Proceedings of the
FIRST INTERNATIONAL CONFERENCE
ON
FLUIDIZED BED COMBUSTION
Sponsored by the
National Air Pollution Control Administration
at
Hueston Woods,State Park, Oxford, Ohio
November 18-22, 1968
U. S. Department of Health, Education, and Welfare
Public Health Service
Consumer Protection and Environmental Health Service
National Air Pollution Control Administration
-------
INTRODUCTORY REMARKS
by
Paul W. Spaite
Robert P. Hangebrauck
Process Control Engineering Program
National Air Pollution Control Administration
Cincinnati, Ohio 45227
Presented During
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
-------
SUMMARY OF NAPCA CONCERN AND PLANS
FOR FLUID BED 'COMBUSTION EVALUATION
by
R. P. Hangebrauck and D. B. Henschel
Process Control Engineering Program
National Air Pollution Control Administration
Cincinnati, Ohio 45227
Presented During
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
-------
SUMMARY OF NAPCA CONCERN AND PLANS
FOR FLUID BED COMBUSTION EVALUATION
Potential for Air Pollution Control
In addition to promise of achieving more compact and efficient com-
bustion systems, several features of fluid bed combustion contribute to
our interest in the process as a potential means for controlling air pollu-
tion. Such systems offer one of the very few potential pollution control
solutions for small and intermediate industrial boilers. Sulfur oxides
control is receiving the most attention in our program at present, and the
positive features of fluid bed combustion evident for control of S0? are
as follows: (1) good reaction temperature for limestone/dolomite additives,
(2) excellent gas-solid contactor, (3) possibility of sufficiently long
additive residence time in bed to give good additive utilization, (A) attri-
tion of additive particles in the fluid bed may wear off unreactive shells
(for example, CaSO/) that are believed to limit limestone/dolomite utiliza-
tion when injected into conventional boilers. Developing means for con-
trolling nitrogen oxides emission is also important, and the promising fea-
tures of fluid bed combustion system of interest here are lower combustion
temperatures than found in conventional boilers and the possibility of lower
excess air requirements. Some promise is seen in the area of improved parti-
culate control considering that fluid beds may be operated at temperatures
that cause ash to agglomerate. Also, the possible reduction in excess air
usage would lower the volumetric gas rates handled by dust control equip-
ment. . .. - . . . -
NAPCA-Process Control Engineering Goals
In line with the National Air Pollution Control Administration's goal to
-------
-2- . ...;._-
reduce air pollution from combustion of coal and other fossil fuels, we
have established a preliminary five-year fluid bed combustion program with
the--following goals-: - .'••'-"•.. - . " ." •'.'': ; ' ....•-...
1. Define the potential of fluid bed combustion as a new combustion
process haying inherently lower sulfur oxides, nitrogen oxides., and parti-
culates than conventional small and intermediate industrial boilers and
power station boilers.
2. Demonstrate the technical and economic feasibility of fluid bed
combustion as a means of controlling pollution by designing, constructing,
and testing prototypes of small and intermediate industrial boilers.
Program to Meet Goals
During the period 1968 to 1970, we have or are planning bench-scale, pilot
design, and -evaluation studies. We first became aware of fluid bed com-
bustion potential in 1967 in connection with an Office of Coal Research :.
study undertaken by Pope, Evans and Robbins. The study involved pilot-
scale studies aimed at development of a 500,000-lb/hr coal-fired package
industrial steam boiler. In 1967 NAPCA extended the scope of Pope, Evans
and Robbins' work via an interagency transfer of funds to OCR. This project
is entitled Characterization and Control of Air Pollutants from a Fluidized
Bed Combustor and consists of pilot and prototype module studies to deter-
mine the level of air pollutant emissions from the high velocity fluid bed
system developed by Pope, Evans and Robbins. The study is aimed at deter-
mination of operating and design features that economically reduce sulfur.
oxides and nitrogen oxides to levels well below those found in conventional
combustion systems. ..._.-.
-------
Particular attention is being given to evaluation and optimization of lime-
stone Injection into the fluid bed as a means of sulfur oxides control. The
"following variables are under investigation: •-"•"-' : " '. '.:' •"-'-• .-."
1. Operating variables - bed temperatures, bed depth, bed composition,
excess air, flyash recirculation.
.2. Fuel variables - ash, sulfur content.
3. Additive variables - nature of additive (limestone and dolomite,
calcined and uncalcined), additive stoichiometric ratio, additive particle
size, and steam injection. ...-•-•
This work will he completed this coming spring with the development of a
conceptual design and an economic evaluation of a coal^-fired package fluid<-
bed boiler plant designed to release an economic minimum of SO , NO , parti-
X X
curates.,, and products o.f incomplete combustion.. Further work_will be con-
ducted to evaluate more fully the effects of coal composition, additive com-
position, and bed gas velocity before prototype development will be considered
by the National Air Pollution Control Administration. This study will be
discussed more fully at this conference by Pope, Evans and Robbins.
To complement the work at Pope, Evans and Robbins, the National Air Pollution
Control Administration has entered into an interagency agreement with the AEC
for a laboratory and bench-scale project to allow exploration of wide range
of design, operating, and additive variables. The work is being conducted
at Argonne National Laboratory under the project title Reduction of Atmos-
pheric Pollution by the Application of Fluidized Bed Combustion. Objectives
for the project include:
1. A search for and laboratory-scale thermo and-kinetic evaluation of- --
potential additives suitable in fluid bed applications for the capture of S02,
-------
H9S, and/or NO . ....
£- X
2. The testing of these additives in a coal-fired bench-scale fluid
bed combustor and evaluation of: the effects'.of" adjusting operating variables.
3. The investigation of various design and operating techniques in-
cluding operation of the bed at reducing conditions and pulsation of the com-
bustion air. . ..... ....... .. .
4. A study of the fluid bed combustion of oil.
5. More detailed studies on promising systems including development of
preliminary designs for optimum systems. - .
6. The-development of a-research and development plan .for work required
beyond the bench-scale stage leading .to the design, construction, and testing
of prototypes. . .-..-.
7. .Support work where required for the Pope, Evans and Robbins project.
A meeting was designed to promote cooperation.and exchange of information,
to stimulate work in the area, to define potential of fluid bed combustion,
to define barriers to commercialization, and to shorten the time and level
of effort required to obtain and apply the potential benefits. This meeting
is the First International Conference on Fluid Bed Combustion. ,
Another project was recently started and is being carried out by A. M.Kinney
Inc. The project is entitled Techno-Economic Assessment of Fluid .Bed Combustion
Process. The study consists of an independent, critical engineering evaluation
.of the state-of-the-art of fluid bed combustion and of the. technical and .eco-
nomic feasibility of fluid bed coal combustion for application in the design
of industrial-sized steam boilers and of power plant-sized boilers.
Finally, we hope we will be able to participate in prototype studies in 1971
-------
-5- - - ......
to 1973 including design, construction, and testing of an industrial-sized
packaged boiler with a steam capacity of 500,000 pounds per hour and deter-
mination of its economic feasibility as a low pollution system for steam
generation.
Viewpoint of NAPCA
In closing, I would like to briefly review our viewpoint. Fluid bed com-
bustion offers many potential advantages from the standpoint of air pollu-
tion control. The potential of fluid bed combustion requires further study -
to overcome the technical and economic barriers to commercial realization
still remaining. No matter how promising fluid bed combustion may be from -•
the air pollution control standpoint, the technique must prove to be an
economically competitive method for generating steam arid power before it..
-can ever be accepted in practice. We hope to see eventual scale-up of
.fluidized bed combustors to utility size if the industrial steam unit proves
successful.
-------
Session I
Small Scale Development
.Tuesday, November 19
Discussion--of bench- and pilot-scale studies designed to
characterize fluid bed combustor operation, to optimize
heat release rates and combustion efficiency, and to in-
dicate and overcome mechanical problems associated with
such combustors-. Theoretical considerations.-
-------
OUTLINE
FLUIDIZED BED BOILER
SMALL SCALE DEVELOPMENT
by
J. W. Bishop
Pope, Evans and Robbins
Alexandria/ Virginia
Presented During Session I
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968 •••-
-------
OUTLINE .
FLUIDIZED BED BOILER
SMALL SCALE DEVELOPMENT
1. The object is to develop high capacity coal fired boiler
of minimum size and cost. - , . : ;
2. Fluidized bed combustion, where the oxidizing fluid bed
transfers heat directly into the heating surfaces, has
been selected to meet the pbjective,_ in view of:
a. High rate of direct contact heat transfer (up.to ._
60 Btu/0F/ft2/hr) . _ .'_
b. Smaller space required for completing combustion
(up to 400,000 Btu/ft3).
3. Operational data and process development have been
accomplished in rectangular water cooled columns.
4. Boiler design parameters have been developed-on atmospheric
pilot boiler and a single.module prototype boiler, the .
latter a full scale version of a multicell packaged boiler.
5. Balancing the requirements for high output, .sufficient
fuel residence time and fan power, heat -input: rate-is
based on 106 Btus- per hour per ft2 of bedr
6. Optimum bed heights' at this condition range from 2% "to 3 .ft,
iS A.KTO F
-------
7. Combustion air velocity range from 3 fps, cold air to
6 fps preheated air (at 500-600°F).
8. Superficial velocities at 12-15 fps with bed temperatures
of 1700 to 2100°F.
9. Under these conditions, an 8 x 0 mesh, 2.4 spg inert
bed material ( 8 x 20 after elutriation), is acceptable.
10. The logical source of bed material is the ash in the coal
being burned. This precludes the need for adding inert
to makeup for attrition losses. Excess ash is removed,
preferably on a classified size/density basis. Classified
bea removal has not yet been satisfactorily developed.
11. The system operates smokelessly with 5%. excess air, but
carbon carryover is excessive.
12. Reinjection of flyash into a bed operating at 40% excess
air reduces carbon heat loss to within acceptable limits.
.._..(.!%. or less) . ... ;. ...-. .•;_.. ......... ,1 l^. .......
13. Any rank of coal is an acceptable fuel. .Coal inventory
within the bed while burning a high volatile bituminous
is on the order of 2%. With anthracite, this inventory
rises to 6-7%.
-------
14. Initial lightoff is•accomplished by oversupplying the
bed with coal (15-20% inventory) and injecting a gas
flame into one spot.
15. Lightoff of adjacent beds is by means of allowing live
. bed from the live cell to intermingle with cold material
in the adjacent cell/ followed by coal injection into
the cold cell. . .
16. Coal feed ..is by injection Into the base of the bed. : In
- .-a 2^-3 ft. .high .bed, one inj.ection point can serve, a bed
6' long with a maximum 50°F AT throughout the bed.
17. A 'b:1 turndown can be accomplished by reduction of air
and coal. The lower portion of the bed becomes static,
impervious to coal injection and. cold.
18. Further turndown can be had by cutting out cells. Sudden
cessation of coal and air reduces load to 10% within
one minute. .
19.' Imbedded superheaters require about a quarter of the
surf a-ee needed when-superheaters are placed In an open.
furnace or gas pass.. Overall ;heat. transfe.r coefficient
measures 45 Btu/°F/ft2/hr. -
-------
20. The 6' long, 17~.5" wide single cell boiler has produced
4500 - 5000 Ibs of steam per hour without convection
cooling and 1600 — i700°F exit gas. A 12' long, 25"
wide cell with gases cooled to- 350 - 400°F will produce
18,000 Ibs/hr.
-•'l.?^?.!. ETVA.KS ".AMD .
-------
COAL BURNING IN A SELF-AGGLOMERATING FLUIDIZED BED
- by
W. M. Goldberger
Battelle Memorial Institute
Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
presented during Session I
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
Summary
Bench scale and pilot plant experimental study has demonstrated
that pulverized coal can be burned efficiently in a fluidized bed with
simultaneous agglomeration of the ash within the bed. The self-agglomerating
combustion system can produce hot gases low in fly ash for possible use
with the open cycle gas turbine.
Factors influencing the efficiency of combustion and the ash
collecting efficiency were studied in a laboratory 4-inch ID fluidized
bed furnace and in a pilot plant furnace tapered from a 6-inch ID base
to a 12-inch ID top. Two types of coal were used in the study: a
Pittsburgh No. 8 seam coal and a sub-bituminous coal from the Lake de
Smet, region of Wyoming.
It was observed that collection of ash and agglomeration and
growth of ash particles in the bed occurred at temperatures as low as
1400 F. However, ash collection efficiency was not high below 1900 F.
The ash collection rate increased rapidly above.1900 F and approached
90 percent collection under certain conditions. Above 2100 F, the
-------
sticking tendency of the bed was too great to maintain stable fluidization.
Fluidization stability was also noted to be a function of fluidization
velocity, bed depth, and the size consist of the fluidized bed.
Fluidization velocities in the range of 5-7 fps were used. A means
was developed to correlate the measured ash collection efficiencies.
In combination with a conventional cyclone collector for
removal of any entrained coarse bed particles, the burner-cyclone
system was found to reduce dust loadings in the hot gases to less than
5 grains per 100 scf. In comparison with other coal firing methods
considered for use with the open cycle gas turbine, only the experimental
gas producer would discharge a comparably clean gas. In addition, the
operating flexibility of the fluidized bed method and possible use of
the sensible heat content of the bed solids are isasKBer advantages that
deserve attention in evaluating the merit of industrial application of
this method of burning coal for power generation.
Acknowledgement
This research was undertaken through support of the Union
Carbide Corporation. Patent No. 3,171,369 ^"Combustion of Carbonaceous
Solids" .disclosing the process concept has been issued to Union Carbide.
-------
A FUNDAMENTAL STUDY OF FLUIDIZED BED COMBUSTION OF COAL
by
A.A. Oming or C.R. McCann
U.S. Bureau of Mines
Pittsburgh Coal Research Center
Pittsburgh, Pennsylvania
presented during Session I
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
Objective of the Pittsburgh Coal Research Center fluidized
combustion project was to elucidate the properties of the fluidized
combustion process tha: govern its operation and are essential to an
evaluation of its potential usefulness. A study was made of the - -
thermal balances controlling temperatures within the fluidized bed. -
It was assumed that the bed was in a steady state with the heat
liberated either absorbed directly from the bed or left remaining in
.the hot combustion productMeaving the bed. In order for the bed to
remain in a steady state, the flow of combustion products must be in
balance with the input flows of coal and air and the heat released
in the bed must appear either in gas leaving the bed or be trans-
ferred to sink within the bed.
The rate of burning, at otherwise constant conditions, depends
upon the amount of fuel in the bed. Increased fuel in the bed tends
to increase the rate of burning; decreased fuel in the bed decreases
the rate of burning, so that the amount of fuel in the bed automatically
tends to move towards that needed for a steady state condition.
-------
The bulk volume of fuel needed to keep the steady state burning
rate equal to the coal feed rate may be less than an adequate bed
volume. This volume must be adequate for proper fluidization and to
give sufficient coverage of heat sink surfaces to provide the desired
heat transfer. A refractory grog, a crushed and double screened inert
solid, may be used as needed to fill the bed to the desired volume.
If temperatures are too low, the volume needed for fuel alone to
maintain a steady state may exceed the available space. The coal feed
rate then becomes a dependent variable which must be adjusted as needed
to maintain the bed volume within the available space.
Operation of the fluidized bed with excess air requires either
a low temperature so that the rate of burning is too low for complete
use of all the air supplied or a limited amount of fuel in the bed so
as to give the same result at higher bed temperatures. The latter
condition might involve either almost instantaneous burning of the
fuel as it enters the bed or rapid burning of volatile matter with
slower burning of coke so that some residual fuel can remain in the
bed with excess air at higher temperatures.
As the fuel air ratio is increased into the deficient air range,
the behavior must change. In order to burn more coal with the fixed
air supply, H2 and CO must appear. The percentage of H2 plus CO
increases with increase in the fuel-air ratio in the deficient air
-------
range. Rates of the reaction producing CO and H£ are slow as compared
to the rates of burning to C0£ and H20. The production of CO and H£
will be controlled by temperature and the amount and kind of fuel in
the bed. The required amount of fuel may give an adequate bed volume
without use of the refractory grog.
Calculated thermal balances showed that, as the fuel-air ratio is
increased at constant air flow, the amount of heat that must be trans-
ferred to sink in order to maintain constant temperature, passes
through a maximum. Conversely, an increase in fuel-air ratio under
constant heat sink conditions will increase the bed temperature in the
excess air range and decrease the temperature in the deficient air
range.- v" ... .
Experimental work was conducted in a water cooled reactor, 17" I.D.
with 6 to 12" beds of coal and crushed refractory brick. Experimental
results were generally in accord with the results of the thermal balance
calculations. The bed temperature increases and oxygen in the flue gas
decreases with increased fuel rate in the excess air range. Work in
the deficient air range was limited to an experiment for measurement
of heat transfer rates but an opposite mode of operation was indicated.
-------
Heat transfer coefficients of 35 and 96 Btu per hr., sq. ft., °F
were observed depending on particle size, the fine size giving the
high coefficient* Heat release rates over 8000,000 Btu per hr.,
cu. ft. of bed volume were observed at one "percent excess air and'
1750°F bed temperature.
Full development of the advantages will require additional and
better information in several areas. Loss of combustibles in solids
carryover from the bed should be minimized. It is probable that
operation at higher bed temperatures would reduce the loss but this
should be verified. The fluidized bed can be used as an ash
agglomerator. This requires good control of temperature and develop-
ment of means for continuous extraction of solids from the bed with
size classification and return of fine material. Control of bed
temperature requires control of the portion of the heat release that
is absorbed directly from the bed. This is roughly controlled by
equipment design. Fine control might be obtained by changing the bed
volume so as to vary the amount of heat transfer surface in contact
with the fluidized bed.
There is much yet to be done, but fluidized bed combustion of coal
has attractive features that may make this the next major development
in fuels technology.
-------
PROCESS DEVELOPMENT IN FLUID-BED COMBUSTION
bv
Paul S. Lewis
U.S. Bureau of Mines
Morgantown, West Virginia
Presented during Session I, Small Scale Development
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
INTRODUCTION
Fluid-bed combustion has potentialities for reducing the cost
oC sLeain ^euerat-ion because low grade coals and coal chars can be burned
and boiler designs could be simplified. Because combustion temperatures
are in the range 1,400° to 1,800°F, boiler tube corrosion and fireside ash
deposition should be less extensive than experienced with high temperature
combustion; boiler costs may be reduced as a radiant heat section would
not be required and less costly construction materials would be required.
Emission of air pollutants may be reduced because smaller quantities of
nitrogen oxides are produced at the lower combustion temperatures.
This paper describes the results obtained thus far in a,continuing
pilot study to obtain data for the design and evaluation of a conceptual
process.
-------
EXPERIMENTAL WORK
Description of Apparatus
The conbustor, figure 1, consists of a steel shell .lined with insulating
brick which is faced with castable refractory. Inside diameter is two
feet and the length is six feet. The fuel bed is supported on a conical
shaoed plate perforated by 1/2-inch diam holes through which the fluidizing
air is introduced. Additional fluidizing air is admitted through tuyeres. ..
Two heat exchangers made of 3/4-inch pipe are located 15 inches and
A3 inches above the air distributor. . . .. . -
The flow diagram of the system, figure 2, shows that the coal feed
is mixed with recycled bed material and injected into the base of the
fuel bed, .The residue is removed at the apex of the conical bed support
by a screw conveyor. Products of combustion pass through two stages of
centrifugal separation where most of the entrained solids are removed
and returned to the combustion zone.
Startup is accomplished by burning natural gas in the combustor and
in-|ecting coal into the combustion chamber. After a bed of about one
to two feet is established at temperatures of about 1,200°F, the natural
gas is shut off. Establishing the bed requires about two hours. Coal
feed is continued at a high rate until the desired bed level is attained
and then regulated to maintain bed conditions .compatible with the particular
superficial air velocity.
-------
Combustion of Anthracite
Initial operations for developing a reliable system and operating
procedures occurred in 1967 using a feed of anthracite passed through
a 3/64-inch round hole screen (76.7% FC, 6.3% VM, 9.0% ash, 8.0% moisture)
Results are shown below. . . . . . ...... .
