1
I
£
—
01
00
o
o
ID
1
1
!ft
^
A
•:!;
1
$
1
1
Sir
CJ>
^ r^
*" "7
s §
00 ^-
HO ~
40 -
^R
Q
_J
C/)
fi
u
£ 25-
U
i
D
I'"'
J
° 15 •
U_
o
0
1- ,., .
L) 10
cr
LL
5 •
_
ra
C
jj
S
-I
if
S
s
•I
'1
.V
£
fj
;.'
•i
•j
?
r.
C
fl
"£
&
if.
'/.
%
£
5
Q
LLI
cr
D
t—
1/5
^1
:
1
_^
in o
1^ 00
CT) 01
—
1
[j
•J
r
j!
^
>:
^
CN
O
O
CO
™ ;
»"
r
•!
L
1
jl
j
ii
CN
CO
in
CN
in
pi
1 Cl
| la
CN
in
CM
(D
(i
|i
C '•
• n_ ^
S J ri
1 i rfei i
|; 1-1 K g f,
? I M £ H ;
;i 1 f 'J i'; i3 3
cc
i^ oo 01 o
in tn uo m ui
CN f^ CN ^ CN
^ (^. OO 00 O^
BOILER OPERATING PRESSURE RANGE, psig
Figure 2. Projected distribution of
U.S. industrial boiler sales — oper-
ating pressure.
BOILER OPERATING TEMPERATURE RANGE. °F
Figure 3. Projected distribution of
U.S. industrial boiler sales — oper-
ating temperature.
-------
i, V
g 18
6".
"' 16
Q
O
en
S 14
2
E 12
D
D
2 10
O
CO
"- 8
U
<
Q-
u 6
w
DC
LU
- 4
O
CQ
< 1
|_ 2
O
0
?ta
mmm
^
i
i
&
§
jjt
1
i
1
1
1
1
— 1
r
.'•*.'
1
lV-""*
1
&?
1
1
P
1
1
1
1
S
—
S -—i
is
55 Sy ^
>*ni
?#
i
fK>
£?
VJ;
^Vi
**yj
1-
!
S
1
^ x:
f- CO "
r- <— !
5«?
1
i
1
1
i
S
1
1
•«.«
'•'<•
'.*".
1
1
•^
'**'.
il
!
—
i
! 1
""
— '
^ —
8 1
1 100 1200 1300 1400 1500
70
UNDER 100 200 300 400 500 600 700 800 900 1000
100 BOILER CAPACITY, megawatts
Figure 4. Projected U.S. utility boiler sales -- steam generating capacity.
60
u
0.
u
cc
50
40
<
D
20
u.
O
z
o
I—
u
<
cc
10
,'\ S V 1500 - 2199 psig ,
I \ > ^ ~~~
I » / *\ /'
! \ • v r-'
,' \l \ yo- 1499 psig
(0
O
110
100
90
80
70
60
50
40
30
20
10
1970 72 74 76 78 80 82 84
YEAR OF INSTALLATION
Figure 5. Projected distribution of U.S.
utility boiler installations-steam pres-
sure.
I I I I I I I I I I I I I I
No. 2 Diesel Oil
— 0.3°o S Residual Oil-
0.3% S Crude Oil-
— Natural Gas
iV|% S Residual Oil —
% S Crud* Oil
ligh S Crude Oil —
•ligh-S Residual Oil
^Mine-Mouth Coal —
I I I I I I I I I I I I I I
1970
1985
1975 1980
YEAR
Figure 6. Projected fossil fuel prices
for U.S. electric industry -- 1971-1985.
V-2-1
-------
PRIMARY
CYCLONE
FLUE GAS: 11824 Ib-mole/hr
PARTICULATE: 1.71 ton/hr
/ ASH: 62.0 w/o
< COAL: 22.8 w/o
I L.S.: 15.2 w/o
FLUE GAS:
PARTI CULATE:
e ASH:
< COAL:
I L.S.:
8810
Ib-mole
hr
2.44 ton/hr
33.4
58.5
8.1 w/o
SOLID:
SOLIDS:
(COAL:
LIMESTONE:
1.63 ton/hr
CaS04: 11.4
FLUIDIZED-
BED
COMBUSTION
WITH
SULFUR
REMOVAL
SPENT LIMESTONE:
6.32 ton/hr
CaS04: 33.7 w/o
I
AIR:
r
COAL: 11 ton/hr
LIMESTONE:
5.60 ton/hr
CaSO4: 11.4 w/o
w/o
58.5 w/o)
CARBON
BURNUP
CELL
COAL: 1.85^0.
AIR: 2900
Ib-mole
hr
TO SECONDARY CYCLONE
& ELECTROSTATIC
PRECIPITATOR
FLUE GAS: 3014
hr
PARTI CULATE: 1.59 ton/hr
/ ASH: 64.0 w/o
< COAL: 20.2 w/o
< L.S.: 15.8 w/o
SPENT LIMESTONE:
1.85 ton/hr
CaS04: 34.4 w/o
FLUE GAS: 361.4
(S02: 7.42mole
SOLIDS: 0.14 ton/hr
SPENT LIMESTONE:
8.17 ton/hr
CaS04: 33.8 w/o
LIMESTONE
REGENERATOR
7.23 ton/hr
COKE: 0.48 ton/hr
Ib-mole
AIR: 321
hr
6.96 ton/hr
CaS04: 13.2 w/o
CaS04: ll.4w/o^
0.97 ton/hr 0.70 ton/hr
CaS04: 13.2w/p
MAKEUP
LIMESTONE
DISCARDED
Figure 7. Overall material balance.
V-2-18
-------
3000
o
LU
CC
2000 —
1000
3560/
Btu/lb FUEL
SENSIBLE HEAT FROM LIMESTONE 0.65'
SENSIBLE HEAT LOSS FROM ASH 31.2
CaO/SO2 REACTION 246
I I I I I I I I I i356?/
Btu/lb FUEL
SENSIBLE HEAT FROM LIMESTONE 0.65
hSENSIBLE HEAT LOSS FROM ASH 31.2
CaO/SO, REACTION 246
SUPERHEATER
(SH)
SURFACE ABOVE
BED
CONV.
PASS
ECON-
ONVECTION PASS
SUPERHEATER
SURFACE (SH)j
ABOVE BED
350 EVAPORATOR
I
EVAPORATOR
I I I
23456789 10 11
ENTHALPY 1000 Btu/lb FUEL BURNED
Figure 8. Temperature-enthalpy
diagram for vertical bed tube design.
1 23 4 56 7 8 9 10 11
ENTHALPY 1000 Btu/lb FUEL BURNED
Figure 9. Temperature-enthalpy
diagram for horizontal bed tube design.
3000
uJ 2000
cc.
LU
I
1000
1900
ABOVE BED
PRE-
H EATER
672/
Vfe
FBC
i
7501
.ECONOMIZER
SH 489| 489
. -ECONOMIZER
f°f,5° EVAPORATOR
~6~ 1 2 3 4 5 6 7 8 9 10 11
ENTHALPY 1000 Btu/lb FUEL BURNED
Figure 10. Arrangement of functional
units of a fluidized-bed boiler — al-
ternative No. 1.
3000 —
200( —
iS
i
LLI
Q.
I
1000
12
SUPERHEATER
(EVAPORATOR)
123456 789 10 11
ENTHALPY 1000 Btu/lb FUEL BURNED
Figure 11. Arrangement of functional
units of a fluidized-bed boiler — al-
ternative No. 2.
V-2-19
-------
3. DESIGN OF ATMOSPHERIC
FLUIDIZED-BED COMBUSTION
STEAM GENERATOR
R. V. SEIBEL
Erie City Energy Division, Zurn Industries, Inc.
INTRODUCTION
If steam were an end product, studies of
capital expenditures, operating costs, and
maintenance costs for various steam-generator
configurations would have precise meanings
to all. Since steam is not an end product and a
steam generator is never the only process in a
plant, it follows that the worth of a steam
generator to any particular user is a complex
subject. In one case, it does not have the same
meaning to all users.
Basically, a steam generator converts chem-
ical energy (available from oxidizable organic
materials) into heat energy (in a form that is
transmittable to a remote process). Heat, in a
convenient form, is the product of a steam
generator. Producing heat for a process by
converting chemical energy is the work of a
steam generator. How much of the available
chemical energy is converted? What procedure
is best for making this conversion? How much
does it cost to make this conversion? How
much reliability is demanded in making the
conversion? All these questions encompass
the study of steam-generator design and appli-
cation.
CONVENTIONAL STEAM GENERATORS
The four questions above (with the second
and third combined for convenience) can be
used as the basic thoughts in the study of
fluidized-bed combustion techniques as
applied to industrial applications in other
terms in atmospheric design. This paper dis-
cusses basic facts and experience with conven-
tional steam generators, and then summarizes
initial work and thoughts on a fluidized-bed
combustion (FBC) steam generator.
Quantity of Chemical Energy Converted
For all fuels, a terminal temperature of the
combustion products (selected by experience)
is the determining factor of boiler efficiency,
defined as heat output divided by heat input.
For natural gas, oil, and coal, the final or
"stack" temperature for optimum boiler effi-
ciency ranges from 300 to 350°F. Sensible
and latent heat losses vary, depending on the
fuel; at the same input and under identical
conclusions: natural gas has the lowest effi-
ciency; oil, the next; and coal, the highest.
Table 1 summarizes heat losses for three
different fuels where the final combustion gas
temperature is 300° F, and the sensible heat
base is 80° F. Practical cases are shown where
conventional design allows for varying excess
combustion air and unburned combustible
losses. No unburned combustible is generally
found with the oxidization of natural gas and
residual fuel oil. Practical designs of combus-
tion methods of coal do not allow complete
V-3-1
-------
combustion of coal. The 3-percent combus-
tible loss for coal in Table 1 represents spread-
er stoker firing for a bituminous coal.
Table 1 shows that boiler efficiency is
greater for configurations of heat transfer
surface on fuel burning equipment that allows
lower final combustion product temperature,
lower combustible losses, or reduced excess
combustion air.
Process and Cost of Energy Conversion
Conventional steam generator designs
promote rapid combustion reactions. The
rapid combustion reactions produce very high
temperatures; in fact, flame temperatures
closely approach those produced at adabatic
conditions, which are a factor in combustion
chamber construction. It has become com-
mon practice to water-cool the combustion
chamber to control maintenance from the
high flame temperatures. Heat absorbed in the
combustion chamber water-cooling system is
recovered by generating steam.
In burning low ash fuels, the size of the
combustion chamber is dependent on the fuel
and the combustion equipment. A slow burn-
ing fuel or a low turbulence burner may dic-
tate a larger residence time at high tempera-
tures to complete combustion. Where ash in
the fuel is molten at higher temperatures, the
combustion furnace chamber water-cooling is
used to lower (through radiant heat transfer)
the combustion product gas temperature
below the ash fusion temperature. In conven-
tional design, flue gas temperatures leaving
the combustion chamber range from 2000 to
2500°F.