Table 1. - Combustion of Anthracite, 3-foot bed depth
Duration, hr
Bed temp. , °F
Sup. vel. , ft /sec
P.O.C. analysis, %:
CO
CO ......
Burning rate
Air/coal ratio, scf/lb
Percent carbon in
72
1,505
1.41
5.2
16.4
0.0
38.3
110.0
60-68
(63)
45-57
(52)
35-42
(39)
21
1,509
1.94
6.1
15.3
0.0
53.4
109.0
54-77
(67)
49-58
(54)
40-54)
(50)
75
1,540 '
1.77
4.6
15.9 -
N.D.
42.7 -
122.0
61-80
(72)
53-68
(62)
53-84
(66)
59
1,517
2.27
5.6
..-. 15.8
0.1
51.8
132.0
51-78
(67)
51-64
(58)
56-70
(63)
Bottom, rain. -max.
(avg.)
Cyclone 1, min.-max.
(avg.)
Cyclone 2, min.-max.
(avg.)
^Carbon loss from stack not accounted for.
The only operating difficulty experienced with anthracite was some
clinker formation that occurred when bed temperatures exceeded 2,200°F.
Feeds of bituminous coal, even a weekly caking coal, could not be introduced
directly into the fuel bed because it always agglomerated before becoming
mixed with the noncaking inventory in the bed. However, even a strongly
caking coal, such as Pittsburgh seam, could be burned without agglomer-
ating in the bed provided it was mixed with the recycled bed material,
as described above..
-------
Heat Transfer Results
Heat transfer .coefficients have been determined for two modes
of fluid-bed combustion. In one, the coal is injected into a bed of
noncombustible material called "grog"; consequently, carbon content of
the bed is low. In the second, an inventory of noncaking coal char
is maintained in the bed, and raw coal is added continuously; in this
case, a high carbon inventory exists.
Before the tests using low-carbon bed could be made, a suitable grog
for the bed had to be found. Several materials were tried, including
sintered" fly ash, sintered ash from a chain-grate stoker, and sand, but
none of them worked well. Crushed mullite (A120_ + Si02), -8 + 20 mesh
sieve size was found to be satisfactory except that loss by attrition
'was considerable. .-. . ...„_. - ...
Water flow rate through the exchangers was varied for each test.
Overall coefficients were calculated directly from heat transfer data
by the relation Q = UoA0At, where ...
0 = heat removed, Btu per hour
Uo = overall heat-transfer coefficient based on outside area
of pipe, Btu/hr~ft2-°F
Ao = outside area of pipe, ft
At = average temperature difference from water to fluid bed, °F.
A series of tests was made with a low carbon inventory in the
bed. A mixture of 100 pounds of anthractie (#5 Buckwheat) and 100
pounds of mullite was used to start the bed. Then, a mixture of 200 pounds
of hvab coal, Pittsburgh seam, (-1/4 inch) and 100 pounds of mullite
was added. Thereafter, only bituminous coal was added.
-------
Overall coefficients for the first test series were only 25 to 30
litu/hr-ft--°F, which was lower than expected, probably because the heat
exchanger was not completely immersed in the bed. In subsequent tests
with a deeper bed, coefficients were higher, ranging from 42 to 58
Br.u/hr-ft2-°F. In the latter tests, a 100% minus 14-mesh mullite was used
and superficial gas velocities were somewhat lower (4 instead of 6 ft/sec)
than in the earlier tests. Coal rates ranged between 70 and 90 lb/hr. These
results are shown below.
Overall heat-transfer coefficients with a 2-ft bed of mullite and hvab
coal bed velocity 4.0 ft/sec
Bed
temp,
°F
1782
1786
1796
1802
1803
1550
Water
rate,
lb/hr
1368
1959
1093
1246
1783
1908
Lower
Water
temp ,
T r>
72.0
70.7
70.7
70.9
69.9
66.6
tube
0 F
Out
106.0
97.5
111.4
107.0
97.0
91.4
Uo
Btu/hr
ft2-°F
50
56
47
48
51
58
Water
rate
lb/hr
1431
1881
1030
1238
1766
1908
Upper
Water
temp,
In
69.0
68.7
68.5
68.0
69.9
66.0
tube
°F
Out
94.0
88.0
103.0
98.3
89.0
85.0
Uo
Btu/hr-
ft2-°F
38
38
38
40
35
45
Values for heat-transfer coefficients for beds containing a high carbon
inventory were obtained using a 3-ft bed of anthracite. These values are shown
below.
Overall heat-transfer coefficients with a 3-ft bed of anthracite (no grog)
Bed
temp,
1617
1485
1450
1405
Bed
vel,
ft/sec
1.5
1.5
.75
.75
Water
rate
lb/hr
1299
2016
583
1225
Lower
Water
temp,
In
75.6
75.8
75.9
74.0
tube
°F
Out
97.3
88.6
107.2
93.6
Uo
Btu/hr-
ft2-°F
33
33
24
33
1252
1966
750
1241
Upper tube
Water
rate,
lb/hr
Water
temp,
In
°F
Out
Uo
Btu/hr-
ft2-°F
73.9
74.1
74.0
72.5
87.2
82.2
95.0
84.8
20
20
21
21
-------
6
CONCLUSIONS
Results thus far have demonstrated that a wide range of coal can be
burned in a fluidized bed. Improvements in the experimental system are
needed to eliminate dust losses through entrainment in the stack gas.
Bag filters have not been satisfactory, and a scrubber of water screen
design has been installed. Gas analyzers giving continuous recording of
CO, C02 and 02 have been installed. The diameter of the fuel bed has been
reduced to 18 inches from 24 inches, making it possible to reduce the amount
of air neec?e
-------
P.O.C.
CO
10
I
Castable insulating
refractory
Port for
thermocouples
u - inch
carbon steel
Insulat i ng
f i rebri ck
Castable
refractory
Thermocoupl e
Water
Thermocouple
Water
Thermocouple
Igniter port
Heat exchangers,
- s.s. pipe
Ai r distributor,
316 s.s.
Thermocouple
Sight
glass
Air
Natural gas
Coal 4- Ai r
Ash
Fliud-Bed Combustor.
-------
CO
«5
I
Ai r to tuyeres-
Air
Natural gas
Inert gas
Combustor
\
\
P.O.C.
1
Cyclone
Coal and Char
Sampl e
Coal
hopper
2-inch screw
f\/\/\/\/\/\/\/\/\/
Coal
Air
3-inch screw
M
Flowsheet Ho. 2 for Fluid-Bed Combustor
-------
GRID PLATE DESIGN FOR PARTIAL CHAR COMBUSTION
by
Martin E. Sacks
FMC Corporation
Princeton, New Jersey
presented during Session I
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
FMC Corporation has studied the partial combustion of coal char
as part of the COED process, an acronym for Char Oil Energy Development,
sponsored by the Office of Coal Research. The COED process produces a
synthetic crude oil, hydrogen or high Btu gas, and char by the pyrolysis of
co?l'in four or more fluidized beds in series. Char is combusted with oxygen
in the last stage to provide heat for the process. This paper reviews grid
designs used in the last stage of pyrolysis.
To give a saleable product gas, it is desirable to fluidize the last
stage with pure oxygen. For process heat balance the operating temperature
is 1600 °F. To date, it has not been possible to fluidize with pure oxygen in
FMC's 100 Ib. per hour Process Development Unit in Princeton, New Jersey,
because of the formation of ash clinkers in the last stage. These clinkers
probably occur because of poor fluidization in the area of the grid. This
results in local overheating of char particles to above their ash fusion
temperature.
-------
Five different grid plate designs were used in the 8-inch diameter
last stage vessel of the Process Development Unit. The first design employed
seven, 1/4 inch tubes embedded in a refractory. This design quickly proved
inoperable. The second design consisted of an approximately 3.5 inch diameter
multi-orifice plate connected to the 8-inch diameter vessel by a conical
transition piece. This design proved unsatisfactory because of the formation
of clinkers on the walls of the conical transition piece. The third design was
a 7-3/8 inch diameter multi-orifice plate. This plate has been used for the
bulk of the experimental work conducted at Princeton. Clinker formation
has been minimized with most coals, but it has not been possible to operate
with pure oxygen. To attain the desired fluidizing velocity at the grid, streams
containing 80 to 90 percent nitrogen must be employed. Design 4 was a 7-3/8
inch diameter porous plate made from sintered stainless steel. Design 5 was
a cap distributor plate. The caps were supplied by Pope, Evans and Robbins.
Designs 3, 4 and 5 were evaluated under comparable operating
conditions. Fluidizing velocities and grid plate designs were varied from
run to run. After 24 hours of operation, clinker formation in the last stage
was observed for each combination of variables.
At a fluidizing velocity of 1.1 ft. per sec., approximately four times-
the minimum fluidizing velocity, clinker formations were observed with all
three designs. The multi-orifice plate, .design 3, gave the smallest clinker
buildup. However, the clinker was hard and unevenly distributed across the
-------
plate. Clinker formations with the porous plate and capped plate were ap-
proximately the same thickness. The clinker formed when the porous plate,
design 4, was used, was soft, friable and evenly distributed over the cross-
sectional area of the fluidized bed. The clinker formed when employing the
cap-type plate, design 5, was soft and friable, but unevenly distributed across -
the cross-section of the bed. At a fluidizing velocity of 1. 8 ft. per sec., about
six times the minimum fluidizing velocity, negligible clinker build ups were
noted with all three grids. The pressure drop across the grid was greatest
for the multi-orifice plate, about three times greater than for the porous
plate and about six times greater than for the capped distributor plate.
In fluidized beds, regions of high bed density had been reported to
exist near tne grid, indicating a poor quality of fluidization. Some investigators ..
have reported that the height of this high density region is maximized by the
use of cap distributors and minimized by the use of porous plates. Clinker
formation at 1. 1 ft. per sec. indicates that the quality of fluidization near
the grid of the last stage of pyrolysis is best with the multi-orifice grid.
This confirms the literature reports that a cap-type grid results in a large,
high density region in the vicinity of the grid.
It was concluded that in this system a high pressure drop distributor
is required to maintain a good quality of fluidization in the vicinity of the grid.
It appears that the turbulence caused by a high pressure drop-type grid is
more beneficial to a good quality of fluidization than even distribution of the
-------
gas. It also appears that velocities greater than six times the minimum
fluidizing velocity should be employed to minimize clinker formation.
At the present time a 36 ton per day COED pilot plant is being
designed. The last stage vessel of the pilot plant will be a two diameter
vessel. In this way it is hoped to maintain a fluidizing velocity at the-grid
greater than six times the minimum fluidizing velocity employing pure
oxygen.
-------
COMBUSTION EFFICIENCY AND HEAT TRAHSFER STUDIES
by
D. F. Williams
National Coal Board
Coal Research Establishment
Stoke Orchard, Glos., England
Presented during Session I
First International Conference on Fluidised Bed Combustion
Hueston Woods, State Park, Ohio
November 18-22, 1968
The National Coal Board is interested in the development of
fluidised combustion primarily for large scale power generation.
Accordingly our initial bench-scale work was concerned with combustion
efficiency, as a high level of efficiency is required. In this work
an unwashed coal containing 25$ of ash was used as feed, as the ash
particles then form the fluidised bed.
Combustion was carried out in a 6 inch diameter bed. The
fluidising air was distributed by a low pressure drop perforated
plate base covered with refractory pebbles. A low-rank coal crushed
to minus 1/16 inch was fed pneumatically just above the base, and
the heat produced was removed by a water-cooled coil in the bed.
Temperature control to - 2°C was achieved by this means. The bed
height was determined by pressure probes, and was automatically held
constant by running off surplus ash through the base. The gas
leaving the coabustor was dedusted by two cyclones in series, with the
facility for recycling the primary cyclone fines to the bed if
desired.
-------
- 2
Over the rnn^s or aonrjT+.nor»« Investigated the only
combustible gas present above the bed vas carbon monoxide. Its
concentration at TOO C with a fluidising velocity of 2 ft/sec
and with 2% of excess air was 0.7$, and at 800°C the concentration
was less than 0.1$. Loss of unburnt carbon from the bed occurred
almost entirely by elutriation. The loss was reduced fron 6.5$
by weight to 0.15$ by recycling primary fines to the bed. Thus
at 800°C» fluidising at 2 ft/sec with fines recycle, the total
combustibles loss was equivalent to only 0.5$ carbon. Further
experiments were carried out under these conditions. When the
proportion of excess air was lowered below 20$, the loss of
combustibles did not increase markedly until stoichiometric
conditions were reached, the CO concentration then increasing to
l.U$. A change in bed height between 1 ft and 3 ft had no effect
on the loss of CO and carbon: neither did a change from a low-
rank coal to a Pittsburgh coking coal, although anthracite was much
less reactive.
The bed ash contained less than 0.1$ carbon, and its
fluidised density was 35 Ib/cu.ft. Little degradation of the ash
took place, and there was no difficulty in maintaining bed height
with the coals of 15 to 25$ ash content that were tested. No
sintering or caking occurred, even with the highly swelling
Pittsburgh coal, and coal could be fed through the base without
difficulty. It was also shown that bed temperature could be main-
tained constant by automatic control of coal feed rate when the coal
feed varied widely in ash content.
Eeat transfer coefficients were measured in a 1 ft square
section combustor, between the bed of minus l/l6 inch particles and
a horizontal tube immersed in it. Coefficients between 80 and 100
B.t.u./ft'li F were measured, depending upon whether the tube was
-------
cooled with water or air. This result agrees well with predictions
made froa literature data. It indicates that a large saving in
the cost of" high pressure steam tubing can be made by immersing
the tubes in the bed.
Experiments have also been carried out in cold beds of ash.
The elutriation rate of particles was measured as a function of
fluidising velocity, particle terminal velocity and the concentration
of the particles in the bed. In other experiments the elutriated
particles were recycled to the bed and the recycle rate was measured
when it had reached a steady value. The results followed the same
relationships as had been found for the elutriation rate. The mean
residence tine of the fines per cycle, in a bed fluidised at 2 ft/sec,
was found to be between 1 and U minutes.
In a 3 ft square bed, heat transfer coefficients were
measured between a fluidised bed of ash and lj inch o.d. tubes
arranged horizontally in a triangular pattern. Hot water was passed
through the tubes and the bed was fluidised with cold air. The
results shoved that increasing the fluidising velocity fron 2 to
U ft/sec increased the coefficient by 10$, and increasing the tube
spacing fron 2.5 inch to 10 inch increased the coefficient by 25$.
Higher coefficients were measured when the tube bundle was near the
top of the bed. Changing the maximum particle size from minus l/l6
inch to minus 1/8 inch and recycling fines to the bed did not affect
the coefficient significantly.
A 5 ft dianeter vessel has also been constructed to neasure
the rate at which particles move laterally through a fluidised bed.
This will give information about the distance between coal feed
points in a wide bed. With this equipnent it will also be possible
to study the air distribution achieved with different designs of base.
-------
Construction of a 3 ft square combustor is nearing
cosipletion. It will be possible in this coiabustor to measure
coabustion efficiency,.after-burning, and beat transfer coeffic-
ients, end to acquire distribution data over a vide range of
fluidising velocities and particle sizes. Cooling tubes can be
accommodated at different orientations and at different tube
temperatures; alternative fluidising bases can be fitted and
the freeboard height can be altered. It will also be possible to
operate vith or without fines recycle.
-------
Pilot Scale Studies of High Intensity Combustion
in Fluidised Beds.
.by
S.J. Wright
BCUKA Industrial Laboratories
Leatherhead, Surrey, England
Presented, during Session I
First International Conference on Fiuidised. Bed Coabustion
Eueston Woods, State Park, Ohio,
. November 18 - 22, 1968.
-------
Gas-solid fluidised beds have two major attributes, both arising
from the rapid agitation and nixing of a relatively dense particle phase;
there are high heat and. mass transfer rates-between the gas and the solids
and there are relatively high heat transfer rates "between, the bed and surfaces
. -. 1.2:
m aAQ arounu j.~if
Interest in the use of fluidised beds for the combustion of solid
fuels has been prevalent, mainly on the continent of Europe, for about
fifteen years. The research and development effort has been largely directed
towards efficient release of the heat from various forms of solid fuel, such
as anthracite fines , lignites '"^ , oil shale and coal v/ashery tailings °,
which were almost incombustible by conventional methods. I-iich of this work
has been successful, resulting in at least one commercial fluidised-bed stean-
bciler system J, The basic approach of almost all investigators to the problem
of heat recovery has, however, been conventional in that they sought to heat
the combustion, gases to the highest obtainable temperature and to recover this
heat by passing the gases through conventional water-tube boiler systems. That is,
they utilised only the property of high heat and mass transfer rates between
gases and particles in fluidised beds to release heat from otherwise intractable
solid fuels.- It was found that, for fuels having a combustible content in excess
-------
of about 35/->, c O.T. oust ion was so rapid that the equilibrium bod temperature
was in excess of the melting point of the ash . Generally this feature has
oscn exploited by allowing agglomerated ash to accumulate at the bottom of
the bed, whence it can be conveniently extracted.
V.'ith one exception, no attempt was made to utilise the other major
property of fluidised beds, the relatively high heat transfer rates between
the bed and surfaces in and around it. This was because the maintenance
of cooling surfaces in a fluidised bed containing molten ash is an almost
insurmountable problem and also because the basic approach to steam raising
was conventional. The exception was a specialised application where the
oil in oil shale provided heat not only for stean raising but also to
..
calcine the shale for cement manufacture . JTomally the combustible content
of the shale was too low for the equilibrium temperature of the insulated
bed to exceed the softening point of the shale but, in cases where this was . .
not so, cooling tubes were inserted to keep the temperature down and relatively
high heat transfer rates were obtained,
*
In 1964 the-SCUBA inaugurated a literature survey of fluidised
9
bed combustion , in order to assess the potential of this method of burning
coal for increasing combustion intensities in industrial shell boilers in the
steam rate range 2,500 to 18,000 kg/h. It was concluded that it should be
possible to use the fluidised bed not only for combustion, but also as the
primary heat exchanger by utilising the good. hea> transfer properties of
fluidised beds, thus making possible both a high heat release rate and an increasec
ho&t transfer efficiency as compared with conventional shell-boiler firing
systems. The result should be a smaller boiler, with a lower maximum combustion
temperature for a given steam rate, and the sane overall efficiency as a
conventional shell boiler. • • -
-------
Parallel to the BCURA work, the CB3B has pursued ctudies of
conbustion in cooled fluidised beds,, in order to apply the principle to
water-tube boilers for power generation. This has led to a major effort
towards this application now being carried out by the National Coal Board.
Work is also proceeding at BCURA. on the use of pressurised fluidised beds
for the power generation application.
2. Small Scale Ex
Apart from the fact that heat-transfer coefficients at surfaces
within fluidised beds were considerably higher than at surfaces in a gas
stream alone, there was no information in the literaturerelevant to fluidised
conbustion beds heavily cooled by surfaces inserted within them. Therefore,
a series of saall scale experiments was initiated in the BCURA , in an open
topped combustor 305 cm in diameter with a fluidised bed 305 cm deep, to obtain
sufficient data whereby a pilot-scale apparatus could be designed for a thorough
investigation of the characteristics of fluidised conbustion beds burning
British coals. It was found that, if a fluidised bed of coal was heated to
combustion temperatures in the presence of near-stoichiometric quantities of air,
the combustion rate was so rapid that the bed temperature could not be controlled
below the ash softening point; extensive agglomeration and sometimes complete
solidification was unavoidable. It was known that the bed temperature could be
controlled by limiting the oxygen supply to the bed but that this resulted
in high carbon monoxide concentrations in the flue gases, since the fluidised
bed then behaved as a gasifier and this did not seem a useful way to control
a combustion appliance. Research on static combustion beds has shown that
the final formation of CO is due mainly to the reduction of C0? in the
presence of carbon, and, therefore, any reduction in the carbon content of
the bed may be expected to suppress the formation of CO. That this would be so
in fluidised combustion beds was suggested by the fact that those fluidised
combustion systems using a fuel having a very high ash content could maintain
bed .temperatures below the ash softening point whilst using etiochiometric or. ....!.