Convection heat transfer is the principal
way to reduce the furnace combustion
products exit temperature to the final or
stack temperature. Although convective heat
Table 1. COMPARISON OF HEAT LOSSES AND BOILER EFFICIENCY FOR
THREE DIFFERENT FUELS3
Heat losses, % of heat input
Natural Residual Bituminous
gas oil (#6) coal
Sensible heat in dry flue gas
Heat loss due to water in (and formed from)
4.2
4.5
aAssumptions: combustion product final temperature—300°F;
natural gas excess combustion air-10%;
oil excess combustion air—15%;
coal excess combustion air—25%; and
type of coal firing-spreader stoker.
5.0
the fuel
Heat loss due to moisture in combustion air
Unburned combustible
Radiation and unaccounted 'for
Total heat losses
Heat output (boiler efficiency)
10.6
0.1
—
1.5
16.4
83.6
6.0
0.1
—
1.5
12.1
87.9
4.2
0.1
3.0
1.5
13.8
86.2
V-3-2
-------
transfer depends somewhat on the local con-
ditions of gas viscosity tube configuration and
some other minor factors, it depends mostly
on velocity. However, two factors limit ob-
taining very high coefficients.
One limit stems from conventional design
practice which has been changing over the
years in an attempt to optimize power and
heat absorption rates for gas and oil fuels:
power requirements for air-moving equipment
are a direct exponential function of velocity.
The other limit is the presence of particulate
material in the flue gas, with erosion of con-
struction materials a necessary consideration.
For examples of the above, velocities in the
convective zones for oil- and gas-fired steam
generators have increased to 120 140 ft per
second; for spreader stoker firing, velocities
are normally limited to 75 ft per second.
Convection heat transfer rates can be much
greater for a particulate-free gas stream, or
when the erosiveness of the particulates is
less.
The temperature of flue gas leaving the
steam generator (that is, upstream of the
separate heat trap) is largely determined by
practical design factors that are not apparent
to the casual observer; optimization of
supports, drum ligaments, fabrication capa-
bility, and assembly costs are just a few of the
many factors that must be considered by the
boiler designer. It is normal, however, to be
able to achieve gas temperatures of about
700°F in industrial boiler practice. Useful
heat remaining in the combustion product gas
stream can be roughly computed from the
above and corresponds to a 400° F tempera-
ture change. Often only one heat trap is
necessary to produce the desired thermal
efficiency of the>steam generator system.
An important philosophy, here, is that the
steam generator is a combination of combus-
tion chamber and convection zone. Utiliza-
tion of all the heating surface configuration
that is physically possibly adds little more
cost than only partial use. Heat traps, separate
items, are an additional expense. An increase
in heat duty assigned to a heat trap has a
direct effect on the cost so that, if boiler
heating surface is sacrificed to where heat
traps must be used, overall costs must be
increased. For example, if there is no convec-
tion surface in parallel with furnace tubes
from the steam drum there is only partial
utilization of the steam drum physical capa-
bilities.
In the discussion of heat traps, there are
two types of combustion air heaters and also
flue gas contacting feedwater heaters or
"economizers." There are regenerative and
tubular air heaters; economizers may be "bare
tube" or "extended surface." Where a com-
bustion air heater is used, the heat absorbed is
directed to the combustion chamber and
convective boiler surface; where an econ-
omizer is used, there is no increase in heat
input to the combustion chamber.
The coefficients that can be obtained and
the heat transfer temperature differences
favor the application of economizer, but a
limiting factor is the amount of heat that can
be absorbed in an economizer. Conventional
design avoids obtaining saturated tempera-
tures in the feedwater. Normal industrial
applications have feedwater inlet tempera-
tures of about 220°F, allowing a temperature
rise of over 100°F in the economizer which is
sufficient to achieve the desired boiler effi-
ciency.
Combustion air heaters find a limit on the
maximum possible air temperature as it
affects the mechanical design. It is common,
for example, to limit stoker combustion air
temperatures to 400°F. Another limit is the
"cold end" problem where metal parts in
contact with flue gas are at the dew point. It
is common to find a separate heater used to
limit the lowest metal temperature to some-
thing above the flue gas dew point.
V-3-3
-------
Operational Reliability
It is impossible for an unscheduled shut-
down not to affect the economy of plant
operation. Maintenance may be a trade off for
operating economy or for lower capital cost,
but designs that do not have predictable reli-
ability are unacceptable.
It is very difficult to describe normally
expected reliability. Reliability can vary from
indefinite periods of operation approaching a
full year (typical in petroleum refining) to
periods of 12 hours on weekdays only (as in a
carpet mill). Reliability is not necessarily
related to maintenance. For example, whereas
a petroleum refinery may be willing to com-
pletely rebuild a boiler once a year if the reli-
ability is re-established, a carpet mill would
expect virtually no maintenance costs for
several years.
If reliability were affected by a mainte-
nance problem (such as burner wear or clean-
liness) it is a normal standard to have on-line
replacement capability to restore perform-
ance. If the problem were continual, full-load
capability may be expected even with main-
tenance being performed. It is normal to
expect no maintenance problems with the
steam generator in regard to tube life or
material considerations. The typical industrial
boiler customer prefers to have no steam-side
maintenance, including circulation pumps or
other items that could affect steam-generator
reliability.
burning techniques. In addition, if there is
reduction of sulfur gases in the combustion
products, a lower final temperature may be
possible.
How the chemical energy is converted to
useful thermal energy is somewhat different
in an FBC boiler. Combustion takes place at a
controlled temperature (now thought to be
1650°F for optimum pollution control).
There is a great deal of discussion on the pre-
diction of submerged heating surface. Some
experimenters feel that the overall coefficient
of 50 Btu/ft/°F, commonly assigned to the
submerged heating surface, is too conser-
vative. Suffice it to say that the mechanics of
predicting bed temperature are not well de-
fined. Radiation heat transfer capability from
flue gases at the bed temperature as they
traverse the combustion chamber signify that
further temperature reduction of combustion
products must be through convective heat
transfer.
Using the previous outline, heating surface
in the bed and in the walls of the fluid-bed
chamber requires a system of steam-generat-
ing tube headers and steam drun^ comparing
to the water-cooled furnace in the convec-
tional unit. Further utilization of the steam
drum and separators, by adding parallel con-
vective heat transfer surface, should be possi-
ble and is a factor in the temperature reduc-
tion of the combustion gases. It seems reason-
able that an equivalent reduction to conven-
tional designs can be made; so, a single heat
trap would be sufficient.
FLUIDIZED-BED COMBUSTION STEAM
GENERATOR
The above analyses set the stage for a dis-
cussion of the application of fluidized-bed
combustion (FBC) to industrial boilers.
More chemical energy can be converted in
an FBC boiler than in conventional types. The
outlook is that excess air values and com-
bustible losses may be less than in other coal-
V-3-4
Erie City has begun to develop FBC boiler
designs to test certain features. A market
survey of industrial boilers has confirmed ,a
firm market position for steam generators
having 250,000 Ib/hr capacity, producing
750°F steam at 600 psig. To begin our anal-
ysis, we used these requirements as design
basis. Also, we established that the physical
configuration of the FBC steam generator
would lend itself to shop assembly and to
shipment by railroad.
-------
We tried two different fluidized-bed de-
signs. One utilized modular shop-assembled
boiler units, with a fluid-bed chamber shipped
separately from the boiler convection surface.
The two would be joined together in the field.
In all arrangements, we proposed that
combustion gases would flow in the freeboard
area to one end of the chamber. They would
be collected and passed through the convec-
tion zone. In one case, we considered sloped
or vertical surface penetrating the grid and
passing through the freeboard area. In another
case, we considered only heating surface in
the bed in the combustion chamber.
In still another trial, we set the goal as a
totally shop-assembled unit. Where there are
sloped tubes in the bed and the bed free-
board, we were able to duplicate conventional
boiler performance if there was sufficient
heating surface to have only one heat trap to
produce the desired boiler efficiency. Table 2
compares performance values of the two-piece
modular arrangement and of the single-piece
unit. Although the advisability of the flow to
one end may be questioned, it does have very
real advantages. Heat transfer surfaces may be
installed in the freeboard area and (most
important) a particulate removal system may
be installed to protect against excessive un-
burned carbon recirculation through both the
convection pass and the heat trap. In accumu-
lating all gases, convection pass heat transfer
coefficients may be improved using conven-
tional configurations of heating surface.
Figures 1 and 2 show two arrangements of
FBC boiler components demonstrating the
use of an intermediate particulate removing
system.
So far, we have not mentioned transfer of
heat to produce steam to elevate the useful
energy that is transmitted to the process. It is
fundamental that the superheater must be as
carefully designed in FBC boiler applications
as it is in conventional designs. Compared to
conventional design, the fluid-bed system is
largely an on-off device; consequently com-
pletely submerged superheater surface does
not seem appropriate when low loads are
considered. A combination of submerged in
the bed and conventional surface appears to
be better. Figure 3 shows one way to protect
against excessive superheater submergence in
the bed at low steaming rates: steam flow is
biased toward the submerged surface to pre-
vent overheating at low steaming capacity.
CONCLUSION
Reliability of the FBC industrial boiler
appears to be the heart of the design problem.
Although we have not designed the details of
a fuel injection system or grid, we have set
three goals:
1. The fuel injection system must have
on-line replacement capability.
2. The grid system must have positive cool-
ing to protect against temperature excur-
sions approaching bed temperature.
3. The grid must be replaceable without
major alterations to the steam generator.
We feel that the fuel feed systems and grid
design are the principal hurdles to industrial
steam generator customer acceptance of an
FBC boiler design. We feel that the other
objections can be overcome by a design that
sacrifices shop assembly.
At this point, we are not at all confident
that the FBC atmospheric steam generator
will win acceptance over the present alter-
natives without full utilization of the poten-
tial for pollution control. With pollution
control, however, lower cost and more readily
available fuel supplies can be used. It is in this
area, that air-fluidized-bed combustion tech-
niques will have worthwhile economic com-
pensations.
V-3-5
-------
Table 2. COMPARISON OF DESIGN VALUES FOR TWO FLUIDIZED-BED
COMBUSTION BOILER DESIGNS
Characteristic
Fuel
Steam flow, Ib/hr
Steam pressure, psig
Steam temperature, °F
Boiler efficiency, %
Heat input, M Btu/hr
Heat absorption M Btu/hr
Combustion chamber & superheater
Convection pass
Economizer
Total heat absorption
Gas temperatures, °F
Leaving combustion chamber
Leaving convection pass
Leaving economizer
Combustion air flow, Ib/hr
Excess air, % of theoretical
Flue gas flow, Ib/hr
Fluid bed press, drop, in. wg
Freeboard press, drop
Convection press, drop
Total press, drop, in wg.