-------
greater air rates without excessive CO formation •*' . Attempts to control the
co-buc-oion of e fluidised bed of coal only were abandoned, therefore, in favour
of an attempt to control the combustion of coal in a fluidised bed of inert
_„ j. - „„• « -i
Eci werj.&i. . .
This fluidised bed consisted of crushed refractory sized 5.?. mm -
0.8 re; and it v:as found that, once this bed was heated to coabustion temperature,
the combustion of coal in it was very rapid allowing heat release rates in excess
p
of 5.15" iv-'/m (based on distributor area) compared with the maximum for a
o
chain grate fired shell boiler of 1.9 I'^/m (based on grate area). The heat
•v-
capacity of the inert bed and the fact that the carbon content of the bed
at any instant was only froa 5fi ^° 10/o allowed the bulk bed temperature to
be maintained between 800°C and 900°C, by extracting up to 50?o of the heat
release directly from the fluidised bed. - . .
The results from the small-scale experiments were utilised to
cesi^Ti a more completely instrumented pilot-scale coabustor capable of heat
p
release rates up to 3.S"Iv2''/m , equivalent to the heat release rate of the .
smallest shell boiler envisaged. The plant is shown diagramaatically in Fig. 1.
This apparatus was commissioned in July, 1966 and considerable information has-
since been amassed on the coabustion and heat transfer characteristics of \
fluidised beds over a wide range of conditions.
5. Ccrr.bustion
. Experiments have been conducted burning four different coals, one
washed and three untreated, whose main characteristics are shown in Table 1.
-------
TABLE 1. COAL ANALYSES
AS FIRED BASIS
•
C.R.C. No.
Moisture . %
Volatile Matter c/a
Fixed Carbon f*
Ash °fa
Chlorine Content %
Calorific Value kJ'kg
Ash Deformation OG
Temperature
B.S. crucible
swelling iTo.
w
Size analysis
i— i
6.4 - 3.2
3.2 - 1.6
1.6 - 0.8
0.3 - 0.5 ' -
0.5 - 0.2
- 0.2
i
WASHED
902
30.8
51.7
1Q. 9
'26,900
.,. 1290
1 .
.
17.4
25.3
24.1
.8.7
16.1
8.4
J
t
\ 701/801
! 3.9
28,4 ,
47.2
20.5
••
25,400 "
1250
3 -&'
18.0
24.6 •
23.8
9.0
12.8
11.8
UNTREATED
900
5.4
26.4. .;
45.7
22.5
0.8
23,600
1190
i-1 -
26.3
24.0
17.4
8.9
9.0
14.4
'
i
.
' 900
11.3 -
. 25..2
43.0
20.5
0.8 .
i
22,600
1250
.1
28.2
18.9
19.1
9.1 .
• 11.6
13.1
All coals were sized 6.4 mm to zero, as this appears to be about the
limiting size range which can be supplied without recourse to crushing on site.
•
When the washed coal was being used the fluidised bed consisted of particles
of previously prepared crushed refractory sized 3«2 - 0.5 QQ. The movement of
the fluidised bed caused continuous attrition and elutriation of the refractory,
requiring periodic additions of fresh refractory to maintain a constant bed depth,
since little of the coal ash remained in the bed. When burning untreated coal,
however, the shale in the coal supply was more than adequate to replace any losses
-------
due to attrition and the bed depth was controlled "by continuous drainage over one
or other of the weirs. The fluidised bed, under these circumstances, consisting
of shale particles sized 6.4 - 0.5 mm.
The range of combustion conditions investigated is shown in Table 2. •
• • • . ' TABLE 2. RAJTGE OF COMBffSTION" CONDITIONS
Bed depth
Mean bed temperature
Heat input rate
i Excess air
,
Percentage of solids
cyclone recycled
mm
°C
. 305
GOO
M//m2 1 • 1.26
»
*
from
zero
•
zero
- 610
- 900
- 5.76
- 90£
-~80#-*
The figures ^efer to the setting of a flap valve below
the cyclone outlet. The percentage division of the
solics stream was affected to some extent by the flow
pattern out of the cyclone which was in turn affected
by gas and solids loading. Therefore, the figures
are indicative only.
It was found that increasing the bed depth, from 305 to 610 mm,
not only decreased the heat lost from the combustor as carbon.monoxide and
elutriated carbon but also increased the proportion of the heat release
occurring within the bed. Therefore, except for early experiments using the
washed coal, the bed depths were always between 450 - 610 on. If the coal feed
was stopped the bed temperature began to fall steeply within 15 seconds, .
indicating a combustible content of less than 5$ i& a "bed weighing approximately
135 '*g. Samples taken from the operating bed, which, could not be rapidly quenched,
contained less than 2$ combustible matter and material drained over the weirs
had a negligible combustible content. . • .
-------
The mean bed temperature was usually controlled between 870 C and
900 C, above which the bed was liable to sinter. This suggests that the
temperature of some particles in the bed is considerably in excess of the
mean temperature. The bed temperature was kept as high'as possible to
obtain iraxicua conbustion rate and hence maximum combustion efficiency,
particularly for the smaller coal particles which are rapidly elutriated from
the bed. . . . .
With.the combustor operating at around its maximum designed heat
input (5.15 KV/in ) a bed temperature of 870°C could be maintained with 10
to 20 percent excess air. If, when burning untreated coals sized 6.4 ram - 0,
the heat (coal) input was reduced and the air rate reduced, in an attempt to
maintain constant excess air, the larger shale particles began to settle
out progressively at the base of the bed as the fluidising velocity fell. This
settling out reduced the heat-transfer efficiency of the lower part of
the bed and the amount of heat being extracted from the bed fell faster than
the heat input, resulting in a tendency for the bed temperature to rise. The
bed temperature was controlled between 870 C and 900 C by increasing the excess
air as the heat input was reduced. Thus, at the lowest heat input investigated
2
(1.26 Jv//m ), the excess air percentage was between 80 and 90.
The temperature of the fluidised bed responded very rapidly to any
change in the coal feed rate, owing to the low combustible content of the bed-
at any instant, and also to any change in fluidising air rate, owing to the
high rates of heat exchange between fi^s and particles. Control was normally
obtained by adjusting the rate of air supplied, to accommodate any changes in the
quality of the coal being fed at a fixed rate.
It was found that the carbon monoxide content of the flue gases
at
610 mm above above the bed surface was about l^/stoichiometric combustion
conditions,.falling below 0.5/tf, for 20 percent excess air and falling below
0.2$6 for greater than 50 percent excess air. The carbon monoxide.level "•
-------
appeared unaffected by the combustion rate end vertical tranverses approaching
the bed surface detected only slight changes in flue gas composition,
suggesting that only a small proportion of the heat release was occurring above
the fluidised bed.
The heat lost as carbon monoxide vas always less than 1$ of the
heat supplied and the major combustible loss was as solid carbon elutriated
from the fluidised bed. If none of the solids trapped by the cyclone were
recycled to the combustion bed then the heat lost as elutriated carbon could
be &s high as 25$ of the heat supplied. However, when between 50$ and 80$
of the solids were recycled, via the coal feed, such losses were substantially
reduced. ' . . .
Fig. 2 shows, for the untreated coals, the heat lost as elutriated
carbon, expressed as a percentage of the heat supplied, plotted against the
coal feed rate and the superficial fluidising velocity; the data points are
differentiated according to excess air percentage, coal swelling number and
the proportion of solids recycled. Each data point represents an average over
at least one hour arid more usually two hours of operations, the longest
continuous test to date being of 30 hours duration. It can be seen that the
loss increases with both coal feed, rate and superficial fluidising velocity
but that any decrease caused by increasing excess air is slight. Increasing
the proportion of solids recycled from 50$ to 80$ had only marginal effects.
The nediun swelling coals give higher losses owing to an increased combustible
content of the elutriated solids.
Although the data were widely scattered, the heat lost as elutriated
carbon increased from about 5$ to 12.5$ up the range of heat inputs, which
compares with 1-^$ to 3$ over an equivalent proportion of the heat input range
in a chain grate fired shell boiler. The minimum, excess air percentage of 25$
to 30$ required for & chain grate compares with 10$ to 20$ in a fluidised-bed.
-------
The. method of recycling elutriated solids via the coal feed system in
the present experiments is not ideal as the more reactive fresh coal is
likely preferentially to absorb the available oxygen. If solids wore to be
recycled separately a reduction in the combustible loss might be expected.
Chemical analysis of the elutriated solids according to size fraction showed
that if all elutriated solids sized greater than J6 micron were recycled
to the coabustor, smaller particles being caught by a secondary Collector,
then the heat lost as elutriated solids could be reduced to between fy/» .and 5fo
over the coal feed rate range shown in Fig. 2. Thus, bearing in mind the
reduced excess air requirement, the combustion efficiency of fluidised beds
should be ccspotitive with that of conventional firing appliances for shell
boilers,
4. Heat Transfer from the Fj-ttidised Bed.
Hea-u transfer rates nave been measured to various arrangements of tubes
immersed within a fluidised bed burning washed coals sized 6,4 mm - 0, and
within a bed burning two sizes' of untreated coal, 3.2 mm - 0 and 6.4 mm - 0.
All particles smaller" than 0.5 ram were elutriated from the fluidised bed, so
that the range of particle sizes in the bed was 3.2 mm or 6.4 mm - 0.5 mm.
Fig. 5 shows a representative selection of the data plotted as
heat transfer rates against superficial fluidising velocity. This latter
quantity is related to the heat release rate by virtue of the operating
criterion of having minimum excess air. All data points represent an average
over at least one hour, and more usually two hours, of steady conditions. The
'tubes' were 60.3 mm O.D. inserted horizontally in rows of five on 15? mm
centres, each data point represents the average of the middle three tubes, as the
effective area of the outer tubes was too small for accurate measurements.
The 'coil1 consisted of a continuous length of 34 iwa O.D. horizontal tubing
set on a 76.2 am triangular pitch filling the cross section of the combustor.
-------
?ig. -j shows that up to 152 cm from the distributor the heat transfer
rate is proportional to the fluidising velocity, with a bed sized 6.4 .no - 0,5"'cm«
There is no significant difference between the tubes and the coil and the velocity
effect is probably due to the tendency of the larger particles to settle out at
the bottom of the bed resulting in only sluggish fluidisation of this part of the
bed. Between 152 mm and 505 ran from the distributor the heat transfer rates to
the coil are higher than in the lower section, apparently independent of fluidising
*
velocity, and of the same magnitude for. both particle size ranges in the fluidised
bed. They are higher and independent of fluidising velocity because the
quality of fluidisation is always reasonably good at- this levelr regardless of
velocity. Tho fact that heat transfer rates are similar for both bed-size
distributions is probably due to segregation when the bed was sized 6.4 mm - 0.5,
the larger particles sinking to the bottom leaving a particle-size spectrum
close to 5.2 mm - 0.5 nun in the upper part of the bed. The data are confined
to the lower 505 ma of a fluidised bed some 600 mm deep, but heat-transfer rates
in the upper portion of the bed will be similar to those measured. 152 - 502 mm .
from the distributor, if even fluidisation ia maintained. This would give an
9
average heat-transfer ,rate of about 180 kW/m to surfaces immersed throughout
. the volume of a fluidised bed whose superficial fluidising velocity was around
n
4 m/s, equivalent to a heat-release rate of approximately 5.15 MW/m , based on
distributor area, for a bed 0.6 in deep.
For all the data shown in Pig. 5 the mean bed temperature was between
6pO C and 900 C and the heat transfer surfaces were vater cooled having an
average water temperature around 60 C. Under these circumstances, calculations
fron the reasonable assumption that thermal radiation to surfaces in a well
agitated fluidised bed is as from a black-body , suggest that radiation
contributes between 50j£ and 50$ of the heat transfer. For all experiments the
proportion of the heat input which was extracted directly from the fluidised bed
and the walls of the carryover space directly above it fell from about 55$ at low
-------
heat-release rates to 40y£ at the highest rate. I-Iuch of this fall is due to
tho increasing loss of carbon elutriated from the "bed, and any increase in combustic
efficiency will increase the proportion of the heat input extracted from the bed.
5. Conclusions
It is concluded that the combustion of coal in fluidised beds should
allow the heat release rate and heat transfer efficiency of coal-fired shell
-> ' . _
boilers to be increased, with a consequent reduction in size and capital cost
for a given steam rating. The remaining problems are nainly of an engineering
nature, and are to be investigated by BCURA on an industrial scale. To this end a
shell boiler, of novel design permitting utilisation of the fluidised
coabustion principle to raise 4000 kg/h of steam, has been designed and will 'be
installed at Leatherhead in the latter part of 1968. . •
6. AcknowledCements
The author wishes to thank the 3CURA for permission to publish this
paper, Mr. D.J. Keating for his invaluable work during the design and
cocnissioning of the pilot-scale combustor and Messrs. B.F. Pell, 1-I.C. Rogers,
P.D. Brown, P.J. Allen, D. Fitzgerald and B. V7aylen for their untiring
assistance through the various stages of the project.
7. References '
1. Zenz F.A., and Othmer D.F.: Fluidisation and Fluid Particle Systems j
I960, Rheinhold, New York.
2. Leva M. i Fluidisation; 1959, McGraw Hill, New York.
3. Godel A.A.i C.S.A. International Meeting; 1963, London, Doc. No. 7595.
4. Panoiu N. and Cazacu C.: Revue d'electrotechnique et Snergetique; 1962,
Ser. B, 2j 7.
5. Kovotny P.» Prace Ustavu Pro. Vyskum Paliv; 196J, j5, 116.
-------
6. Novotny P.: S.N.T.L. Technical Digest; 1965, No. 12, 885.
7. Priese von G.: Erdol und Kohle; 1961, 14. 702.
8. Ifessotte A.D.H.L.: Brit. Pat. Spec.; 1961, No. 858107.
9. Tea£ue D.S. and Wright S.J.: Private Communication, BCURA Keobers1
Information Circular; 1966, No. 301.
10. Wright S.J. and Keating 2). J.: J6th Int. Cong. Chera. Ind.; 1966,
Bruxelles, !_, 627. ,
-------
b stock
lecit-
?xchangor
Mstributor -
'Ambient
1108
Jacket 7
e
0 )••
^ , 5
L.
§C
, ,: ^ f\
,o.2
>>•*-»
L. U
• «. Q)
n ^
O)1
cv
CM
CM
CO
wv
o / • H
i X^ • /a\
a:
U
0 "
i)
o
()
(}
w w w w
1
1 030
0
, O2O
O
o
o
o
I
-------
20-
o
» .,
.i
air 0-20%
1: excess air 20-50%
'./: „ „ >50%
.10
CX
C
a
o
O
o
c
o
JD
£_
a
T3
O
1^ or\
A
A o
ACL
x x
1
75 100 125 150. 175
Coal feed rate : kg/h
20O 225
! A srs&Uina n°3-51£ \
o
oT
a
-c->
tO
O
a
o
o
A
x
„ „
f
50%recyc/e
„ 1 ' 57% recycle
« «
/ : 80%
10-
..
A •
A A A
.
A
*• •
x
0
2-O • 3-O 4-0
Superficial fluldising velocity : m/s
5-O
Fig.2. Heat lost as elutriated carbon when burning untreated.
coals sized 6-4mm -0
-------
270-
240
i i i
o Co/7 0-152mm from distributor
L Coi(152-305mm ,.
2 Tubes 0-152mrn from distributor
D Tubos 0-152mm from distributor
A Coil 152-305mm „ „
Bed sized 5- 4mm -0-5 mm
CvJ
c
^210-
o
o ^°-
£_
0
o
•P
;
r—»
Ui
o
o 30-
o . i
o '
60-'
30-
0
L
D
D
A
.
A
BO°O
DD
A
A
Q
~
2-0 2-5 3-0 3-5
Superficial fluidlsing velocity: rn/s
4-0
Fig. 3. Heat transfer rates to tubes immersed v/ithin the flu id/sect bet
-------
ORIGIN OF THE " IGNIFLUID COMBUSTION PROCESS" AND
CONCEPTION DESIGN. FOLLOWED BY THE DESCRIPTION
OF THE PROJECT OF A NEW FLUID BED COMBUSTOR
by Albert A. GODEL
President of the "Societe Anonyme Activit"
Paris
presented during Session I (November 19.)
First International Conference on Fluidized Bed Combust'ion
Hueston Woods State Park, Ohio
November 18-22, 1968
-------
- 1 -
Since I was invited to take part to this first international
Conference on Fluidization organised by NAPCA, I feel that to begin with
I must make clear my position in regard to fluidization :
The fluid bed technique was used for the first time in France
in 1947 at our factory of Vernon byoup Societ6 Anonyme Activit.
to manufacture activated carbon .
Apart from this industrial production, we have developed
on a semi-industrial scale from 1950 to 1953 a fluid bed process for the
reduction of iron-ore .
Then from 1953 onwards, iri cooperation with the Swiss Firm
CIPA (Compagnie Industrielle de Precedes & d'Applications SA) and with
our own funds, we have developed a fluid bed process applied to the
combustion of solid fuel .
The latter was thoroughly tested at Vernon with a great
number of fuels, from anthracite to highly coking bituminous coal ,
and the results were sufficiently promising to justify the licensing of
the process under the trade-mark "IGNIFLUID" to Babcock-Atlantique,
-------
- 2 -
former Societ6 Frangaise des Constructions Babcock & Wilcox, and to
various other boiler manufacturers in Europe, South Africa and Asia .
The first report on the above method of combustion was
presented to the "French Society of Engineers" in 1955, an extract of which
was published in "Power" July 1956 .
The Ignifluid process thus industrialized was since then
applied to boilers of increasing capacity for industrial use and power
production .
I will confine this first talk to the description of essential
and theoretical features concerning the so-called Ignifluid process , after
w hich I shall disclose a new method of burning coal in a fluid bed .
This last prospective method is only at its very first stage
of development, I must say, as the cortesponding U. S. Patent has only just
been granted .
The very origin of the first above-mentioned Ignifluid process
proceeds from the following observation :
In a turbulent bed of solid fuel in combustion, as soon as more
or less molten ash particles come into contact , they stick one to the other,
but on the contrary, molten ash particle will not stick to a coal particle .
This is how particles of pure ash increase in size till they sink by gravity
through the bed on to the grate .
-------
- 3 -
By the ascending motion of this grate towards the rear, the
slag is automatically driven through the surface of the bed to the ash pot and
this is one of the essential features of the process .
Of course the grate is divided into sections in order to permit
the adjustment of the pressure in dependence with the depth of the fluid bed .
Fine coal of any type and grading below 3/4" is fed continuously through
an air blown injector located in the front of the combustion chamber, so as to
m aintain a constant level of the fuel in the bed . Primary air , in the
proportion of about 50% of the total amount, is blown through the ~rate under
the pressure of 12" water-gauge . x _ '
As it has been proved possible to control the temperature by
adjusting the bed to a given depth, no severe difficulty has been encountered
in burning coals with either low or high fusibility ash .
Secondary air , in the proportion of approximately 50%, is
injected over the fluid bed in order to complete combustion and all the flue
the, ash of
dust collected in the gas is reinjected so that the entire amount of/the coal
agglomerates into slag containing 3 to 5% carbon .
The stoker is of such reduced dimensions that it can be adapted
under boilers of any size including power plant boilers, even under pressurised
boilers, so that we have often compared such a stoker to a "fluid bed burner",
the difference is that it handles minus 3/4 fines instead of pulverized coal .
-------
_ 4 -
As a conclusion, I think that the very simple combustion method
thus realized has allowed us to dispense with all innovations and further
development work concerning the boiler proper ; funther the scheme has
proved exceedingly profitable, napaely when applied to the level of utility
power plants .
I admit that we have not made full use in our present industrial
program of all the advantages pertaining to fluid bed, namely heat transmission
possibility . But undoubtedly the level of industrialisation now reached is
e nti rely due to our realistic approach to the complex problem of fluid bed
combustion . ' .
HOY/ such fluid bed technique may help solving pollution
problems is, I understand, anohter essential point of this conference but
I must say that we have done little development work on this subject , the
major reason is that nearly all French coals are of a reduced sulphur content
(round 1%) .
Another reason, of course, is that in France, pollution problems
Have only become specially acute in the last few years, with the exception
of dust emission problems .