Gas velocity, ft/sec
Through furnace
Entering convection pass
Superheater press, drop, psi
Heat transfer surface, ft2
Submerged in bed
Surface above bed
Convection pass
Superheater
Total heat transfer surface, ft2
Approximate height/length/
width, ft
Modular design
(no freeboard
heating surface)
Coal
250,000
600
750
87.7
337
195.6
68.5
24.9
289.0
1530
672
350
275,000
10.0
298,000
40.0
.0
6.2
46.2
45
100
50
3,200
1,'TOO
7,710
860
12,870
16/40/17
Single unit (with
sloped tubes in bed
and freeboard
heating surface)
Coal
250,000
600
750
87.7
337
244.0
20.1
24.9
289.0
925
672
350
275,000
10.0
298,000
40.0
4.6
7.1
51.7
67
90
50
3,200
4,970
3,510
970a
12,650
16/40/13
aDoes not include 620 ft* in the bed.
V-3-6
-------
FLUIDIZED
BED
CHAMBER
^-\ "
FAN
CAR-
BON
BURN-
UP
CELL
ECONOMIZER
.__!__.)
CONVECTION
ZONE
CARBON-RICH
~~"ASH "'""
I I
MECHANICAL
Li DUST T-*-
COLLECTOR
.W
ELECTRO
STATIC
PRECIPI-
TATOR
W
STACK >
ASH TO
DISPOSAL
Figure 1. FBC system components - arrangement No. 1.
XXX
FAN
I I
ELECTROSTATIC
PRECIPITATOR
'STAClO
1 J
ASH TO
DISPOSAL
Figure 2. FBC system components - arrangement No. 2.
SATURATED STEAM
FROM STEAM DRUM
SUPERHEATER
ELEMENTS ABOVE
FLUIDIZED BED _ -
•SUPERHEATER ELEMENTS
SUBMERGED IN
FLUIDIZED BED
FLOW-BIASING
VALVE
SUPERHEATED
• STEAM TO
PROCESS
SUPERHEATER OUTLET
HEADERS
Figure 3. Superheater arrangement.
V-3-7
-------
4. DESIGN FEATURES
OF PRESSURIZED
FLUIDIZED-BED BOILERS
R. W. BRYERS
Foster Wheeler Corporation
INTRODUCTION
As part of its effort on a NAPCA-funded
study* to evaluate the fluidized-bed combus-
tion process, Foster Wheeler is studying
conceptual designs of fluidized-bed boilers.
Although atmospheric-pressure fluidized-bed
boilers for a 600-MW plant are also to be
studied, work thus far has concentrated on
pressurized boilers for a 300-MW plant. These
pressurized 300-MW designs are the subject of
this presentation.
ADVANTAGES OF PRESSURIZATION
Application of fluidized-bed combustion to
a utility steam generator requires selection of
a cycle and conceptual design which will
make best use of the improved heat transfer
offered by fluidization. Considering the rela-
tively low average temperature at which
combustion takes place in the bed, the com-
bined steam-turbine/gas-turbine power plant
with a pressurized boiler (see Figure 1) must
be considered an immediate candidate. The
higher heat transfer coefficients and high gas
densities of pressurized operation should offer
an excellent opportunity to reduce the first
cost of the steam generator.
Pressurizing a conventionally fired boiler in
this manner reduces the amount of heating
surface required, which reduces its size and
' Westinghouse is prime contractor on this program.
weight. Ultimately these reductions are re-
flected in a reduction in first cost as comp-
ared to a conventional non-pressurized boiler
designed for the same steam conditions and
net cycle output. The reduction is achieved as
a result of an increase in emissivity of the
non-luminous gases, higher gas densities and
available pressure drop, higher permissible gas
mass flows and convection heat transfer
coefficients, and the reduced boiler duty
requirement which results from splitting
power generation between the gas turbine and
steam turbine. Fluidized-bed vcombustion
should offer further improvements in first
cost by making use of high heat transfer
coefficients throughout virtually the entire
exchange of energy from the combustion gas
to the steam, which is true when gas-turbine
inlet temperature closely approaches the
fluid-bed temperature.
CYCLE SELECTION
Cycle selection is ultimately dictated by
requirements for gas side pressure and steam
conditions.
Gas Side Pressure
Selection of the optimum gas side pressure
is, of course, determined by studying the
effect of pressure on overall plant economy.
An examination by Westinghouse of the influ-
ence of pressure on plant heat rate and output
V-4-1
-------
indicates that there is little advantage in
operating the gas turbine at boiler inlet
pressures above 10 atmospheres. For a given
boiler capacity, the cost of the pressure shell
is most strongly related to gas pressure. Sur-
face requirements are a function of steam
cycle conditions and, except for some second-
ary effects, remain unchanged by gas side
pressure. Effects of pressure on cost of auxil-
iaries such as coal handling equipment cannot
be ignored. However, since they are beyond
the scope of this presentation they will
temporarily be overlooked.
The cost of the shell is affected by pressure
level, gas velocity, capacity, and diameter.
The latter two variables are related in that the
designer can resort to modular design at an
appropriate plant capacity to minimize the
effects of a further increase in shell size and
to make use of shop fabrication techniques
rather than field erection. Below a 12-1/2 ft
diameter, a substantial reduction in cost can
be realized by using shop fabrication. Modular
construction, however, is not without bounds.
When more than four or five vessels are con-
sidered, the effective reduction in vessel size
diminishes rapidly, as shown in Figure 2. Gas
velocity is more or less fixed by the combus-
tion process. Any effect it may have on shell
size is secondary compared to the other varia-
bles. As pressure level increases, shell size
decreases rapidly due to the increase in gas
density. Above 10 atmospheres, this effect on
shell size diminishes, as shown in Figure 3.
This diminished effect is compounded by the
increase in shell thickness required to with-
stand the higher pressure. Considering that
the incremental cost of auxiliary equipment,
such as the coal handling system, also in-
creases with an increase in pressure, 10 atmos-
pheres appears to be an optimum pressure
level.
Steam Conditions
The economics can be enhanced by apply-
ing tne pressurized boiler to either the super-
V-4=2
critical or subcritical pressure steam cycle and
using the once-through principle. Doing so
eliminates the large and costly steam drum
required by a natural circulation boiler, as
well as the risers and downcomers which
would otherwise require numerous penetra-
tions of the relatively small pressure contain-
ment shell. Once-through circuitry permits
greater freedom in surface arrangement and
location.
In factors influencing boiler cost, the sub-
critical cycle offers advantages (over the
supercritical) of thinner wall tubes, lower
tube metal temperatures, and some improve-
ment in mean temperature between the gas
and the steam.
Westinghouse thermodynamic and cost
analyses of plants using subcritical and super-
critical steam cycles combined with 10-atmos-
phere gas turbines also show that using super-
critical steam conditions is not advantageous.
Based on the foregoing considerations, the
design conditions of Table 1 were selected as
a basis for the conceptual design of fluidized-
bed boilers for a 300-MW plant. The approxi-
mate steam conditions in the boiler are shown
superimposed on the temperature-enthalpy
diagram in Figure 4.
CONCEPTUAL DESIGNS
Basic Criteria
Conceptual arrangements of the pressurized
fluid-bed boilers are based on several funda-
mental criteria.
1. Where possible, modular construction is
used to take the greatest possible advan-
tage of shop fabrication techniques and
(thereby) of lower cost.
2. A minimum of four completely self-
contained modules per boiler facilitate
the highest possible turndown ratio in
the shortest permissible time.
-------
3. The fluid beds are further divided by
heating functions (i.e., pre-evaporative
heating, evaporation, superheating, and
reheating) to simplify startup proce-
dures, reduce startup time, and improve
control of the various heating functions.
4. Pre-evaporative heating and evaporation
may be combined in a single bed, where
vertical tubes are being considered to
minimize tube bending and manifolding
and to avoid possible tube overheating
due to stratified flow of steam and
water. (Under these circumstances a
minimum of 12 beds are required of two
different sizes.1)
Pressurized operation offers the advantage
of higher permissible pressure drops through
the bed without appreciably affecting plant
economy. This means it is possible to go to
higher gas velocity and much deeper beds. For
example, at 1 atmosphere and gas velocities of
4 ft per second, the bed may be about 2 or
2-1/2 ft deep. A comparable bed at 10 atmos-
pheres and 15- ft per second could be 30 ft
deep. Going to higher pressures changes the
ratio of height to plant area considerably.
Going to high gas velocities runs the risk of
erosion of tube surface, elutriation of large
quantities of fly ash and carbon, and increas-
ing the carbon loss. Instead, conceptual de-
signs are based on stacked beds with gas veloc-
ities of about 4 ft per second; the volume
occupied by the bed remains the same as if
gas velocities of 15 ft per second were
selected. This same volume is accomplished
by splitting the air flow and putting four beds
in parallel on the air/gas side rather than pass-
ing all the air through a single bed.
Preliminary Layouts
Figures 5 and 6 show two possible arrange-
ments of the boiler. In both these arrange-
ments, air enters the shell at the bottom of
the vertical boiler and flows upward through
the space between the shell and a finned tube
water wall which encloses the boiler proper.
The required portion of the air is drawn off at
each bed level, and passed into a plenum
chamber housing the tube headers, and
through a grid plate into the bed. Gases leav-
ing the fill space above each bed pass through
an opening in the inner water-cooled wall into
a central channel and thence downward out
of the vessel. The gases are returned to the gas
turbine in an air-cooled double-lined pipe in
which the air from the compressor is the
coolant.
This arrangement not only provides air-
cooling of pressure parts, headers, and coal
Table 1. DESIGN CONDITIONS FOR COAL-
FIRED, PRESSURIZED, FLUIDIZED-BED
BOILER FOR 300-MW PLANT
Design Characteristic Criteria
Primary steam flow, Ib/hr 1,727,000
Superheater outlet pressure, psig 2500
Superheater outlet temperature," F 1000
Reheat steam flow, Ib/hr 1,644,000
Reheat inlet temperature, °F 650
Reheat outlet temperature, °F 1000
Reheat inlet pressure, psig 600
Reheat outlet pressure, psig 580
Feedwater temperature, °F 578
Fluidized-bed pressure, atm 10
Air flow to boiler, Ib/hr 2,340,000
Air temperature to bed," F 575
Fluid-bed temperature, ° F 1700
Gas temperature leaving boi ler," F 1600
Bed-to-tube heat transfer coefficient, 50
Btu/hr/ft2/°F
V-4-3
-------
conduits, but also ensures that in-leakage of
air (rather than out-leakage of hot gas) will
occur in the event of any breach in the finned
tube enclosure wall. It also reduces the total
external piping required by carrying, the hot
gases within the pressure vessel a part of the
way to the gas turbine.