As regards this last, but important point, we are in a position
t-o ascertain that Igniflv.id bed combustion technique offers a considerable
advantage over pulverizer! firing since the amount of micronic particles
contained in the flue gas rs much smaller, making precipitation easier .
-------
This enables the use of multicy clones instead of electric precipitators in many
cases, namely for top sized industrial boilers or even small utility power
plant boilers .
I may quote the following figures which were offically controlled
at the stack of a 50-55 t/h Babcock -Atlantique Ignifluid fired boiler burning
bitunsinous coal :
O, 4 grams per Nm3 , that is 0, 174 grain : per cubic foot cold
( which corresponds to O, 38 Ib per million BTU) when reinjecting 84% of
the grit, and 0, 5 grams per Nm3 when the total amount of grit is reinjected ,
that is to say 0, 218 grains per cubic foot cold,( which corresponds to 0. 47 Ib
The above figures result from an official test control made by
the CERCHAR which is, as you know, the Development and Research Departmetf
of the French Coal Board .
Since I am offered this opportunity to mention the CERCHAR,
to say
I wan1/how much we are indebted for the help they gave us on many occasions,
namely when official controls were needed for the commissioning of plants,
also recently for bench scale experimentation made in their laboratory in
reference to our new method for reducing flue dust emission over a fluid bed .
On electric precipitation, we have yet no official figures .
However we can assume from very satisfactory results obtained at a 60 MW
plant burning anthracite fines and slurry, which is the only Ignifluid fired plant
equipped with electric precipitators, that there will be no problem to fulfill
the guarantees given : 0, 26 gr. per cubic meter normal (that is to say 0, 113
-------
- 6 ~
grain per cubic foot cold ( which corresponds to 0, 25 Ib per million BTU .
As I said before, we have little experience on the problem
of controlling sulphur oxydes in flue gas, but we are now considering the
following experimentation on a 3 t/h boiler of Babcock-Atlantique at La
Courneuve (France) :
Finely granulated dolomite is fed into the fluid bed in admixture
with the coal, so that it participates in the turbulent motion of the fuel ; it is
expected that the action of the dolomite will be somewhat double : operating
first during its rather long residence time in the fluid bed and afterwards
as a consequence of its erosion in the flue gas .
In this same test plant, we have also begun to study the problem
of neutralising hydroclorine acid in/lue gas emitted by the combustion of
town refuse containing plastic matter waste .
Tests were made on demand of the French Coal Board,
by simply adding a certain proportion of coal ( 20 to 25%) itn admixture
/town
with the refuse burned in the furnace .
Neutralisation was caused by the basic nature of ash in the
coal , so we obtained the following figures :
When burning 25% of low volatile coal from Northern France,
.i*i±h 75% of town refuse, there is a reduction of 50% of the content of
hydroclorine acid in the flue gas . This falls from 10 per million to 5 per
million when compared with the combustion of pure refuse .
-------
- 7 -
Having now done with the essential of the Ignifluid Process,
I may enter the field of prospective development mentioned before, by
disclosing the principle of our new project .
It has been conceived, as you will see, to benefit both from
inte: nsification of combustion and high heat transfer coefficient in the fluid bed.
In this new conception, two or more fluid beds are superposed
in the same steam generator operating preferably in the following manner :
The first fluid bed at the bottom of the boiler is of a conventional
Ignifluid type, thus operating in reducing atmosphere ; it is made up of minus
3/4" or finetcoal, as received, which is burned at a sufficiently high
temperature around 1100°C/ 130CTC ( i. e'. 2012/2372 ?F) to ensure slagging
of the ash .
No heat transfer tubes are incorporated in the bed, but the
lateral walls of the furnace consist of radiating panels .
Above the first fluid bed is a second bed operating in a
slightly oxydising atmosphere, made up with refractory granulated fragments
containing a certain proportion of ash-like carbon particles, coming with
the flue gas from the first bed and, if needed, from an effective incorporation.
The temperature of this second bed round 850° C is lower
/around 850°C.
than that of the first ; it is controlled in order to avoid slagging and yet
allow combustion .
Heat transfer tubes constituting an essential part of the
evaporating system of the boiler are incorporate in it .
-------
- 8 -
Above this second bed has been devised a thirdand even a fourth
bed made up of the same refractory material mixed with incoming ash like
carbon particles and provided with heat exchange tubes for superheating the
steam and eventually heating feed water .
These superposed beds are at decreasing temperature inferior
to the ignition point of the coal . The effluent gas then passes through a multi-
yclone or an electric precipitator and all the flue dust collected is reinjected
into the first high temperature fluid bed , thus securing the maximum possible
efficiency.
Without entering into any detail. I may say that the above boiler
and it is
is planned to use forced circulating water,/working in a pressurized atmosphere
- - _ Since extensive experimentation and industrial application have
been fulfilled by a number of our members on both slagging and non-slagging
fluid beds, I venture to hope that at the issue of this conference, it will be
deemed beneficial to initiate a cooperation .
jn my mind, this cooperation would be essentially aimed on
experimental development work at semi-industrial scale to determine the
possibilities and merits of the at»ove-mentioned combination .
For my part, I want to say that if such a cooperation was set up,
I would expert my utmost efforts to find the best way in which we could
participate for the benefit of all .
-------
Session II
Conceptual Design and
Economic Feasibility
Wednesday morning, November 20
Discussion of prototype and full-scale studies, and of
conceptual design of a full-scale combustor. Economic
comparison of envisioned fluid-bed boiler plants (pack-
age through large utility size) with existing conven-
tional plants. '
-------
POWER STATION BOILERS; PRELIMINARY COSTING AND
DESIGN CONSIDERATIONS
D. F. Williams
Rational Coal Board
Coal Research Establishment
Stoke Orchard, Glos», England
Presented during Session H
First International Conference on Fluidised Bed Combustion
Hueston Woods, State Park, Ohio
November 18-22, 1968
The interest of the N.C.B. in fluidised combustion stemmed
from the difficulty in nicking significant reductions in the cost of
p.f. systeus in order to naintain the competitive position of coal-
fired pover stations in the U.K. A preliminary assessment indicated
that the principal component of the saving offered by fluidised
combustion lay in the cost of the high pressure tubing for the boiler,
superheater and reheater. Taking a bed temperature of 800 C and a
heat transfer coefficient of 50 - 100 B.t.u./ft h°F, at least a four-
fold reduction in tube cost vas estimated. In addition, fluidised
combustion offered savings in coal preparation, since the ccal would
be crushed rather than milled. A further potential advantage arose
from the reduction in combustion temperature, vhich could be expected
to lead to a reduction in fouling and corrosion of the tubes and hence
an increase in plant availability. On the debit side vas an increase
in the fan cost, as a result of the back pressure exerted by the bed
and the air distribution system. On the basis of these items alone,
the net saving in the capitalised cost of a 500 MW station, including
-------
the present vorth of the operating costs, was estimated to be
just under 10$. This vas a promising start because it brought us
vithin reach of our aim to provide a systen vhich would produce
power at not nore than 0.5d per unit in 19T5-
At this stage, consultants have been engaged to provide
a conceptual design of full-scale power stations of output
660 11W and 120 MW, leading to a nore couplets and accurate cost
evaluation. The preliminary calculation shows that if the
fluidising velocity is 3 ft/sec at full load, a bed area of about
2
25tOOO ft is required for a 660 M5-7 boiler. This could be provided
by 15 beds, each 80 ft by 20 ft. A bed height of about 2 ft is
envisaged, with tubes arranged in tvo or three horizontal layers
in the bed, and coal feed points spaced perhaps 10 ft apart along
the fluidising base. The relative merits of arranging the beds
horizontally or vertically are being considered, particularly in
respect to prefabrication and standardisation of components and
to reductions in civils and erection costs that nay well be
possible with this systen.
Other natters requiring attention include the even
distribution of air across each bed, including the design of the
off-gas outlets and the choice of freeboard height so as not to
interfere with this distribution; the control and even distribution
of coal to each feedpoint; the containment of each bed - whether
to cool the walls and roof or to use refractories; the nethod of
start-up, and the method of accommodating large changes in load.
In conjunction with this study a prototype is being
designed, and a decision to build will be taken in 19&9. The
estimates nade so far, and the preliminary views of our consultants,
indicate that there is good reason to press ahead with the programme
with all speed.
-------
The Development of (a) Industrial Fluid-bed Shell Boilers
and (b) Fluid-»bed Combustion under Pressure
by .'
G.G.Thurlow
BCUBA Industrial laboratories
Lea therhead, Sui.-ey, England
Presented during Session H •
First International Conference 'on Fluidised Bed Combustion
Hueston Woods, State Park, Ohio,
e
November 18 - 22, 1968-
BCUEA's work on fluid bed combustion is closely integrated with
that carried out by the National Coal Board. Consideration of
conceptual designs is divided between the two establishments, the
N.C.B. concentrating on the development of designs for large water-
tube units operating at atmospheric pressure whilst BCUKA. is working on
* ' *
(a) small industrial boilers, '
an (b) the extension of the combustion system to pressurised
operation, primarily with large boiler plant in mind.
* . »
Industrial boilers
1. Work so far has concentrated on the "shell" type of unit though
consideration is now being given also to the particular problems
of small water-tube boilers, probably burning £"-0 coal (rather .
than 1/16"-0 preferred for the large water-tube boilers).
-------
2. The shell boiler is popular in the U.K. due to its relative
low cost. Whereas packaged oil-fired and gas-fired shell
boilers are made up to at least 50,000 Ib/h steam, packaged
coal-fired shell boilers are generally limited to about 30,000 Ib/h.
Relative capital costs of coal, oil and gas firing are often in
the ratio 1.35, 0.98, 1 respectively. Market surveys suggest
there is a market for a 30/^,000 Ib/h cheap coal-fired boiler.
3- Research started (at a low level) at BCURA in 196^ to develop a
fluidised combustion system capable of both cheapening coal-fired
shell boilers and of extending their output - at least to the limit
of current oil-fired boilers. The effort increased significantly
. in 1966. .."'.,
k. Fluidised combustion offers cost savings mainly because the high
combustion and high heat transfer rates in the combustion bed lead
to a substantially smaller (-J to 9") overall size of boiler.
Other practical advantages are:
(l) the system is relatively tolerant of variable and poor
quality fuels,.
(2) the low bed temperature (about 850 C) avoids the release
of sticky solids and hence minimises the blockage of smoke
tubes,
(3) the absence of local overheating prevents high thermal stresses and
hence minimises the risk of furnace tube distortion (and
incidentally permits the ready possibility of using the
boiler for indirect chemical process heatingBusing an oil-
based heating fluid).
-------
5. Following encouraging results from an experimental combustor
(burning rate up to 500 Ib/h coal) an experimental prototype
four-pass vertical 8000 Ib/h steam shell boiler has been
designed in collaboration with a manufacturing firm and an
order placed for its manufacture and delivery in February 1969- • . .
6. The fluid-bed will be contained in a k ft diameter by ^ ft 5 in long
furnace tube. The bed will be 2 ft deep and pierced by thirty 2^ in
O.D. tubes at an angle of 10 to the horizontal, allowing about 5®%
of the heat release to be extracted from the bed; tube design has
been tested at up to twice the heat transfer rates expected in the
boiler by studies on a single inclined tube in a' test rig. The
boiler will incorporate a wholly enclosed pneumatic coal feeding
system.
?. The overall height of the boiler (including smoke box and supports)
will be 11-g- ft and the diameter 8 ft. This compares with an 8000 Ib/h
steam chain-grate Economic boiler (heat transfer in furnace tube
less than ^Q% of heat release) with dimensions 15? ft in length
and 85- ft in diameter. The actual volumes occupied by the pressure
shells are respectively:
fluid-bed fired boiler - *K>8 ft5
chain-grate " " - ?66 ft5
8. The boiler will burn a high-ash £"-0 coal to produce an estimated
&i lb steam per Ib coal. This represents a target of 80?£ thermal
efficiency for a gas outlet temperature of 200 C and not more than
5% heat loss due to carbon-in-ash.
9. The actual capital cost considerations are somewhat complex. For
example, a saving of £850 sterling on the chain grate is offset
by the added complications of drilling the furnace tube to receive
-------
the inclined water tubes, and of welding the tubes. Similarly,
.there are other items that tend to balance one another. Perhaps
the firmest indication of capital cost is the cost of the prototype.
If the price of this one-off prototype can be cut, as should be
possible, by 20% when production methods are employed, it will be
slightly cheaper than the equivalent oil-fired version.
10. It is difficult at this stage to make precise estimates of running .
costs since these will be influenced by:
price of coal (assumed throughout that the coal can be
delivered on site -£"-0 size, preferably untreated and dry)
thermal efficiency of boiler (.80% assumed but actual value
has to be determined)
cost of disposal of ash
manpower costs (only one if pneumatic coal system.is used)
electricity costs (fan power etc.)
maintenance costs (including instrument maintenance)
»
11. Postive answers to the various unknowns in capital and running
costs and to some technical questions such as to the greatest
turn-down ratio that can be attained and the range of tolerance
of coal quality, are planned to be available by the end of 19&9-
Instrumentation and automatic control should be completed by the
third quarter of 1970.
12. An estimate of possible reduction in construction time will be
available after the manufacturers have the experience of building
the prototype boiler. Active design and development is also in
hand to find the upper size at which this boiler can be built
to test simplified coal feeding systems, etc.
-------
13. Tests on a two-stage system in which the conventional smoke
tubes used in the present design are replaced by a second
fluidised bed may lead to an even more compact design.
1^. Consideration is now being given to extending the scope of the
basic design of combustion chamber into the packaged water-tube
boiler range up to 1000,000 Ib/h steam.
Fluid-bed Combustion under Pressure
1. Pressurised fluid-bed combustion systems are envisaged for
combined gas/steam turbine driven power generating plant sized
from about 15 MW upwards; the benefits may well be greater in
the sizes suitable for large industrial and central power
stations (e.g. larger than about 80 MW).
Advantages of operating under pressure
2. Fluid-bed combustion at atmospheric pressure offers potential
technical advantages over conventional p.f. combustion for coal-
fired power generating systems. Even greater benefits should
result if. fluid-bed steam generators can be operated under
pressure and if the energy of the combustion gases can be
recovered in gas turbines (or similar heat engines).
3. Operation under pressure facilitates a major reduction in plant
size. It should be possible
(a) to reduce the volume of a fluid-bed boiler at least in proportion
to the square root of the absolute pressure; a conservative
estimate suggests, for example, that a fluid-bed boiler
for a 500 MW power plant operated at 15 - 20 atmospheres could
be contained in a vessel 20 ft diameter and 80 ft long -
which is 1/25th the volume of a conventional boiler of the
same output.
-------
(b) to reduce the cross section area of the fuel bed in
proportion to the increase in absolute pressure; distribution
of coal and air, and collection of steam, will as a
consequence be simpler than at atmospheric pressure.
b. Efficiency of power generation using combined gas/steam turbine
cycles should also be higher than that of conventional steam cycles,
the precise extent of the improvement (1 - k%) depending on the
temperature pressure and details of design of the steam plant;
these can only be defined in the "light of the requirements of
specific applications. . •
5. Finally the pressure process should be able to accommodate load
changes more easily
(a) cycle pressure will change in the same direction as the
load; some reduction in load will therefore be feasible
without a reduction in fluidising velocity, ...
(b) because higher pressure losses are acceptable in the air
. distribution system the velocity can. be reduced to a greater ..
extent before there is a risk of unsatisfactory distribution,
and
(c) because CO formation is less likely to occur under pressure
it may be feasible to operate at a lower fuel' bed temperature
before the CO content of the combustion gases reaches an
uneconomic level.
Investigations in Progress
6. Before these additional benefits of operating under pressure can
be quantified, notional designs for fluid-bed power plant have to
be drawn up; . '
-------
7. The main.objective of the programme of experimental work in
progress at BCUEA is to demonstrate the feasibility of the process;
the crucial factors are:
(a) whether the combustion gases from a fluid-bed can be
sufficiently well cleaned to avoid uneconomic erosion of .
I 6as turbine blades, without also incurring unacceptably
A.-J
high power loss and capital cost, and
(b) whether volatilisation of constituents of the mineral matter
occurs to a sufficiently small extent to avoid serious fouling
or corrosion of turbine blades.
Gas turbines have hitherto found products of combustion from coal
unacceptable because of excessive deposition of, and erosion by,
the mineral matter0 It is confidently expected that the products from
a fluid-bed combustor operating at under QOO C will not present
comparable problems.
8. The main experimental equipment being built to investigate these
problems is a pilot scale combustor designed to burn 1000 Ib/h of
1/16"-0 coal at a pressure of 5 atmospheres. The rectangular fuel
bed (operating at 650 - 800°C) V long, 2' wide and about 2.5' deep
contains 1" outside diameter tubes spaced on 3" centres; the tubes
will absorb about 60$ of the heat on the fuel. The gases leaving
the bed pass through three high efficiency cyclone dust collectors
in series and over a cascade (3" x 2") of turbine blades. The
erosion/corrosion/deposition characteristics of the gases will be
assessed by detailed examination of the blades and cooled metal
specimens that .follow them.
-------
9. The experimental work will also provide information on:
(a) deposition/corrosion/erosion of tubes immersed in the
fluidised fuel beds; specially instrumented tubes are
provided for this purpose; conditions are more exacting
under pressure because of the higher heat release/unit
volume,
(b) fuel and air.distribution,
(c) engineering design of tube arrangements in fuel beds,
(d) problems of part -load operation and control.
10. Supporting investigations are in progress on:
(a) erosion of tubes using a 2 ft square fluid-bed, and
(b) release of alkali and sulphur using a small bench scale
externally heated fluid-bed.
- - (c) tube packing and air distributor design using a full
scale model.
-------
OUTLINE
FLUID!ZED BED BOILER CONCEPTUAL
DESIGN AND ECONOMIC FEASIBILITY
by
J. W. Bishop
Pope, Evans and Robbins
Alexandria, Virginia
Presented During Session II
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
-------
OUTLINE
FLUIDIZED BED BOILER
CONCEPTUAL DESIGN AND ECONOMIC FEASIBILITY
1. General design parameters are as follows:
a. Maximum width of 12', maximum height of 16' to permit
factory assembled boiler shipment to around 80% of
all potential customers.
b. Maximum capacity within these limits (estimate 300,000-
350,000 Ib/hr with a 50-60' boiler length).
c. Lowest practicable capital cost.
d. Lowest practicable operating cost, all factors considered.
e. Capability to operate with a wide range of fuels.
2. Specific parameters currently being used to govern our
design are:
a. Use of common bed material, specifically the ash
received with the coal.
b. Size consist, density and shape factor of bed material
optium for operation at 0.8 to 1.2 MBtus per square
foot fuel input. (Average 3 fps ambient air velocity,
12-15 fps superficial velocity) (-8 mesh coal ash,
2.4 spg currently employed).
-------
c. Heat transfer rates on the order of 50 to 60 Btus/
ft2/hr/°FAT MTD where the bed is in direct contact
with the boiler ' tubes; 45 Btus/f t2/hr/°F MTD for
imbedded superheaters.
d. Beds of 18" static height and 36 - 42" effective
expanded height.
e. Limit of 20" w.g. draft pressure at the bottom of the
bed, fan power not exceeding 1.2 hp per million Btu
input. Grid loss 30% of bed draft pressure. I.
f. All direct contact heat transferring surfaces vertical.
g. — Combustion chambers of rectangular cross section
capable of easy access and maintenance with ratio of
periphery (ft.) to bed surface (ft2) on the order of
(1650) - (Tj., - Ts) , where Ts is the saturation
temperature of the steam and Tfc is the bed temperature.
h. Under conditions described above, bed temperatures
between 1600 and 2000°F. (1800-1900°F is actually
being used) .
i. Preheated air temperature at 500-600°F.
j. Any single screened coal % x 0 to h x 0, 4 to 40%
volatile, minimum 2100°F ash softening temperature,
sufficiently "unwashed" to preclude bed level shrinkage.
-------
k. Coal fired sections to operate with 3 to 10% excess air.
1. Flyash fired (reinjected) section to operate at 30 to
40% excess air; all flyash reinjected into this zone,
only.
m. Minimum overbed combustion space, but sufficient to
preclude elutriation of larger particles.
n. Pneumatic feed of coal at bottom of bed, 6" intervals.
o. Direct overbed ignition of one "lightoff" cell,
4MBtu/hr required.
p. Lightoff of subsequent cells by means of limited
openings in intercell boiler tube walls.
q. Classified removal of excess bed material (+8 mesh)
to permit use of larger size consist coal.
r. Automatic control based on excess 02•
s. Reduction of exit gas temperature to 350-400°F by
air preheater and economizer.
t. All particulate in excess of 44 microns will be collected
in a low draft loss dust collector for reinjection; the
balance which exhibits low carbon content will pass on to
the pollution collector or scrubber. Where stringent SC>
-------
regulations are in effect, the scrubber may incorporate
SC>2 removal features.
3. A design of a 250,000 Ib/hr factory assembled railroad
transportable packaged boiler has been prepared. This
consists of 13 boiler cells, approximately 12' long x 25"
wide plus a carbon burnup cell. All reinjected ash is
into the carbon burnup cell.
4. A 250,000 Ib/hr packaged fluid bed boiler will cost less
than a component assembled oil or gas fired boiler of
equal capacity.
5. Based upon a boiler installed in an existing plant with
existing fuel handling and storage facilities, no
differential in fuel price is needed to justify coal
over gas.
6. Based upon a completely new facility, a 2£ per million
Btu differential in fuel price (in favor of coal) is
needed to justify fluid bed firing, whereas a 7-8C
differential is required for a spreader stoker when figured
on the same basis.
7. A 100,000 Ib/hr prototype unit is now being developed
for correction of deficiencies and for making fluidized
bed boilers marketable.
-------
THE "IGNIFLUID" FLUID BED COMBUSTION PROCESS •
(
i
AT ITS PRESENT STAGE OF INDUSTRIAL DEVELOPMENT [
by Albert A. GODEL
President of the "Sioci^te' Anonyme Activit"
Paris
presented during Session II ( November 20)
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
-------
Yesterday I explained the principle of the Ignifluid process
of combustion., today I shall go into the industrialisation of the process.
The 34 units completed up to date by three of our licen- .
cees of France, Great Britain and Japan, cover the following range of
outputs : . " .
- from 3 t/h (6.600 Ib/h) to 115 t/h (250.000 Ib/h),
two of the latter built by Babcock-Atlanti.que are . -_.i boilers "twinned*1
to a single 60 MW turbo-alternator ; now tenders for 250 MW power plants
comprising twin boilers each of 365 t/h (800.000 Ib/h) are being made.
All these furnaces include the same essential items
which I will now ."Enumerate as follows :
- a chain grate stoker of approximately 1/1 Oth. of the surface of a corres-
ponding conventional stoker, rises at an inclination of approximately 8°
so that the upper moving strands will discharge the slag at the rear of the
furnace into the ash pot ; the chain grate is made of self-cleaning steel
links. As already said, primary air is blown into the sections of the chain-grate,
-------
This stoker is placed between naturally formed fuel banks extending
between lateral and front walls.
The heat production developped by the chain is 10 million Kg/cal,
, that is to say 4 million BTU per sq. feet.
The banks play no part in the combustion except for avoiding
slagging along the walls and for providing a wide extension of the fluid
surface.
- Above the level of the fluid bed, a row of nozzles inject secondary air.
Both primary and secondary air being blown by a single forced draught fan.
The temperature of the primary and secondary air is atmospheric in certain
plants or as high as 190°Cfor primary air, and, for instance, 280°C for secundary
air. '
A satisfactory proportion of excess air for the final combustion
in the furnace is 25 %, but promising tests are made to reduce this proportion
to less than 10 %.
- Fly ash reinjection nozzles are usually placed at the rear of the furnace
at an intermediate level between the fluid bed and the secondary air rein-
jection nozzles, that is in the hottest combustion zone of the furnace. They
handle the total amount of fly ash collected from the flue gas. Grit arresters
or electric precipitators have of course to be used.
-------
- 3 -
We have seen that the level of the fluid bed is maintained as
a rule constant and this leads me to describe the control system which is
used for this purpose :
Three essential!are used :
- the first loop controlling the level comprises a water cooled pressure probe
inserted at the bottom of the fluid bed which transmits the pressure to a con-
trolle/t, recorder ; this recorder actuates a servo-raotor regulating the feed of
the coal to maintain constant the level of the bed.
- the second loop controls primary air in relation to the steam pressure.
- the third loop controls the adjustment of the proportion of secondary air
added in view of completing combustion, this is done by the usual method
of comparing through a "boiler-meter" the amount of flue gas with the steam
production.
As usual of course, other loops are used to control the
draught created by the induced draught fan, and to control the steam su-
perheat and the water level.
The control panels may be located near the boilers or in a
special control room in the case of power plant boilers.
Experience has proved that the operation of Ignifluid fired
boilers involves no particular problem and may be entrusted to any person
with normal qualification.
-------
- 4 -
The Ignifluid firing method insures a good flexibility as
the rating may vary between MCR and 15 to 20 % of MCR.
Now about the type of coal to be used any smalls from minus 1/8" to minus
3/4" may be burned ; this from anthracite to bituminous, coking or not, re-
gardless of the ash content .wjjftn below 50%. Of course such versatility helps
standardising. This is v. , :a due to the fact that the fluid bed actually
consists only of "red hot particles of coke" regardless of the nature of the
coal.
To show how near we approach standardizing I may give the
following example : In August 1966 the CERCHAR carried out a fifteen day
sery of combustion tests in a 50/55 t/h Babcock-Atlantique boiler equipped
with Ignifluid furnace at La Rochette (France).
- Three types of coal were burned successively without any stop for
adjustment and without switching off at any time the automatic control..:
total carbon content
hydrogen
nitrogen
sulphur
G.C.V. on dry
fusibility
oxygen
ash content on dry
volatile matter
anthracite
from
Dauphine
0/3 mm
73.5
1.4
0.6
1.1
6210
1210
21.6
5.8
0/6 treated low
volatile smalls
from Northern
French collieries
86
3.6
1.1
.8
8065
1450
2.2
6.6
7.8
0/1 8 bituminous
coking coal
from Freyming
78.5
5.2
1
0.9
7767
1240
9.4
4.3
36.7
-------
I will simply extract from the official report of these tests
the following short sentences : "The performance of the boiler plant func-
tioning at a rate close to its normal capacity are very satisfactory regard-
less of the fuel used..." "... the thermal efficiency is around 88 %...".
A year later, the firm operating the same 50/55 t/h boiler
at La Rochette gave the following data concerning availability which is
also a very important factor :
The percentage of time of interruptions noted from September 1966 for
twelve months onwards is :
. 3 % because of the stoker alone
1.84 % because of other difficulty
-2.14 altogether.
Efficiency is another important point on which I think it is
of interest to give a few figures.
For this I will rather refer to losses of efficiency on net
calorific value :
- losses due to riddlings under the grate are 0,25 to 0,5 %
- chimney losses due to unburnt gas when excess air for com-
bustion is over 20 % are nil
- chimney losses due to flue dust carry over when using carbon
with less than 20 % ash is less than 0,2 %
- heat losses due to unburnt combustible in the slag varies
with the ash content :
-------
- 6 -
for a 19 % .. ash coa^ the losses are 1.40 %
ii M o c o/ ii ii n it ii o cr» o/
o o ^ — / o. D u ^
Of course apart from the above losses which are dependent
on the process, one must take into consideration other unavoidable losses
due to sensible heat of flue gas, such losses should be calculated on the
bas. of 25 % excess air.
As efficiency results I may quote for the control of a 50/55 t/h
boiler burning anthracite smalls containing 19 % ash the following specifications
was 83 % on LCV but official tests have proved 88,7 on LCV.
For more important boiler plants I may simply say that the guaranteed
efficiency gitoen with Ignifluid firing compares favorably with that of pul-
verized firing.
As an example I will cite that the guarantees given for the 60 MW twin
boilers of Casablanca burning anthracite with 18 % ash is : 87.96 on GCV
dry coal. We hope to exceed this figure when controlling the efficiency .
As another example, the guarantees given for 250 MW twin boilers burning
anthracite of another type with 37 % ash, 8 % moisture and 5 % volatile
matter as received is : 88,83 on GCV.
I would like to mention that it is economically feasible to burn coals having
up to 50 % ash because there is no expenses incurred w'hatever with pul-
verizing while this would probably not be the case with pulverized firing.
-------
Now about burning high ash content anthracite coal
I want to seize this occasion to explain how unexpectedly we have become
involved into a scheme studied in Pennsylvania to solve a very important
pollution problem.
Large aeras of the Susquehanna and Lackawanna valleys
encumbered by silt or refuse banks with some on fire, suffer from air pol-
lution by sulphurous smoke or water pollution by acid drainage. Last
month a Pennsylvanian newspaper published the following facts :
it is now being considered to use a 250 MW Ignifluid power plant to con-
sume during the next 30 years 65 million tons of anthracite refuses thus
restoring land for economic development.
The Pennsylvanian authorities are also reported.to be con-
sidering that the same Ignifluid fired power plant could deliver low pressure
steam to a large demineralizing plant of the Westinghouse type ; to clean up
mine drainage in the aera and produce pure water.
If this project becomes reality it will be a proof of the
virtues of fluidisation in solving most unexpected pollution problems.
-------
Session III
Pollution Control. 1
Control by Means of Corabustor Design
and Operating Factors
Wednesday evening, November 20
Discussion of the control of particulates, hydrocarbons,
SO and NO by adjustment of fluid bed combustoi- design-
ana operating variables. . ,
-------
The Emission of Chlorine and Oxides of Sulphur
and Nitrogen from Fluidised Combustion Beds
by
S.J. Wright
BCUEA Industrial Laboratories,
Leatherhead, Surrey, England.
Presented during Session III
First International Conference on Fluidised Bed Combustion
Hueston Woods, State Park, Ohio,
November 18 - 22, 1968
Experiments have been conducted burning %"-0 coal without additives
on the pilot-scale plant at the BCUEA, in which complete chemical analyses
were performed on the solids input to and output from the combustor,
whilst the flue gases were analysed for S0_, SO , NO and Cl. Two coals
were used, .one containing 20$ ash. and having relatively high chlorine and
alkali contents and one containing ~55% ash and having a relatively high
calcium content. The range of the combustion conditions covered was -
2
heat inputs from 1.50 to jj»2^ MW/m (based on distributor area); excess air
percentages from zero to 80^S; mean bed temperatures from 825 C to 900 C '
and percentages of elutriated solids recycled to the fluidised bed from
zero to 809?.
The results showed that chlorine passed into the flue gases in the
same proportion as did the carbon suggesting that, for 1OO# combustion
efficiency, all the chlorine would be found in the flue gases, probably
as EC1.
-------
Analyses of the solid streams entering and leaving the combustor
showed that between 5% and 10% of the sulphur was retained in the solids
along with between 80% ajid 90% of the alkalis. The retention was •
apparently independent of combustion conditions with the exception of
the percentage.of elutriated solids recycled to the bed. An increase
in the percentage recycle tended to decrease the retention of sulphur.
The sulphur was present in the flue gases mainly as S0? but with
traces of SO,.
Oxides of nitrogen were not measured during all experiments but,
2
for a heat input of 2.7 MW/m at 18% excess air and a mean bed
temperature of 880 C, from ^00 to k$Q ppm (v/v) of NO were measured
at the cyclone inlet. . . " . .'. '. . - •
-------
SULFUR-OXIDE CONTROL IN FLUIDIZED BEDS
by
William T. Reid
Battelle Memorial Institute
Columbus Laboratories
Columbus, Ohio
presented during Session III
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
Emission of sulfur oxides from fluidized combustion systems is controlled
by two major factors: (1) temperature, which generally is lower than in other
methods of burning solid fuels, and (2) excess air, which can be kept less than
v;hen coal is burned on grates, in suspension, or in pulverized form. Other fac-
tors, such as reactive materials that can be added to a .fluidized bed to "capture
sulfur oxides, will be treat-ed by others and will not be considered here.
It is important to recognize that all the sulfur initially in the fuel
will be converted to sulfur oxides in the flue gas except for the minor amount
caught by CaO and MgO present in the cOal ash. Typically, the CaO + MgO content
of Eastern bituminous coals is less than 5 percent; in Midwestern coals they may
total 20 percent. If, in each case, the coal is assumed to contain 10 percent
ash and 4 percent sulfur, then the CaO + MgO in the ash could capture at most
about one-tenth of the sulfur as CaS04 and MgS04 for the Eastern bituminous coals,
and somewhat more than a third for the Midwestern coals. The rest of the sulfur
would appear in the flue gas. Whether even these small fractions of the sulfur
would be caught by the ash depends almost entirely on the maximum temperature
reached by the system.
-------
-2-
Temperature
The mineral matter in coal reacts in a series of steps as the coal is
heated. Most important here is the fact that S02 and S03 are evolved when the
sulfur-containing minerals in coal are heated in the range 500 F to 1600 F. In
the case of FeS, reactions begin at about 1600 F, and generally are completed by
the" time the ash reaches 2000 F. Analyses of coal-ash slags, formed in the tem-
perature range of 1800 F to 3000 F, almost invariably show that these slags con-
tain no more than 0.1 percent sulfur oxides, confirming that all the sulfur in
the coal has been rejected by the silicate melt. Organic sulfur in coal begins
to be evolved at about 500 F, and is completely converted into sulfur oxides
when the combustion process is finished.
Many factors affect this evolution of sulfur oxides as a result of the
formation of silicate melts, such as the composition of the mineral matter, the
size of their particles, their distribution in the coal, the rate at which heat-
ing occurs, mixing of the inorganic matter while combustion is occurring, the
maximum temperature reached and how long this temperature is maintained, the gas
composition surrounding the surface of the burning particles of coal, and finally
the total time-temperature history of the ash in the combustion system. Much too
involved for analytical considerations - and indeed, few data are available on
most of these factors - experience mainly has taught furnace designers how to
adapt large pulverized-coal-fired boiler furnaces to the problems of mineral
interactions.
The same knowledge probably does not exist at present for the conditions
existing in fluidized beds. Ash reactions, and hence the factors influencing the
emission of sulfur oxides, needs much further study under the special conditions
existing in moderately low-temperature combustors.
-------
-3-
It is worthwhile to review the influence of temperature on the reaction
between sulfur oxides and lime and magnesia. Based on thermodynamic calculations,
and ignoring the problem of kinetics, the equilibrium concentration of S03 in flue
gas in contact with MgO will be about 2500 ppm at 1520 F. For CaO, 2500 ppm S02
is in equilibrium with the solid at 2180 F. Hence, at these temperatures, the
normal S0a content of flue gas will not be affected by the presence of MgO or
CaO; the temperature must be appreciably lower for these oxides to capture S0a.
To reduce the S02 by one half, the temperature at equilibrium cannot exceed 1480 F
with MgO, or 2150 F with CaO. To reduce the S0a by ten times, to 250 ppm, the
corresponding temperatures are 1420 F for MgO and 2060 F for CaO. Based on these
considerations, it is evident that a fluidized bed operating at 1800 F will not
remove any S02 by reaction with MgO that may be present, but it may remove all
but 2 ppc of the S02 with CaO present in the bed, assuming equilibrium conditions.
It is evident, then, that the temperature of the fluidized bed is an
important variable if reaction between added limestone or dolomite and SOg is ex-
pected. Under conditions usually specified for fluidized bed combustors, low
enough to minimize slagging reactions, lime theoretically offers complete removal
of S02 while magnesia would be ineffective. •
• i .
Excess Air .
The most outstanding development over the past decade in the burning of
residual fuel oil in large central-station power plants has been the realization
that combustion with exceedingly low excess air prevents the formation of flame-
produced S03. Extensive research has shown that S03 is formed.in flames by the
reaction between S02 and oxygen atoms resulting from complex flame reactions.
Once the flame reactions are completed that produce oxygen atoms, no more S03 is
formed. In normal boiler furnaces, about one percent of the S03 is converted to
-------
-4-
S03 in this manner. That S03 is extraordinarily reactive. Experiments at Battelle
35
studying external corrosion of superheaters and reheaters using S as a tracer
have shown that the 30 ppm of S03 normally present is hundreds of times more re-
active than 2500 ppm of S02 when Na2S04 and Fe303 are available and ample oxygen
is supplied.
Low excess air as a means of eliminating S03 in flue gas was investi-
gated earliest in England, beginning about 1956. The art was advanced markedly
in Germany in 1960. It is now firmly established that limiting the oxygen in
flue gas to 0.2 percent (one percent excess air in place of the 15 to 20 percent
commonly used) essentially eliminates all problems with S03, both in corrosion
and in air pollution. Although low excess air has no effect on the S02 content
--••'
of flue gas, it can essentially eliminate acid smuts that can be a major source
of air.pollution. Oxygen admitted after the flame reactions are completed, and
where the gas temperature does not exceed about 1750 F, does not oxidize S0a to
S03 homogeneously, although catalytic surfaces will increase the S03 level if
oxygen becomes available through leakage as the flue gas moves through a boiler.
It is important to recognize that the excess air must approach zero
for its benefits to be obtained. For example, based on laboratory measurements
made at Battelle in a noncatalytic combustor, it was shown that at 2 percent ex-
cess oxygen, the S03 level when burning a 5.5-percent-sulfur liquid fuel was
about 30 ppm. At 1 percent excess oxygen it was 25 ppm or no significant decrease,
but at 0.1 percent excess oxygen the S03 was as low as 2 ppm. With stoichiometric
combustion, or under mildly reducing conditions, the S03 level was zero.
As a means of controlling air pollution from S03, but mainly as a method
of decreasing the dewpoint of flue gases to increase efficiency without corroding
preheaters, low excess air is being given a great deal of attention today. All
of the oil-burning boiler furnaces of the CEGB in England are now operating with
low excess air, although in some cases the oxygen is as high as 1 percent. Low
-------
-5-
excess air has not been taken over so completely in the United States largely be-
cause of the problems with controls to prevent smoke. In the cases in this country
v/here it has been successfully applied, low excess air has been a boon indeed.
So far, the usefulness of low excess air has been limited to oil-fired
equipment. Pulverized coal in its present size consist, and using existing types
of burners, cannot be operated at anywhere near the necessary low levels of oxygen
without leading to excessive losses of unburned fuel. One of the suggested advan-
tages of combustion in fluidized beds is that solid fuels can be burned effici-
ently with essentially no excess.oxygen. If that expectation can be sustained,
the ability to eliminate combustion-produced S03 could well be one advantage of
fluidized-bed combustors perhaps not now generally recognized.
A point worth noting is that the percentage of sulfur oxides occurring
as S03 at equilibrium increases as the temperature is lower. Specifically, at
0.2 percent excess oxygen, about 0.4 percent of the total sulfur oxides will be
S03 at 2000 F, about 0.8 percent at 1800 F, about 2 percent at 1600 F, and about
6 percent at 1400 F. At 1000 F, two-thirds of the sulfur oxides are S03 at this
oxygen level. The significance here is that the S03 fraction increases as the
temperature is lowered, but at low excess air the total amount of S03 is moderate
at the temperature where fluidized beds would operate.
WTRrebk
11-8-68
-------
ABSTRACT
POLLUTION CONTROL BY MEANS OF COMBUSTOR DESIGN
AND OPERATING FACTORS
by
E. B. Robison
Pope, Evans and Robbins
Alexandria, Virginia
Presented During Session III
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
-------
ABSTRACT
POLLUTION CONTROL BY MEANS OF COMBUSTOR DESIGN
AND OPERATING FACTORS
The research program in fluidized bed combustion, conducted
by Pope, Evans and Robbins for the National Air Pollution
Control Administration, has two primary objectives. The
first is to characterize the pollutant emissions from the
fluidized combustion of coal and the second to investigate
the control of emissions, notably sulfur oxides, by addition
of dolomite and limestone. The discussion to follow is
limited to the characteristic emissions observed in recent
tests and to the possibility of emission reduction by design
change.