Annular beds are used in the Figure 5
arrangement, which has the advantage of
making maximum use of the space available
within the shell. It also separates zones of
high differential pressure on the air/gas side
with cylindrical surfaces which minimizes
bracing and reinforcement.
Coal is introduced tangentially at a number
of points on the outer periphery of the beds.
Three injection points per bed should give
adequate distribution to permit blowing the
coal in through tubes located between the
water tubes of the enclosure wall.
Tube bundles in the superheater and re-
heater sections consist of vertical, multi-start,
helical coils that fill the annular beds. The
vertical evaporating tubes are arranged radi-
ally in the annular bed and are connected to
ring-type headers. The enclosure wall tubes
serve as part of the pre-evaporator surface; the
balance is within the evaporator bed.
Helical coils are advantageous for this appli-
cation because of the relatively small number
of tubes leading into and out of the bed. The
same is true of flat spiral coils, which might
also be employed. However, coils have the
disadvantage of high fabrication cost. Also, it
is impossible to replace individual tubes
within a coil bundle; the only recourse is to
plug the ends of leaking tubes.
The arrangement of Figure 6 uses rectan-
gular beds and the more conventional hori-
zontal return-bend tube elements in the super-
heater and reheater. This arrangement pro-
vides for easier maintenance but has the
disadvantage of large flat surfaces which have
to be reinforced to withstand gas differential
V-4-4
pressures. Space utilization within the shell is
not as efficient as with annular beds.
Cyclones and dip legs (not shown in
Figures 5 and 6) are in the free space above
beds to collect elutriated material and return
it to the beds.
Note that the just-discussed concepts are
preliminary. They are two of several bed-and-
surface arrangements that will be critically
evaluated prior to developing a single detailed
design. Other concepts include a single deep
bed in each vessel, and horizontal vessels
containing relatively shallow beds. Much more
detailed study of the manufacturing, assem-
bly, and maintenance problems involved in
these arrangements is required.
Surface Requirements
Heating surface area and bed volume re-
quirements of the boiler have been calculated
by assuming that 2-in. O.D. tubes are used
throughout and that the tubes in the bed are
placed on an average spacing equivalent to a
4-in. square pitch. Although 2-in. O.D. tubes
throughout may not be optimum, they are a
reasonable choice, considering problems of
support, possible vibration, and the number
of circuits and tube-to-heater connections. It
has also been conservatively assumed that the
bed-to-tube heat transfer coefficient is con-
stant throughout the bed at 50 Btu/hr/ft2/°F,
regardless of tube orientation. Table 2 gives
the results of calculations for the annular bed
arrangement shown in Figure 5. For the
rectangular bed arrangement of Figure 6, the
amount of pre-evaporator surface in the bed
would be reduced because of the greater
amount of enclosure wall surface.
Using smaller than 2-in O.D. tubes of the
same pattern and pitch-to-diameter ratio re-
duces bed volume considerably. For example,
for 1-1/2 in. tubes on 3-in. centers and 1-in.
tubes on 2-in. centers, the bed volume will be
about three-quarters and one-third as much,
-------
respectively, as for 2-in. tubes on 4-in.
centers.
COMPARISON WITH CONVENTIONAL
BOILER
Table 3 compares a pressurized fluidized-
bed boiler to a conventional boiler for the
same steam conditions and plant output. The
maximum mean tube wall temperatures are
nearly the same in both units. The cost of
material in the superheater and reheater of
the fluidized-bed boiler is only about one-
fifth of that in the conventional boiler. Sur-
prisingly, no austenitic steel is required in the
reheater. All this, however, is based on the
relatively conservative bed-to-tube heat trans-
fer coefficient of 50 Btu/hr/ft2/°F. Any in-
crease in this coefficient changes the material
picture considerably.
Table 2. CALCULATED SURFACES AND BED VOLUMES-
PRESSURIZED, FLUIDIZED-BED BOILER FOR 300-MW
PLANT3
Component
"Pre-evaporator
cEnclosure walls
Immersed in bed
Evaporator
Superheater
Reheater
Heating
surface,
ft2
,
10,500
3,300
9,780
15,820
8,350
Bed
volume,
ft3
-
700
2,070
3,360
1,770
Overall
H.T. Coeff,
Btu/hr/ft2/°F
20d
47
47
43
39
aAII values are approximate.
^Enclosure wall length based on assumed 15-ftfree height above
each separate bed.
cPortions of the enclosure walls are immersed in the beds.
^Weighted average.
V-4-5
-------
Table 3. COMPARISON OF CONVENTIONALLY FIRED AND
PRESSURIZED FLUIDIZED-BED STEAM GENERATORS
FOR 320-MW PLANT OUTPUT
Characteristic
Conventional
Fluid-bed
Primary steam flow, Ib/hr
Superheater outlet pressure, psig
Superheater outlet temperature, ° F
Reheat steam flow, Ib/hr
Reheat inlet temperature, "F
Reheat outlet temperature, °F
Reheat inlet pressure, psig
Reheat outlet pressure, psig
Feedwater entering steam generator, "F
2,189,000
2,500
1,000
1,936,000
650
1,000
600
580
486
1,727,000
2,500
1,000
1,644,000
650
1,000
600
580
578
Superheater
Surface area, ft2 61,660 15,820
Max. mean tube wall temperature," F 1,095 1,078
Material weights, Ib
Carbon steel 108,000 36,000
C-1/2%Mo 24,000
112% Cr -172% Mo 62,000 17,000
1%Cr-1/2%Mo 19,000
1 -174% Cr -1 /2% Mo 73,000 22,000
2-1 /4% Cr -1 % Mo 430,000 22,000
18Cr-8Ni 20,000 7,000
Total material weight, Ib 693,000 147,000
Relative material cost, % 100 19
Reheater
Surface area, ft2 67,800 8,350
Max. mean tube wall temperature, "F 1,115 1,136
Material weights, Ib
Carbon steel 140,000
1 /2% Cr - 1 /2% Mo 133,000
1 -1 /4% Cr -112% Mo 59,000 24,000
2-174% Cr -1 % Mo 64,000 14,000
9%Cr-1%Mo 21,000
Total material weight, Ib 396,000 59,000
Relative material cost, % 100 18
V-4-6
-------
STEAM TO
STEAM TURBINE
CYCLE
GENERATOR
FEEDWATER FROM
STEAM TURBINE CYCLE
Figure 1. Combined steam-turbine/gas-turbine cycle with pressurized boiler.
1.0
_
uj co
CO UJ
to >
UJ *^
0.8
5 2 °'7
cr t
"- § 0.6
ui 2
E <
D^ 0.5
0.4
1- OC
0.3
0.2
25 o.l
f
T
I I I I I I
I I
I
I I I I I L
5 6 7 8 9 10 11 12 13 14 15 16
NUMBER OF VESSELS SELECTED
Figure 2. Effect of the number of vessels selected on the vessel size.
V-4-7
-------
345
GAS FLOW, 106 Ib/hr
Figure 3. Effect of gas flow and pressure on vessel diameter.
LU
DC
i
1100
1000
900
800
700
600
500
400
PRE-
EVAPORATOR
EVAPORATOR
I
I
I
I
300 400 500 600 700
800 900 1000 1100 1200 1300 1400 1500
ENTHALPY, Btu/lb
Figure 4. Temperature-enthalpy diagram for steam and water.
V-4-8
-------
HELICAL TUBES IN
FLUIDIZED BED
FEEDWATER
PREHEAT
TUBES
VERTICAL TUBES IN
FLUIDIZED BE
EVAPORATOR
SECTION
SUPERHEAT
SECTION
TWO
SECTIONS
(ONE NOT
SHOWN)
VERTICAL TUBES
IN FLUIDIZED BED
FEEDWATER
PREHEAT
TUBES
HORIZONTAL TUBE BANKS
IN FLUIDIZED BED
Figure 5. Pressurized fluidized-
bed boiler — annular beds.
Figure 6. Pressurized fluidized-
bed boiler - rectangular beds.
V4-9
-------
5. SOME ECONOMIC ASPECTS
OF HIGH-TEMPERATURE
STEAM CYCLES
D. E. ELLIOTT AND E. M. HEALEY
University of Aston, England
INTRODUCTION
One of the most important aspects of
fluidised-bed combustion is the very low
corrosion rates which occur when metals are
immersed in the solids; hitherto, steam cycles
have been limited to a temperature of 1050°F
because of fouling and corrosion of super-
heater tubes. Thus the heat rate of steam
plant, which showed a downward trend until
the mid 1950s, has now flattened out at
about 2.5 kW heat/kW electricity. As flui-
dised-bed combustion may allow us to
continue the downward trend of the heat rate
curve, we have undertaken a brief examina-
tion of the cost of achieving better steaming
conditions.
FOULING AND CORROSION
Dainton and Elliott1 immersed a typical
stainless steel probe for 120 hours in a fiui-
dised bed burning Thoresby coal at a tempera-
ture of 1300°F. This coal, notoriously
viscous, has a high chlorine/sulphur content
and would cause a normal power station to
shut down within a day because of savage
fouling. The probe was cooled internally by
air so that the exterior in contact with the ash
varied in temperature from 900 to 1200°F.
The metal surface which was at a temperature
above 1100°F was clear of any deposit or
corrosion; at temperatures below 1100°F,
fouling became progressively worse as the
temperature decreased. The deposit, mainly
V-5-1
sodium chloride, was accompanied by a small
amount of surface pitting. By contrast, the
fouling of superheater tubes in a test rig at
B.C.U.R.A., simulating pulverised fuel firing
and using the same coal, gave massive deposits
which almost bridged across the tube bundle
in a 24-hour experiment.
As the absence of fouling and corrosion is
one of the cornerstones of our present pro-
posals, it is worth looking at some of the
fundamentals involved. Figure 1 shows the
variation with temperature of the vapour
pressure of some of the alkali metal salts
present in most coal ashes. It will be seen that
the vapour pressure levels at normal flame
temperatures of about 2500°F are several
orders of magnitude higher than those in a
fluidised bed operating at 1500°F. As fouling
is a process of deposition from the vapour
phase, it is reasonable to predict much lower
fouling rates at the lower temperatures.
Figure 1 also shows that the vapour pressure
of sodium chloride at 1300°F is significantly
higher than that of most of the other com-
pounds and would therefore be the one most
likely to be deposited; this agrees with experi-
mental results. It is interesting to note that
the vapour pressures of the sulphates, which
are a common cause of fouling in normal
boilers, is negligible under fluidised-bed condi-
tions.