Emission tests have been conducted in two units, one a pilot
scale fluidized bed column, designated the FBC, and the other
a full scale "boiler module referred to as the FBM. Samples
of the flue gas from the units are analyzed continuously for
hydrocarbons, nitric oxide and sulfur dioxide concentrations.
Flame ionization and infrared recording instrumentation is
used with periodic sampling for wet tests.
Early in the development of the fluidized bed boiler it was
learned that a bituminous coal could be burned in a fluidized
bed at near"stoichiometric air flow without smoke in the flue
gas. When emission testing was begun, this operating condition
was found to produce a high concentration of hydrocarbons in the
-------
'flue gas. Increase in the oxygen concentration, however,
sharply reduced the hydrocarbons emission. Data are pre-
sented to show that hydrocarbons emission from the FBC is
reduced to near zero with 3% oxygen in the flue gas. A
concentration of 4% appears to be necessary for the FBM.
These Vcilues correspond to 17% and 24% excess air for the
coal being used. Carbon monoxide emission is not a problem.
Emission of hydrocarbons during the transient light off
;
period is easily controlled .with excess air. No appreciable
effect v/as observed with change in bed temperature and bed
"•'
height. An increase with water injection above the bed
indicates that the final concentrations are not established
in the bed. Emissions with less volatile coals have yet
to be determined.
The fact that a combustion fluidized bed could be operated
at a temperature below that of conventional coal firing
systems held promise of significant reduction in nitrogen
oxides emission. Rapid heat loss from the bed permits
operation in the range of 1500°F to 1900°F. Emission tests
show that nitric oxide concentrations vary in the range of
300 to 400 ppm depending primarily on the oxygen content in
the flue gas. These values roughly correspond to values of
1 to 3% oxygen concentration.
-------
3
Unexpectedly, the nitric oxide concentrations were found
to be independent of "measured" bed temperature. The
term "measured" is used because the analysis indicates the
presence of local high temperatures, possibly at the bottom
of the bed where oxygen concentration is highest. Data are
presented to show the theoretical equilibrium concentrations
of nitric oxide formed with 20% oxygen (at the bottom of
the bed) and 3% oxygen (at the top). The measured NO concen-
trations suggest that even though the bed is operated at a
measured temperature of 1500°F, a minimum of 1800°F exists
somewhere in the bed. Failure of the NO concentration to
readjust to a lower theoretical value at 1500°F can be ex-
plained by slow reaction rate at the low concentrations
relative to the system residence time.
Equilibrium constant data for the oxidation of nitric oxide
to nitrogen dioxide show that if time permitted, all the
NO would be oxidized in the flue passages. The rate of this
termolecular reaction is very slow, however. It is esti-
mated that in the 4 second transit time of the gas thru the
entire FBC system the conversion is less than 0.1%. This
situation is unfortunate since present removal techniques
require its oxidation.
-------
In the matter of sulfur oxides emission there appears to
be no advantage to fluidized bed firing over conventional
methods when sulfur capture additives are not employed.
Most of the sulfur in the coal appears.in the flue gas.
Nitric oxide emission might be reduced by cooling the
bottom of the bed with water cooled projections. A scheme
is described.
Particulate emission from the fluidized bed boiler is a
function of the efficiency of.the ash collection system
since most of the ash in a coal is elutriated from" the'bed.
From the results of tests conducted thus far it would
\
appear the fluidized bed boiler will be operated with 4%
oxygen in the flue gas to limit hydrocarbons emission at a
slight penalty in nitric oxide emission. This oxygen
content is less than the 7-8% normally found in conventional
boilers. For this reason emission of nitric oxide should
be less for the fluid bed boiler even though concentrations
are comparable.
Heat release rates with the fluidized bed combustor opera-
ting at a 14 fps superficial, velocity are higher than have
ever been achieved by any system yet devised for the combus-
tion of coal. At the same time a fair approach to stoichio-
metric combustion is achieved. It is not unreasonable to
-: . Er,\>rAT'-3.—• A 'NTT"-', C? O-p-J pp, T fs.]
-------
expect that a more perfect combustion with less oxygen
and less hydrocarbon emission could be attained with some
sacrifice in the heat release rate. Local high tempera-
tures , which are probably diffusion controlled, might be
reduced and the nitric oxide in proportion.
-------
Session IV
Pollution Control. 2
Control by Means of Additive Injection
Thursday, November 21
Discussion of additive selection, kinetics studies, and
injection tests. Economics of control.
-------
Reduction of Atmospheric Pollution by the
Application of Fluidized-Bed Combustion
by
A. A. Jonke, R. L. Jarry and E. L. Carls
Argonne National Laboratory
9700 South Cass Avenue
Argonne, Illinois 60439
Presented During Session IV
First International Conference on Fluidized-Bed Combustion
Hues ton Woods State Park, Ohio
November 18-22, 1968
A study of fluidized-bed combustion as a means of reducing the
quantity of atmospheric pollutants (oxides of sulfur and nitrogen) released
during the combustion of fossil fuels is under way at Argonne National
Laboratory. Initial emphasis will be on fundamental studies with the
primary objective of optimizing the pollution control aspects of fluid-
bed combustion. Fluid-bed materials and additives that react with
sulfur (and possibly nitrogen) compounds released during the combustion
process will be studied. The program will involve laboratory-scale ex-
perimental studies to evaluate additives, bench-scale fluid-bed tests
to study the effects of operating variables, and appropriate evaluations
and assessments of technology relevant to pollution control and fluid-
bed combustion.
Exploratory studies of a variety of alternative schemes for
fluidized, pollution free combustion will be made. For example, the
fixation of sulfur in a form which will allow recovery of a sulfur
value directly or via a regeneration cycle appears to be an economically
-------
- 2 -
sound and desirable objective. This might be more readily accomplished
if a sulfide rather than a sulfate is formed. The operation of the
fluidized-bed combustor under partial reducing conditions might result
in the formation of a sulfide product. Such a scheme of operation
could utilize a multistage fluid-bed combustor with counter-current
flow of gas and solids to increase the efficiency of SO- capture and
additive utilization.
The initial laboratory-scale experiments for the evaluation of
potential additives for sulfur dioxide capture in a fluidized-bed com-
bustor have been performed. (Most of the previous work of this type done
by others has been at temperatures lower than those of interest in
fluidized-bed combustion.) Preliminary screening for SO- absorption
capacity has been done for the following materials: phosphate rock,
spent oil shale, "red mud", a calcareous shale, the oxides of manganese,
copper, cobalt, lead, zinc, and nickel, and a dolomite (BCR-1337).
The experiments were performed at 1700°F using a gas phase containing
5000 ppm S0? in air at a superficial gas velocity of .0.05 ft/sec. The
elapsed times were noted for the first breakthrough of SO^ from the
additive bed and for breakthrough at an S0_ content in the effluent gas
stream of 20% of the original concentration, as measured by a thermal
conductivity cell. The sorption capacity, expressed as g S02/100 g
additive, ranged from 0.3 to 2 and from 0.5 to 3.2 at the 100% and 80%
capture levels, respectively, for most of the additives tested other
than the dolomite. Zinc oxide and phosphate rock had no measurable
-------
- 3 -
sorption. In comparison, the dolomite (BCR-1337) had sorption capacities
of 5.5 and 14.8 g SC>2/100 g dolomite at the 100% and 80% capture levels,
respectively.
These results indicate that additives other than dolomite are
probably not desirable choices since they did not approach the capacity
of dolomite. However, the possibility of easier regeneration of some
of the lower capacity sorbents will be evaluated, and some of the
potential additives might be preferred on this basis.
A model has been developed which relates the extent of S09 capture
in a fluidized-bed combustor to the stoichioinetric equivalents of CaO
added as limestone or dolomite, taking into account the reaction parameters;
bed depth, superficial gas-velocity, and particle size. The reaction rate
data selected for use were derived from data generated in the NAPCA in-
house program. Two hypothetical cases of gas mixing in a fluidized bed
were considered—no gas mixing and perfect gas mixing. The fractions
of the S0? captured are given by the expressions 1 - e and r/1 + r
for no gas mixing and perfect gas mixing, respectively. For the
dolomite BCR-1337 at 1600°F, the parameter _r is given by the following
expression
. R f L
r = 1.04 x 10 —
where,
R is the reaction rate, Ib SO^/lb calcined stone-min
f is the fraction of the bed as initial calcined stone
L is the bed height, ft
v is the superficial gas velocity, ft/sec.
-------
- 4 -
Using the equation derived from the model and the rate data for
the dolomite BCR-1337 from the NAPCA study, a computer program was
written. The inputs for the computer program were particle size (96,
282, 507, and 1095 pm); superficial gas velocity (1, 2, 4, 6, 10, and
14 ft/sec); and bed height (0.5, 1, and 2 ft).
For all of the cases considered, the percentage of SO- removal is
greater for the assumption of no gas mixing than for the assumption of
perfect gas mixing. The spread between the two values increases with
greater gas velocity, larger particle size of the additive, decreased
bed depth, and fewer stoichiometric equivalents of calcined stone
added. The effects of the variables were analyzed, using average
values of the no gas mixing and perfect gas mixing cases. These cal-
culations, in effect, quantified the expected relationships between the
extent of S0_ removal and the operating parameters of a fluid-bed
combustor. Results indicate that S0_ absorption will be enhanced by
decreasing the additive particle size and superficial gas velocity and
by increasing the bed depth.
The following examples indicate the effect of particle size on
the degree of S0~ removal from the gas stream. To achieve 90% removal
of SO at a superficial gas velocity of 1 ft/sec in a 1-ft deep bed
would require little more than one stoichiometric addition of CaO
(as BCR-1337 dolomite) for a particle size of 282 um, but would require
about 1.7 stoichiometric additions .for a 1095 ym particle size. At a
gas residence time of two seconds or less and a S0? removal level of
90%, the maximum fractional CaO utilizations that can be expected in
-------
- 5 -
the fluid-bed are: 0.99, 0.90, 0.75, and 0.60 for particle sizes of
96, 282, 507, and 1095 urn, respectively.
An extension of the above correlation will be made, using NAPCA-
generated reaction data for BCR-1360, a calcitic limestone. Comparison
of the results of pilot plant tests of other workers with the model will
also be attempted.
Argonne bench-scale experiments on the control of the emission of
S0_ during the fluidized-bed combustion of coal will employ a 6-inch
diameter stainless steel coal combustor. The equipment has been designed
for operation over a relatively wide range of conditions, whether within
or without the range of practicality for full-scale systems, so that the
effects of variables such as bed height, superficial gas velocity, and
particle size of the additive can be thoroughly evaluated. A schematic
diagram of the reactor is shown in the accompanying figure. Spaced
vertically over the bottom two feet of the unit are four annular
chambers (each 2 1/2 inches high) through which a mixture of air and
water can be circulated to effect heat removal in each zone. This
capability will allow for varying the heat removal of the unit so that
it can be operated over a range of coal feed rates (directly related
to fluidizing air velocity) and a range of bed depths.
The fluidizing air, after passage through an electrically heated,
stainless steel ball-packed preheater, will enter the reactor through
a bubble cap type gas distributor. Coal and additive will be fed by
volumetric screw feeders and will be entrained in an air stream for
introduction into the fluidized bed. The off-gas from the unit will
-------
ANL Fluidlzcd-Bcd Cotnbustor
Viev;ing Port
Air-Water Coolant Inlet
Auxiliary Inlet
View Port, or
Overflow Solids
Take-Off
Air-Water
Coolant Inlet
Coal and Additive
Feed Point
Combustion
Air Inlet
Vent Gas to
Cyclone Separators
Openings for
Thermocouples,
Pressure Taps,
and Solids Sampler
Solids Bottom Take-Off
NOT TO SCALE
-------
— 7 —
be passed through cyclone separators to separate most of the solids
before final filtration. Provisions have been made for recycle of
entrained solids to the fluid-bed reactor. A constant bed depth can
be maintained either by withdrawal of solids from the bottom or by
overflow from the top of the bed. The unit will be capable of
operation at superficial gas velocities up to 14 ft/sec and bed depths
up to 2 ft. Continuous analysis of the off-gas will be provided, as
well as intermittent analysis of solid samples withdrawn from the bed.
-------
NAPCA's Dry Limestone Injection Program
E. A. Zawadzki
PCEP, NAPCA, DREW
5710 Wooster Pike, Cincinnati, Ohio 45227
presented during Session IV
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
-------
The National Air Pollution Control Administration through the
Process Control Engineering Program has initiated a comprehensive
research effort consisting of bench-, pilot-, and full-scale testing
of the dry limestone injection process. This process consists of
injection of limestone or dolomite into the combustion zone of a boiler
at a point above the burners. The limestone is calcined and subse-
quently reacts with S02 and oxygen present in the flue gas to form
calcium sulfate. Calcium sulfate and unreacted lime are removed with
the particulates in existing dust collection equipment. Many factors
influence the efficiency of the process. It is the purpose of this
study to resolve these problem areas.
A summary of the activities which NAPCA is sponsoring in the
area of dry limestone injection is as follows:
a) Process Control Engineering Program - In-house research.
This program deals primarily with the characterization of a
wide range of U. S. limestone and dolomites. Samples are
screened into the size fractions and subjected to tests in
a fixed bed reactor, in which the capacity of the stone,
previously calcined at a standard conditions, is measured
synthetic flue gases are passed through the fixed bed,
breakthrough-time and capacity are measured. The other test
-------
is an attempt to measure the kinetics of the reaction of
S0« and lime over lime and magnesia in a differential
reactor. Activation energies and reaction rates are ob-
tained. In support of this activity, measurements are made
by American Instrument Company for PCEE on the pore size
distribution and surface area of the limestones and dolomites
tested.
b) Battelle Memorial Institute under contract from NAPCA has
initiated a program to study the kinetics of the reactions
of alkaline earth oxides, hydrates, carbonates in^a dispersed
phase contact reactor in which temperature and residence time
_. _ . can be accurately controlled. In support of this activity
Battelle has also examined a series of test stones in a
differential thermal analyser to obtain data on the mechanism
of the reaction of SO with limestone.
c) Babcock and Wilcox Research has undertaken a study of the
measurement of the capacity of the limestones and dolomites
to absorb S0_ using a small coal combustion unit. B and W
is examining not only the capacity of the stones for reaction,
but also obtaining data on the degree of interaction of the
stone with fly ash. An attempt is being made to evaluate the
effects of temperature and degree of dead burning of the stone
and as a supplementary program attempting to obtain information
-------
3
on the resistivity of the limestone modified fly ash both
at the test site and in the lab. THis work is being done
by Research Cottrell on a subcontract to B and W.
d) Esso Research and Engineering Corporation at Linden, New
Jersey, has been studying the application of fluidized bed
techniques through the development of a dry limestone injec-
tion process for controlling sulfur dioxide. In addition,
they are looking at the possibility of regenerating calcium
sulfate for reuse in the. fluidized bed with the additional
objective of recovery sulfur value from the spent limestone.
e) TVA Fundamental Research Laboratory at Muscle Shoals, Alabama,
is examining basic chemical and physical properties effecting
the reaction of SO^ with calcium and magnesium oxides, hydrates,
and carbonates. Emphasis is placed on attempting to establish
the mechanism of the reaction and in establishing the role of
various physical properties of limestones and dolomites which
contribute to the reaction.
f) The Coal Research Board, West Virginia University, Morgantown,
West Virginia, is examining the potential utilization of
limestones-modified fly ash and is also evaluating a process
for recovering unused lime from the limestone modified
fly ash.
-------
g) TVA through the Applied Research Branch Division of Chemical
Development has prepared a detailed conceptual design and cost
study of the dry limestone injection process as it is applied
to the removal of sulfur dioxide from power plant stack gas.
h) TVA Division of Power Production will conduct a full scale
field trial of the dry limestone injection process at Shawnee
power plant, Shawnee Unit //10. The study is to be conducted
in the fall of 1969. It will lead to the evaluation of the
process in the field and development of engineering and
operating data which might be used by others in applying
the process.
i) NAPCA is sponsoring a program at the Peabody Coal Company
at Columbia, Tennessee, to obtain test data on dry limestone
injection using a - pounds per coal fired per hour moving
bed stoker. The object of this program is to obtain informa-
tion under controlled test conditions. The objective to
determine extent to which dry limestone injection can be
applied to small industrial stoker units.
•"»
Field tests conducted by PCEP and Florida Power Company have provided
samples of lime which have been subjected to furnace temperatures and
actual flue gas coal fired unit. These samples will be used to study
the "dead burning" characteristics of the stones.
-------
USE OF FLUIDIZED BEDS OF LIMESTONE BASED
MATERIALS FOR DESULFURIZING FLUE GAS
By
A. Skopp
Esso Research and Engineering Company
P. 0. Box 8, Linden, New Jersey 07036
Presented During Session IV
First International Conference on
Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
Among the many processes being investigated for the removal of
sulfur dioxide from power plant flue gas, the injection of finely sized
limestone or dolomite into the furnace offers the advantage of simplicity
and low initial capital charges. These advantages are partially offset,
however, by the apparent inability of injection processes to provide high
sorbent utilization and complete flue gas desulfurization. The reason
for this is that in the short time available to the adsorbent for the
reaction in a boiler, only the oxide at or near the surface can sulfate.
Figure 1 shows the range of results which have been obtained with injection
processes.
As part of its program to study ways of improving sorbent utiliza-
tion, the National Air Pollution Control Administration contracted a pro-
gram with the Esso Research and Engineering Company to investigate the
use of a fluidized bed reactor for contacting flue gas with limestone
based materials. Two sorbents were studied extensively in this program,
a limestone BCR 1359* and a dolomite BCR 1337*. These materials were
evaluated in both a coarse and a fine solids fluid bed.
* Bituminous Coal Research, Inc. designations
-------
- 2 -
Using the coarsely ground limestone and dolomite, high sorbent
utilization could not be obtained, even when the sorbents were calcined
under pressure-induced conditions chosen to produce porous materials.
With the large particles, only the oxides at or near the surface reacted
rapidly. Further reaction took place at reduced rate and for the limestone
was in fact practically non-existent (Figure 2) .
Additional studies with large particles were made using fluidizing
conditions selected to promote attrition of the outer sulfated surface as
it formed. It was felt that this might provide a continuous method of
removing this surface, thereby exposing reactive oxide for further reaction.
Particle attrition was promoted by using increased fluidization velocities
and deeper beds. As shown in Figure 3, such techniques provided only
limited success because the sulfated surface was found to be harder to
break up than the oxide core. Consequently, the fluidized particles
decrepitated in their entirety to produce fines of about the same composi-
tion as the coarse sorbent from which they had been formed.
To demonstrate high oxide utilization, the use of finely ground
sorbent was then evaluated in our program. As Figure 4 shows, the
dolomite when sized to about 100>y provided calcium oxide utilizations
of 7070 to 80% at conditions which provided nearly complete sulfur dioxide
removal. Using the fine particle results, commercial process designs
were established and economic analyses made. Figures 5A and 5B show that
a fine particle fluid-bed process is expensive (~$2 ton/coal) and requires
single reactor sizes well outside the present state of the art.
-------
- 3 -
Currently, the program is proceeding in two directions. First,
conditions are being investigated in which coarse particles or a combina-
tion of coarse and fine particles are used to achieve high sorbent utiliza-
tion. Second, the possibility of regenerating the partially sulfated
sorbent is being explored actively. Regeneration could be coupled with
either a direct boiler-injection process or with a fluid-bed reactor,
shown in Figures 6A and 6B. In a regenerative process, the sale of a by-
product such as H9SO, would affect part of the desulfurization costs.
Studies made thus far indicate a regenerative process to be
technically feasible. At temperatures near 2000°F, sulfated sorbent
can be reduced back to its oxide. This reduction is fast and with good
gas-solids contacting sulfur dioxide concentrations close to those pre-
dicted by equilibrium a^e. obtained (Figure 7) . The regenerated oxide can
re-adsorb sulfur dioxide although at a reduced capacity as shown in Figure
8. Current efforts are being directed at further defining the capacity
loss rate for the cycled sorbents and at determining the reasons for
this loss.