The absence of fouling on the metal sur-
faces which are above 1100°F can be likened
to the phenomenon of deposition of dew
-------
from moist air. Unless there is a sufficient
temperature differential, little deposition
occurs; if there is any, it could well be re-
moved by the mildly abrasive action of the
solids. Only when there is a substantial tem-
perature difference does the deposit adhere. It
should be pointed out that the experiments
were conducted with very fine coal at low
fluidising velocities; we could expect less
deposits when using larger coal, unless the
nature of the fluidising conditions gives rise to
particle temperatures well in excess of bed
temperature. When similar trials were under-
taken with coals which are considered to be
bad (but tolerable) to use in normal systems,
no fouling whatsoever occurred at any metal
surface temperature.
These preliminary encouraging results have
been supported by experience in the larger
combustors now running; although we need
more evidence from much longer duration
experiments, it is useful to enquire if the
potential to go to higher temperatures would
be worthwhile economically.
STEAM CYCLES
Four steam cycles have been investigated:
the first two were based on the design condi-
tions of an existing supercritical p.f.-fired
power station; the other two were cycles
considered by Downs.2 Table 1 gives working
conditions and the efficiency of these cycles,
together with those of a standard 500-MW set
which has been taken as datum. It is not sug-
gested that these cycles would be adopted in
an actual design; however, they represent (at
the lower end) plant which could be achieved
without departing from current steam turbine
technology and (at the upper end) plant
which could possibly be developed in the next
decade.
FLUIDISED-BED CONDITIONS
Tube cost calculations are based on the
fluidised-bed operating conditions shown in
V-5-2
Table 2. These conditions were deliberately
chosen on the pessimistic side; e.g., we have
evidence that heat transfer coefficients in
excess of 70 Btu/ft2/hr/°F can be achieved,
and it is likely that operating the bed above
1500°F would produce a more economical
tube bundle.
In each case it is assumed that all the super-
heater and reheater tubes are in the fluidised
bed to take full advantage of reduced heat
transfer surface. Cooling of the gases after the
fluidised bed is by normal convective sur-
faces. The assumed fluidising velocity is rather
low (2 ft/sec) to correspond to conditions in a
pressurised system; however, the same argu-
ment would hold for higher velocities.
TUBING DESIGN
A wide range of high-temperature alloys
have been developed for use in gas turbines
which may prove suitable for boiler applica-
tion; the final choice would most likely be
made because of weldability and resistance to
internal waterside corrosion, rather than to
optimise the cost of the tubing. Nevertheless
the cost of tubing to perform a given duty can
be estimated as follows:
1. For each metal considered, calculate
values of tube thickness for metal
temperatures in the range 572-1472°F
(300-800°C), using the formula:
t_dip +40(2f + P)
2f-P P(2f-P)
where,
t = wall thickness (in.) but
never less than 0.092 in.
dj = internal diameter (in.)
P = working pressure (psig)
f = working stress of metal at
the appropriate tempera-
ture in psi.
-------
Table 1. ASSUMED STEAM CYCLE CONDITIONS
Cycle
No.
Datum
1
2
3
4
>s. Point of
^vcomparison
>v
Cycle3 ^-v
X^
Fluidised-bed 'a'
500-MW cycle
(b)
(a)
Drakelow C
(b)
Logical single-
reheat
(b)
Optimum-pressure(a)
single-reheat ...
(b)
Optimum-pressure(a)
double-reheat (b)
(c)
Generated
power.
MW
500
375
464
460
528
Steam
pressure.
psi
2300
585
3500
735
3500
770
6500
1430
6500
1430
286
Boiler working fluid conditions
Temperature
Inlet,
°F
493
689
505
683
500
700
500
700
500
700
700
Outlet,
°F
1050
1050
1100
1050
1300
1250
1300
1250
1300
1250
1250
Enthalpy
Inlet,
Btu/lb
476
1344
498
1324
488
1341
489
1292
489
1292
1369
Outlet,
Btu/lb
1488
1545
1495
1540
1629
1651
1572
1638
1572
1638
1661
Steam
flow
rate.
Mlb/h
3.4
2.66
2.405
1.794
2.405
1.804
2.405
1.804
2.405
1.804
1.353
Steam
cycle
efficiency.
%
42.8
45.9
47.9
48.6
49.4
Assumed
sent out
peak
efficiency.
%
39.5
42.4
44.2
44.8
45.6
Reduction
in fuel
consumption.
%
0
7.35
11.9
13.4
15.4
aEach cycle shows separate conditions for: (a) main steam; (b) reheater No. 1; and (c) reheater No. 2.
-------
The second term in the formula is
a corrosion allowance; vary it
according to experience.
2. Using t, and knowing the fluidised-bed
temperature, 0b, calculate the overall
bed-to-steam heat transfer coefficient,
h, assuming:
(a) a bed-to-tube heat transfer
coefficient of 70 Btu/ft2/h/
°F,
(b) a tube-to-steam heat transfer
coefficient of 500 Btu/ft2/h/
°F,and
(c) values of thermal conductivity
for the material under con-
sideration.
3. Knowing 0fc and h, calculate the
working fluid temperature 0f.
4. Knowing 0^, t, h, 0f, and the metal
density, determine the relative weight
Table 2. ASSUMED FLUIDISED-BED OPERATING
CONDITIONS
Characteristic
Condition
Combustion bed temperature, °F 1472
Gas temperature at inlet to air
heater, °F 662
Waste gas temperature at input
to stack, °F 248
Fluidising velocity, ft/sec 2
Excess air, % 20
Base plate pressure drop, in. wg 5
Bed depth, ft 2
Density of fluidised bed, Ib/ft3 50
Bed-to-tube heat transfer coefficient,
Btu/ft2/hr/°F 70
Tube-to-working-fluid heat transfer
coefficient, Btu/ft2/hr/°F 500
Tube internal diameter (I.P.), in. 1
V-5-4
of material to transfer one unit of
heat at the appropriate temperature.
If the cost of the metal and its cost of
fabrication is estimated, the cost to
transfer a unit of heat can be calcu-
lated. This allows the cheapest metal
to be chosen for any given steam con-
dition. Normally it would not be de-
sirable to use more than four different
materials in one boiler in order to
minimise the number of dissimilar
welds. Figure 2 shows how the relative
costs to transfer a unit of heat would
appear using 2-1/4 percent chrome
steel, Austenitic 316, and Nimonic
PE16 with costs per ton of £750,
£3000, and £7500, respectively, for
steam pressures of 3500 psig, and
using 1-in. I.D. tubes.
5. Finally, calculate the individual costs
of metal for the different sections of
the boiler tubing for the various
cycles. Table 3 gives the results, using
the above materials and mild steel.
For comparison, the cost of the tubes in
the datum standard 500-MW, 2300 psi x
1050°F x 1050°F set, using the same
fluidized-bed conditions would be £ 0.6/kW.
(We have chosen to compare the advanced
cycle conditions with those of a fluidised-bed
boiler; we believe the latter will prove more
economic than normal pulverised-fuel
systems.)
In addition to the increased boiler tubing
cost when going to the higher steaming condi-
tions, there is also an increase in the cost of
the economiser tubing, due to the higher
operating pressure. This increase has been
allowed for in the overall comparison shown
later.
FUEL COST SAVINGS
The present-day worth of the savings in
fuel has been calculated assuming an equiva-
lent life time load factor of 50 percent, a
-------
Table 3. CAPITAL COSTS3 OF 1-IN. BOILER TUBES
Tubing
Cycle 1 Cycle 2 Cycle 3 Cycle 4
Main and superheater
NimonicPE16 0.24 0.74 2.28 1.92
Austenitic316 0.20 0.17
2-1/4% chrome steel 0.18 0.11
Reheater
Nimonic PE16
Austenitic316
2-1/4% chrome steel
Mild steel
Total 0.69 1.29 2.58 2.35
aCosts given in £/kW.
^Figures include both reheaters.
0.02
0.03
0.02
0.17
0.06
0.03
0.01
0.18
0.08
0.03
0.01
0.26b
0.1 1b
0.05b
0.01b
station life of 30 years, and a return on
capital of 7-1/2 percent for fuel costs of
3d/therm and 5d/therm. Table 4 shows these
savings together with tubing and other costs.
OTHER COST SAVINGS
The importance of overall efficiency to the
overall cost of a power station is not always
appreciated. The cost of access roads, coal
storage, conveyor and mixing systems, coal
bunkers, mills, fans, gas cleaning plant,
chimnies, foundations, services, cooling water
supplies, condensers, and water purification
plant, all costs come down pro-rata with the
lowering of the heat rate. Since these items
constitute more than half the total station
cost, "other savings" are indeed most im-
portant. The converse (decrease of efficiency
leads to increased capital cost) is also true, as
has been found in numerous studies aimed at
producing a cheap peak-load power station by
cutting out fuel-saving systems; e.g., feed
heaters and high-temperature superheaters.
DISCUSSION
Table 4 shows that the value of the fuel
cost savings far outweighs the increased cost
of the high-temperaiuic boiler tubing and,
together with the other savings, should allow
more than enough margin to pay for the in-
creased cost of the steam turbine. Note that,
although we could realize a moderate increase
in turbine inlet temperature, say to 1100°F,
without altering turbine design, an increase to
1200°F would almost certainly require a
change in construction of the steam turbine
rotor. This would be a more difficult problem
than the design of the turbine blading which
could rely heavily on gas turbine practice.
However, there would not appear to be any
insuperable problems in ultimately producing
1300°F machines.
Although the estimates made in this prelim-
inary investigation are not precise and until a
detailed design study is made, we could not
indicate the optimum conditions; there seems
to be a good case for serious consideration of
high-temperature systems. In addition to
V-5-5
-------
Table 4. SAVINGS3 IN CAPITAL AND CAPITALISED REVENUE RELATIVE
TO DATUM CYCLE
Characteristic
Boiler tubing costs
Economiser tubing costs
Other station costs'3
Fuel costs
3d/therm
5d/therm
Total savings
3d/therm
5d/therm
Datum
500-MW unit
0.6
0.6
23
72.9
110.1
-
Savings
Cycle 1
-0.09
-0.1
+1.8
+3.8
+6.4
+5.4
+8.0
Cycle 2
-0.69
-0.1
+2.7
+5.9
+9.9
+7.8
+11.8
Cycle3
-1.98
-0.5
+3.0
+6.6
+11.0
+7.1
+11.5
Cycle 4
-1.75
-0.5
+3.6
+7.5
+12.5
+8.8
+13.8
aSavings given in £/kW.
bMost likely an underestimate; the full costs influenced by efficiency have not been com-
pletely examined.
design studies, we need further experimental
evidence of both the internal and external
corrosion resistance of high-temperature
materials; accordingly, it is recommended that
in the next round of pilot-scale plant, oppor-
tunity is taken to incorporate at least some
tubes running under advanced conditions.
Only if we can show evidence of good possi-
bility of achieving high temperatures will
turbine manufacturers be justified in under-
taking the expensive development of ad-
vanced machines.