A. SKOPP/jmt
10/25/68
-------
FIGURE 1
DRY INJECTION PROCESS GIVES LOW LIMESTONE
UTILIZATION & INCOMPLETE
5 . 80
o
^ 60
^
'o
2" 40
CvJ
O
oo 20
0
GAS DESULFURIZATION
1
a 3
CaO + S02 + 1/2 0
_
srf\
^WAWi
/^
i
-
2
MOLES CaC03 IN.
2
"
— >
it
1
l>-
CaO + C02
-r
o
***
aSO^
^
^f^1
^-
3
JECTED/MOLE
630 1260
TONS/DAY -CaC03 FOR 800 MW
so2
1890
*>
-'-
•*"
^-*
••
r
' —
—
'
4
IN
GAS
1
2520
PLANT - 3% S
IN FUEL
68-11101
-------
FIGURE 2
SORPTION CHARACTERISTICS OF LARGE PARTICLES
IN A FLUID BED OPERATING WITHOUT ATTRITION
2700
2400
co 2100
LU
LU
1800
150°
- 1200
CM
O
CO
S 900
(X
Qu
600
300
0
INLET S02 CONC.
12-16 MESH
CALCINED LIMiESTONE
BCR 1359
(MOLE % CaO REACTED)
(5.4%)
(5.1%)
[28%)
T = 1600°F
12-16 MESH CALCINED
DOLOMITE BCR 1337
'(42%)
20% SO
BREAKTHROUGH
TIME ON ADSORPTION'
68-11103
-------
..FIGURE 3
INCREASING ATTRITION PROVIDES ONLY SMALL IMPROVEMENT
IN CAPACITY BECAUSE THE ATTRITION IS NON-SELECTIVE
c
I
O)
LLJ
H-
O
V
ACTUAL PARTICLE
BREAKUP
(14.8)
PREFERRED
BREAKUP
FORMATION RATE OF XFRESH CaO SURFACE REQUIRED
FOR STOICHIO
METRIC REACTION WITH INLET SO.
(5.5)
V = 7.0 FT/SEC
T = 1600°F
(%CaO REACTED)
0
12 18
BED HEIGHT - INCHES
24
30
-------
FIGURE 4
HIGH CaO UTILIZATION POSSIBLE WITH SMALL PARTICLES
100
<
M
O
ro
O
50
20
50
I
CALCINED DOLOMITE
AT 1600°F
_L
100 200 500 1000
AVERAGE PARTICLE DIAMETER, MICRONS
2000
68-11105
-------
FIGURE 5A
COSTS ARE HIGH FOR
FINE-PARTICLE FLUID-BED PROCESS
FLUIDIZING VEL.-FT/SEC
FLUID BED D1AMETER-FT
PARTICLE SIZE,M
INVESTMENT, $MM
OPER. COST, $/T COAL
90% LOAD FACTOR
60% LOAD FACTOR
800 MW PLANT - 3% S COAL
6
140
100
12
2.10
2.60
24 :
73
50-100
10
1.90
2.30
68-11106
-------
FIGURE 5B
COSTS ARE HIGH FOR FINE-PARTICLE FLUID-BED PROCESS
BECAUSE OF REACTOR SIZE & SOLIDS ENTRAPMENT
h-
U.
i
0£
LU
H-
LiJ
Qi
O
h-
O
<
LU
o:
500
200
100
50
1.0
2.0 4.0 6.0 10.0 20.0
SUPERFICIAL FLUIDIZING VELOCITY
500
200
40.0
CO
2
O
o;
o
LU
NJ
CO
LU
ioo y
68-11107
-------
FIGURE 6A
REGENERATIVE PROCESS WITH COARSE-PARTICLE FLUID BED
DESULFURIZED GAS
RETURNED TO FURNACE
FURNACE
TO
STACK •*
ADSORBER
REGENERATOR
~2000°F
ACID
PLANT
H2S04
MAKE-UP
SORBENT
FUEL
AIR
68-11110
-------
FIGURE 615
REGENERATIVE PROCESS WITH FINE PARTICLE INJECTION
FURNACE
DESULFURIZED GAS
TO STACK
FINELY SIZED
MAKE-UP SORBENT
GAS/SOLIDS
SEPARATOR
n
ACID
PLANT
REGENERATOR
~ ~2000°F
FUEL
AIR
68-11109
-------
TIGURE 7
EQUILIBRIUM GOVERNS FLUID BED REDUCTION OF CaSO
UJ
Z3
_J
LL-
u_
UJ
o;
o
h-
<
C£
UJ
2
UJ
O
UJ
CM
O
CO
UJ
_l
o
10
8
0
0
Q
PREDICTED
EQUIL.
REDUCING GAS
CO OR H0
20% 'C02 OR H20
70% N0
SUPERFICIAL FLUIDIZING VEL.,
0.8 TO 1.6 FT/SEC
1800 1900 2000 2100 2200
REGENERATION TEMPERATURE, °F
68-11111
-------
. FIGURE 8
RELATIVE CAPACITY OF CYCLED SORBENT
O
J-
O
<
CL
<
O
D_
Ctl
O
CO
o
I—
I
I
ADSORPTION —1600°F
REGENERATION ~2000°F
© ANHYDRITE (CaS04)
DOLOMITE
_L
I
J_
34567
ADSORPTION CYCLE
8
68-11112
-------
A ROLE FOR FLUIDIZED COMBUSTION IN CLEAN POWER SYSTEMS
Arthxir M. Squires
Department of Chemical Engineering
The City College of The City University of New York
New York, New York 10031
presented during Session IV
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
This paper examines a steam cycle believed to deserve consideration
in equipment cost studies for "clean power systems" which utilize low-sulfur
coke in fluidized combustion. Also, an improved processing scheme is
described to provide such a coke. Heat rates are given for a complete clean
power system incorporating the proposed features and using cycle parameters
felt to represent reasonable near-term objectives.
Introduction: An Argument for Higher Steam Temperatures
With development of the fluidized combustion technique, re-examination
of steam cycle parameters would appear to be timely. Most large recent
stations employ either 2,400 or 3, 500 psi steam at about 1000° F. These
steam conditions evolved following experience gained about a decade ago from
units designed for 4, 500 to 5, 000 psi and 1150 to 1200°F. In 1958 Sporn and
Fiala (1) wrote: "Studies of the effect of initial temperature indicated that
all operating temperatures above 1050 required extensive use of stainless
steel in steam generating tubes [and] in piping and turbine parts. In the
range of temperatures from 1050 to 1250 F the gains of increased performance
are not adequate to bear the heavy capital burden of the austenitic materials.
-------
. . . Temperatures higher than 1050 F cannot be economically justified
at this time." One gathers that troubles with both hot-side and cold-side
corrosion in the 4, 500 to 5, 000 psi experimental boilers led to the conclusion
that temperatures as high as 1250°F would be a practical impossibility in a
conventional boiler.
With fluidized combustion, steam temperatures appreciably above
1000°F come again into view. Higher tube-wall temperatures and a higher
heat flux across steam-superheating surface can be achieved -- reducing the
quantity of high-cost stainless steel tubing required -- yet the tubes are not
. / ' '
necessarily subjected to conditions which are dangerous for either hot-side
or cold-side corrosion. The complex, molten alkali-iron sulfates which are
responsible for hot-side corrosion at about 1200° F in a conventional boiler
cannot form at the higher temperatures suitable for fluidized combustion.
Cold-side corrosion should also be less of a problem, since metal temperatures
cannot rise unduly if deposits form within a tube. If the fluidized combustion is
conducted at an elevated pressure — at around 8 atmospheres, say --
a higher heat-transfer coefficient could be exploited in part to permit a still
higher heat flux and in part to reduce the bed temperature, giving greater
insurance against cold-side corrosion.
What is suggested is that equipment cost studies are now in order
to see whether or not the 1200°F and higher steam temperatures contemplated
around 1955 might today seem again an economic proposition. If so, the
improved heat rate which such temperatures can afford may provide an offset
against the cost of measures taken to eliminate sulfur oxides from boiler
stack gases.
-------
A Steam Cycle for a Clean Power System
Both Consolidation Coal Company (a subsidiary of Continental Oil Co.)
and the FMC Corporation (2, 3) have proposed schemes yielding a low-sulfur
coke for power-station use. Such a fuel is nearly ideal for fluidized combustion.
A case to consider in studies of equipment for clean power would be a fluid-bed
boiler fueled with such coke, operating at atmospheric pressure, and serving
a conventional steam cycle at 5,000 psi and 1200°F, say. Two factors, however,
argue the advisability of examining other cycle alternatives: (1) The idea of
"supercharging" the fluid-bed boiler is appealing and indeed may be necessary
to achieve attractive costs for high steam temperatures. (2) Requirements of
water purity are so stringent if 5, 000 psi steam is to be carried to 1200°F that
one may wish to examine cycles employing sub-critical steam pressures on
that account.
FIG. 1 A Steam Cycle to Exploit the Fluid-Bed Boiler's
Ability to Superheat Steam to High Temperatures
STEAM TURBINES
FLUE GAS
LOW-SULFUR
COKE
AIR
LOW-LEVEL HEAT RECOVERIES
-------
12.OO
IIOO
1000
u. qoo
o
8OO
UJ
700
< GOO
CC
HI BOO
a
S 400
UJ
{- 3OO
200
!OO
I.HI IBS. VAPORIZED
/ I.Q IB.TO CONDENSER V
i I i
' I I
1.0. 1.5 2.0 2.5 3.0 3.5
ABSOLUTE ENTROPY, BTU PER °R
FIG. 2.
Temperature-Entropy Diagram for the Steam Cycle of FIG. 1.
[ NB. Entropy is treated as an extensive variable, rather than
as an intensive one in the more usual manner. ]
Figure 1 illustrates diagrammatically a steam power cycle (4) adapted
for sub-critical steam pressure and well suited to receive heat from a fluid-bed
boiler operating at high pressure. The cycle is characterized by the addition
of an unusually large amount of superheat to the steam. Three high-temperature
steam-expansion turbines are provided. Steam from the third turbine is cooled
against water ahead of a fourth, low-temperature steam turbine, which exhausts
to a condenser. Figure 2 is a temperature-entropy diagram of the cycle for
the following conditions: 2,400 psia and 1200°F to the first turbine; two reheats
to 1200°F; 15 psia exhaust from the third turbine (at 767°F); 260°F entering
-------
the final turbine; and 100°F exhaust (at 7.89% wetness).
An advantage of .the cycle illustrated by Figures 1 and 2 is the fact that
the water receives an appreciable amount of low-level heat in addition to the
heat picked up regeneratively from the 15 psia third-turbine exhaust. This
can be understood when it is remembered that water has a higher heat capacity
than steam. This advantage of the cycle is important when combustion is
conducted at high pressure. Since air to such combustion is heated by
compression, cold air is not available for cooling flue gas to a low stack
temperature. The cycle of. Figures 1 and 2 provides cold boiler-feed-water
for heat exchange against flue gas without a penalty in efficiency.
The advantages of the cycle can be understood by comparing Figure 2
with Figure 3, which gives a temperature-entropy diagram for a conventional
cycle using 2,400 psia steam at 1000°F, one reheat to 1000°F, seven
regenerative boiler-feed-water heaters, and 100°F exhaust. The net cycle
efficiencies in Figures 2 and 3 are 47. 1 and 44. 7 per cent respectively
(net work divided by net heat intake to water). All heat taken into the
conventional cycle of Figure 3 is at a level above 477° F, and cold boiler-
feed-water can be provided for heat exchange against flue gas only with a
penalty in efficiency. If heat is furnished to the cycle below 477° F *•-
at expense of reduction in size of all except two coldest regenerative heaters
-- such low-level heat is converted to work at only 29. 2 per cent efficiency.
By contrast, the cycle of Figure 2 receives 10. 5% of its net heat intake at
levels below 477° F. The throttle steam flow in Figure 2 is more than 30%
less than that in Figure 3 for the same net work.
-------
I2OO
IIOO
IOOO
LL qOOr-
jaooh
5~ TOO -
< ©OO
DC
yj 5OO
CL
S 4O°
UJ
»- 3OO
200
too
1.455 IBS. VAPORIZED
.,
1.0 LB. TO CONDENSER \"
!
±
1.O I.5 2.O 2.5 3.O 3.5
ABSOLUTE ENTROPY, BTU PER °R
FIG. 3. Temperature-Entropy Diagram for a Conventional Steam Cycle.
The equipment of Figure 1 could operate in cooperation with a gas-turbine
cycle providing air and expanding flue gas. At 1600° F gas-turbine-inlet
temperature and 300° F stack temperature, a heat rate of about 7, 534 Btu
per kilowatt-hour sent out would appear attainable for a typical low-sulfur coke.
A Process for Converting Coal into Low-Sulfur Fuels
The aforementioned schemes proposed by Consolidation Coal and
FMC (2, 3) would provide a low-sulfur coke as a byproduct of operations
yielding products of higher value: in the case of Consolidation Coal, either
hydrogen or pipeline gas; in the case of FMC, a liquid feedstock for refining
into gasoline and other liquid products. Both schemes rely upon a lime-containing
-------
acceptor to promote the action of hydrogen in desulfurizing a char obtained
from a prior coal-carbonization step.
Figure 4 shows an improved procedure whereby coal is carbonized
and the products are desulfurized in a single unitary operation, having production
of low-sulfur fuels as a sole objective. The fuel-processing vessel depicted in
Figure 4 consists of three zones: (1) a coal "hydrocarbonizing" zone, in which
coal is converted into gaseous products and coke pellets roughly 1/4 inch in
diameter, say; (2) a desulfurizing zone, in which sulfur is removed from both
gaseous products and coke by action of hydrogen and a lime-containing acceptor;
and (3) a calcination zone, in which CaCO^ is decomposed and gaseous
carbonization products are partially burned with air to provide a lean fuel gas
containing hydrogen. The air flow is approximately 11 per cent of the
FIG. 4. - Processing Scheme to Convert Coal into Low-Sulfur Fuels
CALCINATION
LEAN FUEL GAS
(LOW IN SULFUfc)
cuus
SYSTEM
DESULFUR1ZJNG
ZONE'
ELEMENTAL
SULFUR
TO MARKET
COAL
AND COKS 6SLF-
'AGGLOMERATING ZONE
LOW-SULFUR
COKE PELLETS
RECYCLE OF PORTION
OF LEAN FUEL GAS
(CONTAINING HYDROGEN)
-------
stoichiometric for complete combustion of the coal. A portion of the lean
fuel gas is recycled to provide fluidizing gas to the coal hydrocarbonizing
zone. The lime-containing acceptor is much smaller in size than the coke
pellets -- the acceptor might be smaller than 40 mesh, say. Fluidizing-gas
velocity is much higher in the coal hydrocarbonizing zone than in the other
two zones, to prevent the acceptor from sinking into the former zone. The
desulfurizing zone contains an intermingling of coke pellets and acceptor.
Coke pellets circulate at a high rate between the hydrocarbonization and
desulfurising zones, the pellets tending to be ejected from the former zone
into the middle of the latter, and thereafter tending to sink downward near
the walls of the latter because its fluidizing-gas velocity is below that needed
to maintain a high concentration of coke pellets therein.
Finely ground coal is fed to the hydrocarbonizing zone, which operates
at 1400°F, say. A coal particle is heated almost instantaneously to bed
temperature, and almost at once, the coal is split into a gaseous fraction,
comprising mainly methane and hydrogen, and a sticky, semi-fluid residue.
The latter is "captured" by a coke pellet, sticking thereto to form a "smear"
upon the surface of the pellet. Further coking reactions, which occur in the
order of a second or so, transform the sticky smear into dry coke and cause
additional gases and vapors to evolve. Soon thereafter, the given coke pellet
finds itself in the desulfurizing zone, and the freshly-made coke is desulfurized
by action of hydrogen and acceptor. Thus, desulfurization is substantially
simultaneous with hydrocarbonization. Practically speaking, no "aging" of
the coke product occurs between its formation and its exposure to desulfurizing
conditions. Desulfurization should be rapid, because H2S formed by action
-------
of hydrogen upon the coke need not diffuse outward from within micropores
deep inside a char or coke structure, as in the Consolidation Coal and FMC
operations (3).
For operation at 21 atmospheres, a calcination zone temperature of
1740°F is adequate. Many lime-containing solids -- especially many natural
dolomites -- can be subjected to this temperature repeatedly without loss of
reactivity for desulfurization at 1400°F (5, 6). At this pressure, the
coal-treating capacity of the vessel in Figure 4 is such that one can visualize
a single vessel to serve 1, 000 megawatts of electricity-generating capacity.
At a fluidizing-gas velocity of 1. 5 feet per second, the calcination zone would
be 30 feet in diameter. Since only a shallow bed is needed, this zone might
be housed in a sphere communicating to a 20-foot desulfurizing zone below.
Calcium sulfide can be regenerated by action of steam and carbon
dioxide at about 1000°F to form a gas rich in hydrogen sulfide (7), which
is readily converted to sulfur.
A Clean Power System for Coal
Figure 5 illustrates a combination of power-generating equipment
with the coal-desulfurization process of Figure 4. In working up an example
of this combination, the steam cycle of Figures 1 and 2 was assumed. Coal
containing 3. 72 per cent sulfur (moisture-and-ash-free basis) was used. Air
can be compressed to about 8 atmospheres and flue gas can be expanded in
conventional gas-turbine equipment. A gas-turbine inlet temperature of
1600°F and a stack temperature of 300°F were taken. About 10. 5 per cent
of the air was further compressed and supplied to the coal-desulfurization
process. Lean fuel gas was reheated to 1300°F by indirect heat exchange
-------
FLUE GAS
TO STACK
CO* RECOVERY
C02TO
SULFUR
DESQRPTSON
REMOVAL
PUBS. 6AS
FLUID-BSD
LEAH FUEL GAS
COKE PELLETS
COAL DESULFURIZATION
FIG. 5. Power-Generating Equipment to Cooperate with the Scheme of FIG. 4.
and was expanded from about 21 atmospheres to a combustion at about 8
atmospheres.
For generation of electricity at a rate of 1, 000 megawatts, some
features of the plant are as follows:
Air flow = 6, 188, 000 pounds per hour, easily handled by 4 machines
Throttle steam flow = 3, 902, 000 pounds per hour
Coal feed = 272.25 tons per hour (m. a. f.)
Heat rate = 7, 887 Btu per kilowatt-hour of electricity sent out
(allowing 5% for losses and station auxiliaries)
Sulfur production =199 long tons per day
-------
TABLE 1.
Energy Balance for Clean Power System Producing Electricity and Sulfur Only
Basis: Higher heating value of coal (HHV) = 100.00
Net shaft work 45. 13
Heating value of sulfur 0. 93
Heat to cooling water 45. 34
Sensible heat to stack 4. 64
Latent heat to stack 3. 98
100.02
Heat rate (taking credit for sulfur and allowing 5% losses):
3,412. 75x (100. 00 - 0.93)/45. 13x0. 95 = 7, 887 Btu per kilowatt-hour
Table 1 presents an energy balance (before allowance for losses) and illustrates
the calculation of heat rate.
The coal-desulfurization process of Figure 4 might also be operated to
furnish baseload power from the combustion of the lean fuel gas and to supply
low-sulfur coke to other power stations at a distance. If four vessels of the
size already described are provided, the plant would have features as follows:
Electricity rate = 1,082 megawatts
Air flow = 7, 298, 000 pounds per hour
Throttle steam flow = 3, 714,000 pounds per hour
Coal feed = 1, 098 tons per hour (m. a. f.)
Heating value of coke product = 68. 5 per cent of coal
Heat rate = 9, Oil Btu per kilowatt-hour of electricity sent out
Sulfur production = 797 long tons per day
Table 2 presents an energy balance.
-------
TABLE 2.
Balance for System Producing Electricity, Sulfur, and Low-Sulfur Coke
Basis: Higher heating value of coal (HHV) = 100. 00
HHV of coke • 68.46
Net shaft work 12.20
Heating value of sulfur 0. 93
Heat to cooling water 13. 79
Sensible heat to stack 1. 76
Latent heat to stack 2. 87
100.02
Heat rate (taking credit for sulfur and coke and allowing 5% losses):
3,412. 75 x (100. 00 - 68. 46 - 0. 93}/12.20x0. 95 = 9,011 Btu per kilowatt -hour
If the coke product is shipped to stations producing electricity at the
aforementioned heat rate of 7, 534, the average heat rate for all of the
electricity produced from the coal is 7, 990.