One interesting aspect of such development
would be the result of an increased demand
for high-temperature metals: would it create a
shortage and a higher price, or lower material
costs? Our calculations suggest that, even with
the very advanced conditions, about 600 tons
of Nimonic (or other equivalent high-
temperature) material would be used in a
2000-MW station. Comparing this with the 20
tons of such alloys which are needed for the
engines of a Concorde, it would appear that,
although the demand would be large enough
V-5-6
to justify quantity production, it is unlikely
to produce a scarcity market.
CONCLUSIONS
1. The incremental cost of the tubing
needed to achieve high-pressure high-
temperature steam in a fluidised-bed boiler is
relatively low: a 6500 psi x 1300°F x 1250°F
x 1250°F boiler should not cost as much as a
conventional pulverised-fuel-fired 2300 psi x
1050°Fxl050°F boiler.
2. The fuel cost savings which would
accrue from these higher efficiency cycles are
substantial; they would more than pay/for^the
extra cost of the high-temperature steam
turbine.
3. In addition to the reduced fuel costs,
there should also be a saving in capital costs
associated with the higher efficiency, because
all plant items related to fuel throughput have
their costs spread over a greater output.
-------
4. Modest increases in steam temperature
(within the existing capabilities of steam
turbines) could be the prime objective of
pilot-scale plant because they should present
no extra development problem for the
fluidised-bed boiler. Opportunity should also
be taken to incorporate much higher tempera-
ture trial sections in such plant in order to
assess the capabilities of such systems.
BIBLIOGRAPHY
1. Dainton, A.D. and D.E. Elliott. Researches
into Combustion of Coal. Seventh World
Power Conference. Moscow, 1968.
2. Downs, J.E. ASME Paper No. 55-SA-76.
ACKNOWLEDGMENT
The authors wish to thank Mr. D. H. Stock-
well of the Coal Research Establishment for
carrying out detailed analyses and for general
assistance in the preparation of this paper.
They also wish to thank the National Coal
Board for permission to publish the work.
The ideas and conclusions expressed are those
of the authors, not necessarily of the Board.
V-5-7
-------
TEMPERATURE, 104/T«K
9
10
11
1400
900 800
TEMPERATURE, °C
600
Figure 1. Variation (with temperature) of vapour pressure of alkali salts.
10
o
H
Z
z
I
8
uj
I
0.5
NIMONIC PEI6
FLUID PRESSURE 3500 psig
TUBE I.D. 1 'n.
I I
400 500 600 700 .800. 900 1000 1100 1200 1300 1400
WORKING FLUID TEMPERATURE °F
Figure 2. Relative cost of tubing for the high-pressure conditions of cycles No. 1 and 2.
V-5-8
-------
6. COMBUSTION
OF OIL OR GAS
IN FLUIDIZED BEDS
L. REH
Lurgi Gesellschaft fuer Chemie und Huettenwesen GmbH
ABSTRACT
Over the last 10 years, Lurgi has built
numerous plants for combustion of liquid
refinery wastes with additional heating by
refinery waste gas. Bad experiences during
startup of the first plants led to a
fundamental study of the combustion of
liquid and gaseous fuels in a fluidized bed.
The principal results of this work are dis-
cussed in this paper.
INTRODUCTION
About 10 years ago, the Lurgi Company
had to solve a problem for several German
refineries: burning liquid wastes from their
waste water treatment plants. Because of the
ever-changing net heating value of these
viscous waste sludges, the fluidized-bed prin-
ciple was chosen.1 >2 >3
COMBUSTION OF LIQUID REFINERY
WASTES
The plants, built as compact units, had
fluidized-bed furnaces of approximately 4.5
ID. The furnaces burned, for instance, 1400
Ib/hr oil sludge with oil content of 12 per-
cent, water content of 70 percent and inor-
ganic matter content of 18 percent The
composition often changed rapidly, so that
nearly all oil-contaminated water had to be
"burned" and the heat requirement of com-
bustion had to be supplied by injected
gaseous fuel.
Figure 1 shows the principal flow scheme
of such a plant. The incoming air was first
preheated by the flue gases of the fluidized-
bed furnace to 900°F and then split into
primary and secondary air streams. The first
was sent through the grate as fluidizing air to
a sand bed into which the oil sludge was
pumped; the second was blown at a very low
angle of inclination over the fluidized-bed
surface level. When necessary, gaseous fuel
was injected over nozzles in the side wall of
the fluidized-bed furnace at a distance not too
high above the grate. The temperature of the
fluidized bed was held in the' 1500°F range;
however, even with means for distribution of
fuel into different fluidized bed sections, the
temperature of the furnace off-gas ranged up
to 1650°F under normal operating condi-
tions, due to afterburning above the sand bed.
The flue gases, completely freed from organic
matter, passed through a tube-and-shell heat
exchanger. The entrained ash was precipitated
in a following cyclone before the gases left for
final dust cleaning and disposal via the stack.
Figure 2 shows that, between net heating
values of 2000 and 2500 Btu/lb of sludge, the
optimal combustion temperature can be
reached without any addition of gaseous fuel
(which is necessary for lower net heating
values) and without any cooling (needed for
higher heating values).
When, for instance, an oil sludge with
nearly no heating value was burned, heavy
afterburning (caused by gaseous fuel) oc-
curred in the upper part of the furnace. The
bed temperature went down and the flue gas
V-6-1
-------
temperature up, sometimes leading to damage
in the heat exchanger. The same occurred
when light volatilizing fuels were burned.4
This problem led us to the study of the
influence of the radial mixing of gases in a
fluidized bed on combustion.
fluidized-bed furnace. The samples were taken
symmetrically to the injection points of fuel
in two vertical planes rectangularly arranged
to each other. By means of balance calcula-
tions, the mean content of unburned carbon
in the fluidizing gas was found in every hori-
zontal layer of the furnace.
INFLUENCE OF RADIAL GAS MIXING
ON COMBUSTION
In cooperation with the Technical Univer-
sity, Karlsruhe, Germany, first tests were
made with cold models under similar fluidi-
zation conditions as in an actual fluidized-bed
furnace. One of the cold models was the same
size as a 32-in. diameter fluidized furnace
which could be fired with gaseous fuels, diesel
oil, or heavy oil.5
For the conditions existing in a fluidized
bed with grate tuyeres with horizontally
blowing openings, we studied, for instance,
the injection length of the jet into a nearly
two-dimensional fixed bed of sand of approxi-
mately 1/20-in. particle diameter. Even at
high velocities (in the region of several hun-
dred feet per second) the jet reached only the
middle of the model at a width of approxi-
mately 1 ft; in the experiment, the influence
of the jet affected the whole width of the
sand bed, 2 ft above the injection point.
Following the model tests, a fluidized
furnace was constructed, having a grate from
seven horizontally blowing tuyeres into
which gaseous fuel could be introduced either
by the central tuyere in the furnace axis
through the grate or by injecting nozzles by
the side; the side nozzles could be used also
for liquid fuels. Because premixed air/fuel gas
mixtures, entering a fluidized bed by ade-
quate grates to avoid back-ignition, give com-
plete combustion, only those cases were
studied in which mixing and combustion of
fuel gases occur inside the bed.
Concentration and temperature profiles
were measured at different heights of the
Figure 3 is the temperature profile for a
test, where town gas, injected through the
central tuyere of the grate, was burned in a
sand bed (1/20-in. mean grain diameter) with
50 percent excess air at a fluidizing velocity
of approximately 5 ft/sec. The profile, typical
for other measurements, shows clearly that
the temperature in the bed adjacent to the
wall (where, preferably, the air should be
passing) is considerably lower than in the
centre of the bed. Some combustion is still
taking place above the bed level, regulated to
60 in. above the grate.
Combustion inside the bed is strongly de-
pendent on the fluidization velocity, as shown
in Figure 4, in which the content of unburned
carbon for the aforementioned combustion
conditions is plotted versus the axial coordi-
nate of the sample point above the grate. At
higher fluidization velocities, higher bed
heights are needed to reach complete combus-
tion because of a limited radial mixing rate,
which also means longer mixing ways in the
direction of gas flow. Especially at higher
fluidization velocities, the content of un-
burned carbon leaving the fluidized bed at a
bed level 60-70 in. above the grate is rather
high; it reached values of 15-20 percent.
The concentration profiles, which will be
published5 later, indicate an excess 01" air in
the bed adjacent to the wall; this excess is due
to insufficient radial mixing, and nearly corre-
sponds with the temperature profile.
When designing fluidized-bed combustion
systems for gaseous or liquid fuels, the possi-
ble advantage of premixing fuel and combus-
tion air should always be kept in mind.
V-6-2
-------
Different constructions, fulfilling the 2. Reh, L. Gaswaerme, 15: No. 8, p. 265-270,
above-mentioned requirements, are already in 1966.
use.
3. Reh, L. Chemie-Ingenieur-Technik. 39: No.
4, p. 165-171, 1967.
4. Reh, L. Chemie-Ingenieur-Technik. 40: No.
BIBLIOGRAPHY 11, p. 509-515, 1968.
l.Giese, W. and P. Schwarz. Brennstoff. 5. Boehm, E. Dissertation, TH Karlsruhe, not
Waerme, Kraft. 18: p. 227-230, 1966. yet published, 1970.
V-6-3
-------
FLUID-BED FURNACE
STARTUP
BURNER
°//
1650 °F
!•*••
/.•v
'&
'•"•£'
'"V
:•''/
£*,*
,*••"'
;=}>:
V.':
>
\ 1500°F
1 '
P- -
fclr -i
W
•?5
••"•'
j^-
Ji':
I"-?.'
'•S •
"*••'
' •'»
<»•*
."•*
Si
K?
i
.'•5;
^i*j
c$C~'
^
FUEL (GAS)
900 OF
.PRIMARY AIR
TUBE HEAT EXCHANGER
AIR
BLOWER
1000 OF
CYCLONE
ASH
Figure 1. Combustion of refinery waste sludge in a fluidized bed.
2500
2000
Q
UJ
m
Q
m
N
5 1500
5
OPTIMAL COMBUSTION TEMPERATURE
frMMl'X*'
EC
1
LU
CL
1000
500
OIL CONTENTS NET H.V.
5 0
10 910
15 1820
20 2720
ASH CONTENTS 10%
500
1000 1500 2000
NET HEATING VALUE, Btu/lb
2500
3000
3500
Figure 2- Combustion temperature for aqueous oil sludge in a fluidized bed without air
preheating.
V-6-4
-------
90,
C/5
o
Q.
X
<
1560°F
FLUIDIZEI>BED LEVEL
10 —
808
RADIAL POSITION, in.
Figure 3. Temperature profiles, town
gas combustion in sand bed, fluidized-
bed furnace.
30 40 50 60
HEIGHT ABOVE GRATE, in.
Figure 4. Carbon content leaving fluidized bed at various sampling heights.
V-6-5
-------
SESSION VI:
Discussion Panel and Summary
PANEL MEMBERS:
Mr. R. P. Hangebrauck, NAPCA, Chairman
Dr. D. H. Archer, Westinghouse
Mr. Shelton Ehrlich, Pope, Evans and Robbins
Professor D. E. Elliott, University of Aston, England
Mr. D. B. Henschel, NAPCA
Dr. C. Y. Wen, University of West Virginia
-------
1. MINUTES OF THE PANEL DISCUSSION
AND SUMMARY SESSION
D. B. HENSCHEL
Division of Process Control Engineering
National Air Pollution Control Administration
The final session (Session VI) of the
Second International Conference on
Fluidized-Bed Combustion was a summary
discussion, led by a panel, covering the major
topics and questions arising during the Con-
ference.
Panel members were Dr. D.H. Archer, Mr.
Shelton Ehrlich, Prof. D.E. Elliott, Mr. R.P.
Hangebrauck (chairman), Mr. D.B. Henschel,
and Prof. C.Y. Wen.
The subjects discussed were:
1. Atmospheric versus pressurized flui-
dized-bed combustion.
2. Gas/solids distribution problems. i
3. SO2 sorbent regeneration versus non-
regeneration.
4. Fluidized-bed boiler design factors.
5. Methods of carbon burnup for com-
bustion efficiency.
6. Corrosion/erosion/deposition in
fluidized-bed boilers.
7. Fluidized-bed combustion for indus-
trial-size boilers.
8. Potential for NOX control.
9. Gasification versus combustion in the
fluidized bed.
10. Fluidized-bed combustion for new
boilers, versus an add-on to existing
units.
Summaries of the discussions appear below
and are number-keyed to correspond with the
list of subjects, above.
1. Atmospheric versus pressurized flui-
dized-bed combustion.
It was generally agreed that the first genera-
tion of fluidized-bed boilers will be atmos-
pheric units. However, many felt that pressur-
ized operation will be important in the
future. It was argued that some of the items
requiring investigation for the development of
pressurized units can and should be studied
soon, even before the technology for neces-
sary high-inlet-temperature turbines is avail-
able.
Dr. Archer (Westinghouse) felt that pres-
surization may be desirable from the stand-
point of both economics and air pollution
control. Both Dr. Archer and Mr. Nicole (Na-
tional Coal Board) said that work should be
started as soon as possible to develop the tur-
bines needed for such an application. Profes-
sor Elliott (University of Aston) indicated
that much development work on pressurized
combustors can be conducted even before
turbine technology is available. Mr. Broadbent
(National Coal Board) stated that gas turbine
development is the major hurdle to be over-
come for pressurized combustors, and that
VI-1-1
-------
this development can proceed concurrently
with the development of atmospheric
fluidized combustors, since turbine work can
be conducted without a fluidized-bed boiler.
Mr. Walker (Babcock and Wilcox) indicated
that 1500-2000 MW "conventional" boilers
are currently on the drawing board, adding
that if 660 MW fluidized-bed units will not be
available until 1980, then fluidized-bed com-
bustion will be "too late with too little."
Professor Elliott argued that once the 660
MW fluidized-bed unit is developed, scale-up
to larger units can proceed rapidly. Mr.
Ehrlich (Pope, Evans and Robbins) and Mr.
Broadbent pointed out that the development
of fluidized boilers could be accelerated if
more money were available.
2. Gas/solids distribution problems.
Professor Elliott stated that gas distribution
is not a problem, adding that, to study solids
distribution, a sufficiently large unit must be
built and operated. Dr. Archer felt that
greater solids feeding problems (possibly in-
cluding hot spots and localized reducing con-
ditions) would be encountered when coal is
finely ground. Mr. Hammons (Esso Research
and Engineering Co.) indicated that, although
Esso is currently studying pulverized coal
combustion in a fluidized bed of lime, he
would not defend the use of coal as fine as
that presently being studied (average particle
size 200 jum) in later stages of the develop-
ment. Mr. Ehrlich argued that possibly Esso
need not use such fine coal in order to achieve
their goal of complete ash elutriation from
the bed. He indicated that, with one partic-
ular coal that Pope, Evans and Robbins
tested, the ashing of 1-inch lumps of this coal
yielded ash that was, for the most part,
smaller than 20 mesh. Such fine ash would
elutriate from the bed as desired by Esso.
Possibly not all coals would behave this way.
3. SC>2 sorbent regeneration versus non-
regeneration.
VI-1-2
Dr. Gorin (Consolidation Coal Co.) stated
that the purpose of regenerating the sorbent is
not to recover the value of the resulting
byproduct, but rather to alleviate the need for
the vast quantities of sorbent which would be
necessary if it is not regenerated. Mr. Ehrlich
also voiced this feeling: he said that regenera-
tion and recovery possibly would apply more
to utility boilers than to industrial boilers,
adding that industrial units might well be
fired with a clean fuel such as the char result-
ing from coal gasification processes.
4. Fluidized-bed boiler design factors.
Mr. Seibel (Erie City Energy Division, Zurn
Industries) indicated that users of conven-
tional boilers have had difficulties with steam
tubes of the relatively small diameter that is
being considered in some current fluidized-
bed boiler designs. He said that, without going
to such small tube diameters, it would be
difficult to fit the required amount of tube
surface in the bed. Professor Elliott stated
that nuclear power plants incorporate tubes as
small as 1/2 to 1 inch in diameter—as small as,
or smaller than, the diameters proposed for
tubes in fluidized boilers. Mr. Seibel agreed,
but indicated that he felt that the nuclear
industry has a different situation.
Dr. Wright (National Coal Board) said that
the British have observed high heat transfer
coefficients in the bed (up to 100 Btu/hr/ft2/
°F) at high tube metal temperatures. He also
indicated that beds composed of dense fine
particles gave the best results. Mr. Bishop
(Pope, Evans and Robbins) added that coeffi-
cients immediately above the bed can be
fairly high, much higher than would be ex-
pected in the absence of a bed.
Mr. Walker questioned the means by which
heat pickup could be controlled in the various
sections of a fluidized-bed unit during turn-
down. In particular, he wondered what would
happen to the water side of the tubes in one
cell of a multiple-bed boiler if that cell had to
be turned off during turndown. Would the
-------
water/steam continue to be circulated
through the cold tubes in that cell? Or would
each of the many cells have an individual
water/steam valve to shut off the flow to that
cell? The superheater tubes were of particular
interest. Mr. Bishop indicated that, according
to the Pope, Evans and Robbins design, most
of the superheater sections would be in the
carbon burnup cell, which would not be
turned off when the boiler is operating; the
limited superheater tubes in any cells which
were turned off would be run cold. Mr.
Walker was concerned about the effect that
cold operation of some of the superheater
tubes might have on steam conditions and the
steam turbine.
Mr. Walker and Mr. Demmy (UGI Corpora-
tion) were skeptical of the ability of cyclones
to collect the -325 mesh limestone that some
investigators are proposing to inject. Professor
Elliott and Mr. Ehrlich stated that their ex-
perience has shown their cyclones to be
effective in collecting the fine lime.
5. Methods of carbon burnup for combus-
tion efficiency.
The National Coal Board representatives
indicated that, operating at low superficial gas
velocities, they have been able to achieve 99
percent combustion efficiency in their 3-foot
square unit by recycling primary fines. High
dust loadings result. Mr. Bishop reported that,
at the high gas velocities employed in the PER
combustors, the best combustion efficiency
attainable by primary recycle is 94 percent.
By utilizing the carbon burnup cell concept
developed by PER, combustion efficiencies of
98 percent have been achieved. The carbon
burnup cell does not create an increase in dust
loading, since recycling is not involved. The
British felt that the high dust loading resulting
from flyash recycling should not create
serious erosion problems in the overbed
convection passes because the relatively low
operating temperature in the fluidized bed
results in softer flyash that is not fused and
sintered as is the ash emitted from higher-
temperature conventional coal-fired units.
6. Corrosion /erosion /deposition in flui-
dized-bed boilers.
The lower operating temperature of the
fluidized bed should result in a less abrasive
flyash and, at the lower temperatures that
have been studied, less volatilization of the
alkali components and less tendency for the
ash to become sticky. Under these circum-
stances, corrosion, erosion, and deposition
should be less serious.
Mr. Demmy, pointing out that lime is a
high-temperature fluxing agent for the ash,
stated that glass could be formed on boiler
interiors during additive injection. Mr. Ehrlich
agreed that fluxing and clinkering might occur
in a marginal situation, when the fluidized
boiler is being operated at a temperature near
which the ash in the particular coal being
burned will be affected in this way. However,
for the low temperatures at which fluidized-
bed boilers can be operated, such effects can
be avoided with most coals. The National
Coal Board, operating at even lower tempera-
tures than are the Americans, should be even
safer in this regard.
7. Fluidized-bed combustion for industrial-
size boilers.
Mr. Walker stated that there will be no
market for coal-fired industrial units unless
the boilers are extremely simple, fully auto-
mated, and foolproof. Mr. Broadbent agreed,
stating that the market for coal-fired boilers
smaller than 100,000 Ib steam/hr would be so
small that even the development of a control
system for such small units would not be
justified. Mr. Walker felt that even 100,000
Ib/hr would be too small. One disadvantage of
coal-fired industrial units is the need for fuel
storage space.
Mr. Demmy added that environmental
considerations also enter into the small user's
choice of fuel. A small operator does not
want to be bothered by local air pollution
VI-1-3
-------
control officials. Therefore, he will opt for gas
or even electricity.
Professor Elliott indicated that he has
studied gas-fired fluidized-bed combustors
using a 100,000 Btu/hr unit.
8. Potential for NOx control.
Dr. Ulmer (Combustion Engineering) stated
that NOX emissions appear to be affected by
the nitrogen content of the fuel, not just by
the reactions of atmospheric N2 and 62- Mr.
Jonke (Argonne National Laboratory) indi-
cated that when natural gas, containing no
organically bound nitrogen, is burned in a
fluidized bed, NOX emissions are low, near
thermodynamic equilibrium levels for the N2
O2 reaction. When Argonne burned coal in
their fluidized bed, NOX emissions were such
that, if the NOX were entirely from nitrogen
in the coal, then only about one-third of the
fuel nitrogen would have been converted to
NOX. Mr. Ehrlich felt that no one has fully
studied how to reduce NOX emissions from
fluidized-bed combustion. He suggested that
one possibility might be to operate the bed at
stoichiometric air and add secondary air over
the bed; work at PER indicates that NOX
emissions might be cut in half by this tech-
nique.
The question was raised concerning the fate
of the nitrogen in the fuel when the unit was
operated as a gasifier. One suggestion was that
NH3 might be formed.
9. Gasification versus combustion in the
fluidized bed.