Concluding Remark
Study is needed to evaluate the proposals of this paper. The author
fully recognizes that other steam cycle arrangements and parameters are
capable of providing the heat rates stated here. He also recognizes that the
cost to achieve such heat rates in steam cycle equipment, of whatever type,
may prove uneconomic, even with the help of fluidized combustion. It is too
early to conclude that "clean power" may also be cheaper power, but to the
author, this seems at least a sporting proposition.
-------
References
1. Philip Sporn and S. N. Fiala, "Evaluation of Supercritical Pressure Steam
Plants Based upon the First-Time Operating Experience at Philo",
World Power Conference, Montreal, September 1958, Paper 111 Bs/5.
j
j 2. F. W. Theodore, "Low Sulfur Boiler Fuel Using the Consol CO£ Acceptor
: Process: A Feasibility Study", Report from Consolidation Coal Co.
' to Office of Coal Research, November 1967, OCR Contract 14-01-
• 0001-415.
5
i 3. John F. Jones, Michael R. Schmid, Martin E. Sacks, Yung-chuan Chen,
: Charles A. Gray, and R. Tracy Eddinger, "Char Oil Energy
Development", Report from FMC Corporation to Office of Coal
Research, January 1967, OCR Contract 14-01-0001-235.
4. Arthur M. Squires, "Clean Fuel Power Cycles", ASME Paper 67 —
WA/PWR-3, November 1967.
5. Arthur M. Squires, "Cyclic Use of Calcined Dolomite to Desulfurize
Fuels Undergoing Gasification", Advances in Chemistry Series 69,
205-229, American Chemical Society, Washington, D. C., 1967.
6. George P. Curran, Carl E. Fink, and Everett Gorin, "CO2 Acceptor
Gasification Process: Studies of Acceptor Properties", ibid., 141-165.
7. Arthur M. Squires, "Processes for Desulfurizing .Fuels", U. S. Patent
3,402,998 (September 24, 1968).
-------
RETENTICIT OF SULPHUR BY LIMESTOKE
D. F. Williams
Rational Coal Board
Coal Research Establishment
Stoke Orchard, Glos., England
Presented during Session IV
First International Conference on Fluidised Bed Combustion
Hueston Woods, State Park, Ohio
November 18-22 „ 1968
Measurements have been made of the emission of sulphur as .
sulphur dioxide and trioxide, and its retention as sulphate in the ash,
secondary cyclone fines and dust, during the combustion of coal in a
6 inch diameter fluidised ted. The combustion conditions in these runs
were as follows: temperature, 800 C; fluidising velocity, 2 ft/sec;
bed height, 2 ft; excess air, 10 - 20$; primary cyclone fines
recycled to the bed.
It was found that the amount of sulphur retained depended on
the relative proportions of sulphur and carbonates in the coal. Thus,
Babbington coal contained 0.6% sulphur and its stoichiometric equivalent
of calcium and magnesium carbonates, and more than Uo$ of the sulphur
was retained, whereas Goldthorpe coal contained 2% sulphur and 0.5$
carbonate, and only 10 - 15$ of the sulphur was retained. Farmington
Ho. 9 coal contained a similar amount of sulphur to Goldthorpe coal
(2.3$), but more calcium carbonate (l.25$, as CO), and about 25$
of the sulphur was retained.
-------
; In further experiments limestone ground to the same size
I as the coal (minus l/l6 inch) was added to the Goldthorpe or
3 Farmington coal feed and was found to increase the retention of
i
I sulphur. Addition of about twice the stoichiometric quantity of
| lir^estone (12$ by weight) to Goldthorpe coal led to the retention
j of virtually all the sulphur, only about 10 p.p.m. remaining in
] the off-gas. On addition of a similar proportion of limestone to
j Farmington coal, 80$ of the sulphur was retained.
j It was noted that, following the addition of limestone,
\
I the sulphur content of the off-gas reached its new equilibrium
i level much more rapidly than the free lime content of the bed,
I which continued to increase for several hours. Comparison of the
] data from different runs also indicated that the proportion of free
1
line in the bed did not affect the sulphur content of the gas.
\ This suggested that most of this lime lay in the interior of each
particle and was not readily accessible. Accordingly, for a further
run the limestone was crushed to minus 120 B.S.S. mesh (12U microns)
before addition to the Goldthorpe coal, and it was found that less
than lg times the stoichiometric amount of limestone was required
to retain all the sulphur.
-------
The Control of Sulphur Emicsion by
adding Limestone to the Combustion Bed
by
S.J,Wright
BCUBA Industrial Laboratories
Leatherhead, Surrey, England
Presented during Session IV
First International Conference on Fluidised Bed Combustion
Hueston Woods, State I&rk, Ohio,
November 18 - 22, 1968,
Work on the addition of limestone to the combustion bed of the
BCUKA pilot-scale combustor is in its early stages and a. comprehensive
investigation will involve a variety of types and size ranges of
limestone and experinents of up to JO hours duration.
Preliminary results, however, suggest that, whilst the
addition of limestone improves the percentage retention of sulphur
fno t-S-.
in the solids, the reactions CaCo_ -r> CaO ~* CaSO^ are very inefficient
especially in the £"-0 bed, and it will be necessary -to ensure the nost
favourable conditions - high surface area and reactivity together with
limestone additions well in excess of stoichiometric - before a
satisfactory level of sulphur retention is obtained.
-------
OUTLINE
POLLUTION CONTROL BY MEANS OF ADDITIVE
INJECTION
by
E. B. Robison
Pope, Evans and Robbins
Alexandria, Virginia
Presented During Session IV
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
-------
POLLUTION CONTROL BY MEANS OF ADDITIVE INJECTION
A number of potential advantages of the fluidized combustion
process have been pointed out in the past for control of air
pollution by additives. Among these are the following:
(1) The random motion of the fluidized bed of inert granules
could provide an ideal environment for contacting additives
for sulfur capture from a coal flame.
(2) The fluidized bed can be operated in the 1500-1800°F tempera-
ture/ a range which some investigators have found to be optimum
for sulfur capture.
(3) Additional gas-solids contacting is provided by a dilute
phase of solids above the bed.
(4) The turbulence and rapid mass transfer in the fluidized bed
could provide for a uniform distribution of the solid for further
contacting in the flue passages.
(5) The random motion of the bed could possibly erode away a
sulfate product shell and continuously expose an unreacted surface
for further sulfur capture.
One major objective of the current research effort for the
National Air Pollution Control Administration is to determine
the effectiveness of adding limestone to the fluidized combus--
tion process for emission control.
Additive tests have been conducted in both the pilot scale
fluidixed bed column (FBC) and the full scale boiler module (FBM)
-------
with continuous monitoring of sulfur dioxide, hydrocarbons and
nitric oxide emissions. The additive used in the tests were
selected on the basis of reactivity tests conducted by NAPCA.
One additive was a 50-50 dolomite designated BCR 1337 and the
other a high calcium limestone BCR 1359. The additive particle
size was selected at -7 + 14 mesh for retention in a bed
material of -8 -f 16 mesh sintered coal ash.
The continuous monitoring of emissions permitted relatively
rapid change in the operating variables in the course of a
single test. The operating variables included the bed tempera-
ture, bed height, excess air, the additive and coal rateo,
coal type, ash recirculation rate and water injection. The
coal vised in tests conducted thus far has been a high sulfur
(4.5%) high volatile, unwashed coal. The coal and additive
were screw fed into a pneumatic line and injected into the
bottom of the bed.
The test results showed that sulfur capture by the BCR 1337
dolomite is favored by low bed temperatures, high beds and
excess air. The most favorable sulfur dioxide reduction with
this additive in this size consist was 54% a va• obtained
at 3% excess oxygen in the effluent and a 10 inch L>-^d operating
at 1500°F and an additive ratio of 1.4. A 65% reduction was
achieved but at a less favorable stoichiometric ratio.
-------
cor.
c o
iddi
a
v;as
ial
ont
• ch
e t
v,.-.
m o
est
ite
s a
acj;:
ex.
0 0 c i
ved
-------
The BCR 1359 high calcium limestone was less effective.
The best sulfur dioxide reduction values were 28% at a
stoichiornetric ratio of 1.5. The reduction was less
sensitive to bed temperature than the BCR 1337 additive
but was again favored by high beds .and excess air. With
both additives a small improvement was noted with ash
recirculation.
As a rule, the additive shows very little effect on either
the hydrocarbons or nitric oxide emissions. In one excep-
tion, however, -a 44% reduction in nitric oxide emission
was noted in one test while feeding BCR 1337 at a 1.75 ratio.
A check on the possibility of leaks or a gain change in the
recording system indicated that the result was real. About
the same time a clinker formed in the bed as indicated by
•a sudden divergence in the temperature indications of adja-
cent theriaoscouples.
Mixtures of dolomites and ashes are known to form eutectics
with melting points Tower than either component. Melting
of the additive may ." ve interfered with the formation of
NO by limiting a local high temperature in the fusion process
Another possibility is that calcination of the stone could
have reduced local temperatures.
-------
At this point in the program NAPCA suggested thcit we might
improve the sulfur capture by reducing the additive parti-
cle size to -325 mesh. This reduction in particle size
from the -7 + 14 mesh increases the lateral surface per unit
mass by a factor estimated at 40. The only adverse effect
of using fine particles could be a shortening of the particle
residence time in the bed.
*
Tests were begun with the hydrated form of the D.imestone which
occurs naturally in a 97% -325 mesh -- commonly known as
"milk of lime." The hydrate addition produced a sharp reduc-
tion in sulfur dioxide emission. With the BCR 1337 dolomitic
hydrate a reduction of 89% was achieved at a 1.8 stoichio-
metric ratio based on the calcium fraction of the dolomite.
Similar reductions were achieved with the BCR 1359 high
calcium hydrate.
-------
PILOT STUDIES IN FLUIDIZED COMBUSTION
^p» by
-f Paul S. Lewis
U.S. Bureau of Mines
Morgantown, West Virgini:iaj?t
*'•
Presented during Session IV, Pollution Control,
Control by Means of Additive Injection
Firs-t International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
Fluid-bed combustion offers possibilities for retention of sulfur
in the solid .discharge. The capacity of a given coal ash to retain
sulfur will depqsid upon its composition, but this can be adjusted and
™'':' •.•"..
controlled by the ^addition of other material to "trie fuel bed. Our experience
s *" ' ' •
in burr.ir.q hvab coc.1, rfttsburgh seam, having an ash containing 6% CaO,
with mullite grog containing no CaO is that about 10% of the coal sulfur
is in the solid discharge and 90% appears in the stack gas as sulfur
f
dioxide. The distribution of sulfur in the solids for one experiment
burning Pittsburgh coal is as follows:
Sulfur content, percent
Coal feed 2.6
mullite feed trace
flue dust 0.3
raullite discharge 0.02
An investigation will be started in the near future to determine
the effect of limestone addition on retention of sulfur.
-------
wit
nit
sc-1
One
oth
siz
mat
The
rap
s 5. i .
tu-
cc
cc-
(4.
bot-i
The
do] .
exo:
thi£
at :
at 3
achj
-------
Session V
Barriers to Commercialization
Thursday evening, November 21
Discussion of the practical technical, economic, and social
problems associated with developing widespread utilization
of fluid bed combustors for steam and power generation.
-------
ABSTRACT
BARRIERS TO COMMERCIALIZATION OF THE FLUIDIZED BED BOILER
by
A. H. Bagnulo
Pope, Evans and Robbins
Alexandria, Virginia
Presented During Session V
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park, Ohio
November 18-22, 1968
-------
ABSTRACT
BARRIERS TO COMMERCIALIZATION OF THE FLUIDIZED BED BOILER
» .. •.
•A proper analysis of the problem of bringing a new idea to
full fruition must take as its first consideration the
competitive superiority of a new idea or its potential
superiority as far as these can be determined on a realis-
tic plane. This superiority is the driving force toward
final accepteince and utilization.
Major problems in promoting the commercial use of the
fluidi'zed bed boiler stem from the following:
a. The fluidized bed boiler shares the problem common
to the development of all new ideas: the development process
\
from concept through demonstration of commercial feasibility
requires expenditure of a significant amount of money.
b. The trend away from coal in'the direction of oil, gas
and nuclear fuels has gained a momentum which is difficult
to reverse. This makes, it difficult to obtain support for
the development, effort from government or private sources.
c. The new air pollution control regulations being
'promulgated are acting as a deterrent against coal. Even
\
where favorable economics under present conditions can be
demonstrated, uncertainty about the future increases the
risk factor in the minds of supporters and users.
-------
d. Boiler manufacturers have backlogs of orders for
conventional boilers. This destroys motivation for incur-
ring the costs for a ne.w product.
On the positive side, when competitive superiority can be
demonstrated, interest will develop. The fluidized bed
boiler concept has been carried a long way in the develop-
ment phase. A complete picture of its superiority in our
present environment and its potential in the future for
air pollution conti'ol and capital cost reduction is
emerging.
Both of the government agencies sponsoring the development,
NAPCA and OCR, are interested and have indicated the
•:
intention of continued support depending on the funding
situation.
-------
Getting KLuid Bed Boilers into Service
by
G.G.Thurlow
BCURA. Industrial Laboratories
Leatherhead, Surrey, England
Presented during Session V
First. International Conference on Fluidised Bed Combustion
Hueston Woods State Park, Ohio,
November 18 - 22, 1968
1. There are obviously still many technical problems to be solved before
the full benefits of fluidised combustion systems can be realised.
Whilst a better understanding of the processes taking place in the bed
(rate of coal distribution, residence times, local heat transfer rates)
will be most useful, most of the technical problems are of a nature
that can, now, only be solved by working on plant representative
of the actual, commercial unit.
These problems include -
reduction of carbon loss by the optimum design of recycling systems,
turn down ratio
erosion and corrosion
ignition
2. Technical developments must be matched by market research and
economic studies. The conclusions of these will be specific to the
size of plant and to the country in which the studies are carried
out.
3. Design studies on a large number of alternative designs of plant should
j
be a continuing activity. While it is obviously necessary to select a
-------
limited number of what appear to be the most promising designs and to
develop these as fast as possible, it is equally important not to
"freeze" the design of plant prematurely.
k. The commercial exploitation of the industrial boiler (up to say
100,000 Ib/h) is simplified by the fact that it is economically
feasible to build, test and modify prototype plant.
5. The problem is more difficult when one considers power station
applications. While useful design data are and will be gained by
tests on pilot-scale plant (some of which is already of considerable
size'and complexity) it is imperative (at least in the U.K.) to
"prove" the value of this system in the shortest possible time and
this can only be done by building in the next year or two, plant at
least representative of a section of a large boiler. Such a pilot-
scale plant (which will probably be of the order of J>0 MW) will be
costly and it will probably be necessary to finance such projects
by collaborative effort.
-------
" DEVIL'S ADVOCATE" COMMENTS ABOUT BARRIERS
TO COMMERCIALIZATION
By Albert A. GODEL
President of the "Societe Anonyme Activit "
Paris
presented during Session V (November 21)
First International Conference on Fluidized Bed Combustion
Hueston Woods State Park , Ohio
November 18-22, 1968
-------
1 -
As far as barriers to commercialization are concerned,
I am afraid that I will have to take the uncomfortable position of the
'Devil's advocate" in putting forward for your reflection a number of
objections which I do not wholly accept myself.
Though many problems have been solved by American
and British development Organisations, I believe that difficulties still
' /be
remain to overcome. Namely the flifficulty of reaching the highest pos-
sible thermal efficiency which means first of all solving the difficult
problem of heat Change in fluid beds at low temperature. Secondly
burning to its last content of carbon the entire amount of fly ash
carried over by flue gas. On this last point, I understand that serious
improvments have been accomplished in this country and in Great Britain,
but I would be very interested to know whether the problem has. really
been solved by any other method than the one I have proposed of total
reinjection in a high temperature slagging fluid bed. It is only when
this is achieved that industrialisation and regular sales can be thought
of and this will probably require several years ; then who knows what
the trend of the coal industry will be ? .
I think that a distinction should be made between processes
using fluid beds as heat exchange medium and other combustion fluid bed
-------
- 2 -
processes simply built under standard boilers.
To this latter type belongs the Ignifluid process which has
now become more or less conventional so that the only development pro-
blems lies in increasing the size of a conventional unii .
But since we are dealing with such conventional types of boilers, there
will be no serious problem to pass from the present 60 MW to a larger plant.
,.. Yl should mention also that to accomplish the corresponding expansion of
the furnace, the necessary widening of the stoker from 1.40 m to 2 m.
presents little difficulty.
It remains to be determined of course, whether such
simplified process which does not make use of heat transmission in
fluid beds is really competitive with the conventional method of burning coal
namely P.P.
I believe I may be positive on this point : though not pre-
tending to drastic savings, the saving on investment cost would still be
about 12 % to 15 % for a utility power plant boiler burning bituminous coal,
or 18 % to 25 % when burning anthracite, and these figures will be improved
in the future.
Moreover, savings on power consumption for crushing are
to be considered to the extent of 0,5 to 0,75£of the power production when
using treated coals, savings are much more important for high ash coal.
-------
Now coming back to the process with heat exchange tubes
in the fluid bed, I am afraid that there is no evidence that this technique
can be reasonably and profitably applied all along the circuit of the flue
gas. For in fact, if we compare the amount of heat transmitted to radiating
panels at 250°C in a conventional P.P. furnace (or an Ignifluid furnace), at
about 1300°C, to the amount of heat transmitted to vaporizing tubes at 250°C
in a fluid bed at about 850°C, we find the following figures :
Heat coefficient Transferred ambuhf
per Kcal/C°/sq.m./hr. of heat / Kcal/sq.m./hr.
first case (radiating panels) 160 170.000
second case (tubes in fluid bed) 300 180.000
These figures show that the operation is not certainly _ec~
of flue gas passing in a much cooler zone, say from 850°C to 200°C for
instance, we are confronted with another difficulty which is the necessity
of using several beds and this, I admit, is what I have advocated Vo? our
prospective development. My excuse is that it would seem desirable in such
case . to take advantage of high heat transmission coefficient in the bed,
when heat trasmission needs increasing surfaces, but it is a fact that
recovery of heat in flue gas at reduced temperature should also be extended
-------
- 4 -
to its extreme limits even at the cost of grave difficulty.
As regard, this problem of limiting efficiency, it is well
known that when comparing ivestment cost for boilers having different
thermal efficiencies, one generally c.orivarts :-: intouweiVmtnfthe actual
amount of loss on efficiency.
In this respect, the following example will be of interest :
the French Electric Authorityd.etermine^ few years ago that "one point "
lost over efficiency in a 250 MW P.P. power plant, which cost appro-
ximately 215 millions frs., if amortized over thirty years, corresponds
to an actual . cost of : 3 million frs.
This as you see, corresponds to 1,39 % of the cost of the
plant. But if the loss of 3 millions frs. is affected, as it will surely be
/single
to the cost of the ._ . boiler which is approximately 50 millions frs., then
the actual, loss would be 6 %.
This gives a good idear of how necessary it is to pbtain
high efficiency whatever be the cost'. Although the problem is different for
small giants, it would be uneealistic to believe that customers will be
impressed by the low cost of a boiler if the thermal efficiency is not
sufficiently high Therefore, a drastic effort must be made in our deve-
lopment work to comply with this obligation.
Now another observation I want to submit to your reflection
is that I am afraid most of the important boiler manufacturers with order.
^backlogs will hesitate before entering into the study of drastically
new types of coal fired boilers with uncertain future, and the same reluctance
-------
- 5 -
might be felt for exploitation by customers .
This shows how wise you were in the United States to coor-
dinate such development work under government authority.
Last of all, I want to call your attention to the following
.disadvantage • . through developing coal combustion and heat
exchange processes in fluid beds, we are perhaps running the risk of
"making the bed" of competitors interested in burning gas or fuel oill!
Yet, all this taken into account, I am still convinced that
most of the above difficulties will be overcome and also that the fluid
bed combustion technique has an attractive future - because
/reduce
of its ability to pollution problems.
------- |