It was pointed out that gasification and
combustion are different. Dr. Archer sug-
gested that both should be developed. Mr.
Jonke added that char combustion in a
separate fluidized-bed unit will probably have
to be tied in with a coal gasification process.
Dr. Gorin stated that the technology for
gasifying a weakly caking coal, with sulfur
removal, is available today. However, it was
suggested that the current technology might
not be economic.
10. Fluidized-bed combustion for new
boilers, versus an add-on to existing units.
Mr. Walker stated that existing conven-
tional units have been designed for a given
heat release rate. He added that modifying the
heat transfer surface in the boiler, to accom-
modate an add-on fluidized unit, would be
very expensive. He suggested that a new
fluidized-bed unit be built, rather than modi-
fying an existing unit by adding on a fluidized
bed. The Conference seemed in general agree-
ment. Mr. Ehrlich added that ductwork re-
quired for the add-on would be prohibitive if
the fluid bed had to be sited far from the
existing boiler due to space limitations.
Professor Elliott felt that, in the case of
gasification, an add-on process might be more
feasible, specifically the process being de-
veloped for oil by Esso Research in England.
Minutes submitted by:
D.B. Henschel
October 22, 1970
VI-1-4
U. B. GOVERNMENT PRINTING OFFICEl IBM 74S4el/41O3
-------
APPENDIX A
Attendance List
-------
A. APPENDIX
ATTENDANCE LIST
(NOTE: To facilitate their identification, attendees are listed alphabetically together with the
name of the organization they represent. The complete address of each organization represented
at the Conference appears at the end of the list of attendees.)
LIST OF ATTENDEES
Name
Archer, Dr. David H.
Bailie, Dr. Richard C.
Bishop, Mr. John W.
Broadbent, Mr. D.H.
Bryers, Mr. R.W.
Coates, Mr. N.H.
Curran, Mr. G.P.
Demmy, Mr. R.H.
Diehl, Mr. E.K.
Eckerd, Mr. J.W.
Ehrlich, Mr. Shelton
Elliott, Prof. Douglas E.
Feldkirchner, Mr. Harlan L.
Forney, Mr. A.J.
Glenn, Dr. R.A.
Glenn, Mr. Roland D.
Godel, Mr. Albert A.
Gorin, Dr. Everett
Gwynn.Mr. J.D.
Hammons, Mr. Gene A.
Hangebrauck, Mr. R.P.
Hanway, Mr. John E. Jr.
Helfenstine, Mr. Roy J.
Henschel, Mr. D.B.
Jarry, Mr. R.L.
Jonke, Mr. A.A.
Keairns, Dr. Dale L.
Lawroski, Dr. Stephen
Lindquist, Mr. W.E.
Lundberg, Mr. R.M.
MacDonald, Mr. B.I.
Mansfield, Dr. Vaughan
Marshall, Mr. Keith
Meyers, Mr. Sheldon
Moss, Dr. Gerry
Nicole, Mr. T.C.L.
Rakes, Mr. S.L.
Rawdon, Mr. A.H.
Representing
Westinghouse
West Virginia University
Pope, Evans and Robbins
National Coal Board (England)
Foster-Wheeler
Bureau of Mines (Morgantown)
Consolidation Coal
U.G.I.
Bituminous Coal Research
Bureau of Mines (Morgantown)
Pope, Evans and Robbins
University of Aston (England)
Institute of Gas Technology
Bureau of Mines (Pittsburgh)
Bituminous Coal Research
Pope, Evans and Robbins
Societe Anonyme Activit (France)
Consolidation Coal
Balfour-Beatty Power Consultants (England)
Esso Research and Engineering (Linden)
NAPCA (Cincinnati)
Chicago Bridge and Iron
Illinois State Geological Survey
NAPCA (Durham)
Argonne National Laboratory
Argonne National Laboratory
Westinghouse
Argonne National Laboratory
Fuller
Commonwealth Edison
Kennecott Copper
Peabody Coal
Balfour-Beatty Power Consultants (England)
NAPCA (Cincinnati)
Esso Petroleum (England)
National Coal Board (England)
NAPCA (Durham)
Riley Stoker
A-l
-------
List of Attendees (Cont)
Name
Reh, Dr. Lothar
Riley, Mr. Boyd T. Jr.
Seibel, Mr. R.V.
Shannon, Mr. Larry
Shultz, Mr. F.G.
Skopp, Mr. Alvin
Slikas, Mr. Charles A.
Smith, Mr. J.B.
Spector, Mr. Marshall L.
Squires, Dr. Arthur M.
Svoboda, Mr. Jean J.
Turner, Mr. P.P.
Ulmer, Dr. Richard C.
Vogel, Mr. G.J.
Walker, Mr. James B. Jr.
Wen. Dr. C.Y.
Wheeler. Mr. Cecil M.
Williams, Dr. D.F.
Wright, Dr. S.J.
Zoschak, Mr. Robert J.
Representing
Lurgi
Bureau of Solid Waste Management
Erie City Energy Division
Midwest Research Institute
Bureau of Mines (Morgantown)
Esso Research and Engineering (Linden)
American Petroleum Institute
Tennessee Valley Authority
Air Products and Chemicals
CCNY
Babcock-Atlantique (France)
NAPCA (Durham)
Combustion Engineering
Argonne National Laboratory
Babcock and Wilcox
West Virginia University
Copeland Systems
National Coal Board (England)
National Coal Board (England)
Foster-Wheeler
LIST OF ORGANIZATIONS REPRESENTED
Name (Represented by)
Air Products and Chemicals, Inc.
(Mr. Spector)
American Petroleum Institute
(Mr. Slikas)
Argonne National Laboratory
(Mr. Jarry, Mr. Jonke,
Dr. Lawroski, Mr. Vogel)
Babcock-Atlantique
(Mr. Svoboda)
Babcock and Wilcox Co.
(Mr. Walker)
Address
P.O. Box 538
Allentown, Pa. 18105
1271 Avenue of the Americas
New York, N.Y. 10020
9700 South Cass Ave.
Argonne, 111. 60439
48 Rue La Boetie
Paris, 8e, France
20 South Buren Ave.
Barberton, Ohio 44203
A-2
-------
List of Organizations Represented (Cont)
Name (Represented by)
Balfour-Beatty Power Consultant Group
Engineering and Power Development
Consultants (Mr. Gwynn,
Mr. Marshall)
Bituminous Coal Research, Inc.
(Mr. Diehl, Dr. Glenn)
Bureau of Mines, U.S.D.I.
(Mr. Coates, Mr. Eckerd,
Mr. Shultz)
(Mr. Forney)
Bureau of Solid Waste Management
(Mr. Riley)
Chicago Bridge and Iron Co.
(Mr. Hanway)
The City College of the City
University of New York (CCNY)
(Dr. Squires)
Combustion Engineering, Inc.
(Dr. Ulmer)
Commonwealth Edison Co.
(Mr. Lundberg)
Consolidation Coal Co., Inc.
(Mr. Curran, Dr. Gorin)
Copeland Systems, Inc.
(Mr. Wheeler)
Erie City Energy Division
Zurn Industries
(Mr. Seibel)
Esso Petroleum Co., Ltd.
(Dr. Moss)
Address
Marlow House
109 Station Road
Sidcup, Kent, England
350 Hochberg Road
Monroeville, Pa. 15146
P.O. Box 880
Collins Ferry Road
Morgantown, W.Va. 26505
4800 Forbes Ave.
Pittsburgh, Pa. 15213
Twinbrook Building
12720Twinbrook Pkwy
Rockville, Md. 20852
Route 59
Plainfield, 111. 60544
245 West 104th St.
New York, N.Y. 10025
1000 Prospect Hill Road
Windsor, Conn. 06095
72 West Adams St.
Chicago, 111. 60690
Library, Pa. 15129
120OakbrookMall
Suite 220
Oakbrook, 111. 60521
Erie, Pa. 16503
Esso Research Centre
Abingdon, Berkshire, England
A-3
-------
List of Organizations Represented (Cont)
Name (Represented by)
Esso Research and Engineering Co.
(Mr. Hammons, Mr. Skopp)
Foster-Wheeler Corp.
(Mr. Bryers, Mr. Zoschak)
Fuller Co.
(Mr. Lindquist)
Illinois State Geological Survey
(Mr. Helfenstine)
Institute of Gas Technology
(Mr. Feldkirchner)
Kennecott Copper Corp.
(Mr. MacDonald)
Lurgi Gesellschaft Fur Chemie und
Huttenwesen mbH
(Dr. Reh)
Midwest Research Institute
(Mr. Shannon)
National Air Pollution Control
Administration (NAPCA),
U.S. DHEW, PHS (Mr. Hangebrauck,
Mr. Meyers)
(Mr. Henschel, Mr. Rakes,
Mr. Turner)
National Coal Board
(Mr. Broadbent, Mr. Nicole,
Dr. Williams, Dr. Wright)
Peabody Coal Co.
(Dr. Mansfield)
Pope, Evans and Robbins
(Mr. Bishop, Mr. Ehrlich,
Mr. Glenn)
Address
Government Research Laboratory
P.O. Box 8
Linden, N.J. 07036
12 Peach Tree Hill Road
Livingston, N.J. 07039
Research and Development Dept.
Catasauqua. Pa. 18032
Natural Resources Bldg.
Urbana,Ill. 61801
3424 South State St.
Chicago, 111. 60616
161 East 42nd St.
New York, N.Y. 10017
Lurgihaus, Gervinusstrasse 17/19
PostFach 9181
6000 Frankfurt (Main), Germany
425 Volker Blvd.
Kansas City, Mo. 64110
5710WoosterPike
Durham, N.C. 27701
411 West Chapel Hill St.
Cincinnati, Ohio 45227
Hobart House, Grosvenor Place
London S.W. 1, England
301 North Memorial Drive
St. Louis, Mo. 63102
515WytheSt.
Alexandria, Va. 22314
A-4
-------
list of Organizations Represented (Cont)
Name (Represented by)
Riley Stoker Corp.
(Mr. Rawdon)
Societe Anonyme Activit
(Mr. Godel)
Tennessee Valley Authority
(Mr. Smith)
U.G.I. Corp.
(Mr. Demmy)
University of Aston at Birmingham
(Prof. Elliott)
Westinghouse Electric Corp.
(Dr. Archer, Dr. Keairns)
West Virginia University
(Dr. Bailie, Dr. Wen)
Address
90 Neponset St.
Worcester, Mass. 01606
66 Rue d'Auteuil
Paris, XVIe, France
720 Chattanooga Bank Bldg.
Chattanooga, Tenn. 3 7401
247 Wyoming Ave.
Kingston, Pa. 18704
Gosta Green
Birmingham, B4 7ET, England
Research and Development Center
Beulah Road, Churchill Borough
Pittsburgh, Pa. 15235
Morgantown, W.Va. 26506
A-5
-------