EPA-650/2-73-053
DECEMBER 1973
Environmental Protection Technology Series
mmmwmmwwmmymm
I
1
UJ
a
•
fli:i
^^
-------
EPA-650/2-73-053
PROCEEDINGS OF THIRD INTERNATIONAL
CONFERENCE ON FLUIDIZED-BED COMBUSTION
Sponsor:
The Environmental Protection Agency
Office of Research and Development
National Environmental Research Center
Control Systems Laboratory
Program Element 1AB103
Project Officer: P.P. Turner
Chief, Advanced Processes Section
Clean Fuels and Energy Branch
Control Systems Laboratory
Prepared for
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D.C. 20460
December 1973
-------
These proceedings have been reviewed by the Environmental Protection Agency and approved
for publication. Except for minor editing for consistency of style, the contents of this report are as
received from the authors. Approval does not signify that the contents necessarily reflect the views
and policies of the Agency, nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
-------
PREFACE
The Third International Conference on Fluidized-Bed Combustion was held October 29-
November 1, 1972, at Hueston Woods Lodge, RFD No. 1, College Corner, Ohio, under the
sponsorship of the U.S. Environmental Protection Agency (EPA).
The Conference, under the general- and vice-chairmanship of Messrs. P.P. Turner and D.B.
Henschel, respectively, was open to early arrivers Sunday, October 29.
The Conference proper, consisting of six sessions, got underway Monday morning with the
official welcome extended by EPA's Robert P. Hangebrauck. Following Mr. Turner's introductory
remarks, Dr. A.M. Squires delivered the keynote address, "The Clean Fuel Technology Gap —
Opportunities for New Fluidization Procedures."
The six-part Session I, chaired by Mr. A.A. Jonke, followed the theme, "Coal Combustion and
Additive Regeneration." Session II, held Monday evening, consisted of five presentations on the
general topic, "Non-Coal Fluidized-Bed Combustion Processes;" Mr. Alvin Skopp was chairman.
Dr. Everett Gorin was chairman of Session III, titled "Gasification/Desulfurization," which
opened Tuesday's activities. The session consisted of six presentations. "Conceptual Designs and
Economics" was the theme of Session IV. It was chaired by Dr D.H. Archer and consisted of seven
papers.
Wednesday began with Session V which presented six papers on the topic, "Pilot Plant Design,
Construction, and Operation." The Session was chaired by Mr. H.B. Locke. Session VI, the last of
the Conference, was a Wednesday afternoon discussion, led by a panel of six, summarizing
thoughts presented during the Conference and providing a final opportunity for comments from
the floor. Chaired by Mr. Hangebrauck, the panel consisted of Dr. Archer, Professor D.E. Elliott,
Mr. Henschel, Mr. H.B. Locke, and Dr. Squires.
All papers presented during the Conference are included in these proceedings.
iii
-------
TABLE OF CONTENTS
Page
Preface iii
R.P. Hangebrauck
Welcoming Remarks 0-1-1
P.P. Turner
Introductory Remarks 0-2-1
A.M. Squires
Keynote Address: Clean Fuel Technology Gap:
Opportunities for New Fluidization Procedures 0-3-1
SESSION I
1. G.J. Vogel, E.L. Carls, J. Ackerman, M. Haas, J. Riha, and A.A. Jonke
Bench-Scale Development of Combustion and Additive Regeneration
in Fluidized Beds 1-1-1
2. R.C. Hoke, H. Shaw, A. Skopp
A Regenerative Limestone Process for Fluidized-Bed Coal
Combustion and Desulfurization 1-2-1
3. R.L. Rice and N.H. Coates
Combustion of Coals in Fluidized Beds of Limestone 1-3-1
4. S.J. Wright
The Reduction of Emissions of Sulphur Oxides and Nitrogen Oxides
by Additions of Limestone or Dolomite During the Combustion
of Coal in Fluidised Beds 1-4-1
5. A.A.Godel
Selective Extraction of Clinker at the Bottom of a Deep Self-Agglomerating
Fluidized Bed 1-5-1
6. E.P. O'Neill, D.L. Keairns, W.F. Kittle
Kinetic Studies Related to the Use of Limestone and Dolomite
as Sulfur Removal Agents in Fuel Processing 1-6-1
-------
Page
SESSION II
1. H.W.Schmidt
Combustion in a Circulating Fluid Bed II-l-l
2. G.G. Copeland
Disposal of Solid Wastes by Fluidized-Bed Combustion II-2-1
3. G.L. Wade and D.A. Furlong
Fluidized-Bed Combustion of Municipal Solid Waste in the
CPU-400 Pilot Plant H-3-1
4. B.J. Baxter, L.H. Brooks, A.E. Hutton, M.E. Spaeth, and R.D. Zimmerman
Fluidized-Bed Combustors Used in HTGR Fuel Reprocessing II-4-1
5. W.E. Cole and R.J. Essenhigh
Studies on the Combustion of Natural Gas in a Fluid Bed II-5-1
SESSION III
1. J.T. Stewart and E.K. Diehl
Fluidized-Bed Coal Gasification — Process and Equipment Development III-l-l
2. F.G. Schultz and P.S. Lewis
Hot Sulfur Removal from Producer Gas III-2-1
3. D.H. Archer, E.J. Vidt, D.L. Keairns, J.P. Morris, and J.L. Chen
Coal Gasification for Clean Power Generation III-3-1
4. J.W.T. Craig, G. Moss, H.H. Taylor, and D.E. Tisdall
Sulphur Retention in Fluidised-Beds of Lime Under Reducing Conditions III-4-1
5. G.P. Curran and E. Gorin
The COj Acceptor Gasification Process — A Status Report —
Application to Bituminous Coal III-5-1
6. G.P. Curran, B. Pasek, M. Pell, and E. Gorin
Low-Sulfur Producer Gas Via an Improved Fluid-Bed Gasification Process III-6-1
SESSION IV
1. D.E. Elliott and MJ. Virr
Small-Scale Applications of Fluidized-Bed Combustion and Heat
Transfer (The Development of Small, Compact Fluidized-Bed Boilers) IV-1-1
vi
-------
Page
2. D.L. Keairns, W.C. Yang, J.R. Hamm, and D.H. Archer
Fluidized-Bed Combustion Utility Power Plants — Effect of Operating
and Design Parameters on Performance and Economics IV-2-1
3. B.R. Dickey and J.A. Buckham
Application to Combined Cycle Power Production of Fluid-Bed
Technology Used in Nuclear Fuel Reprocessing IV-3-1
4. A.N. Dravid, CJ. Kuhre, and J.A. Sykes, Jr.
Power Generation Using the Shell Gasification Process IV-4-1
5. R.A. Newby, D.L. Keairns, E.J. Vidt, D.H. Archer, and N.E. Weeks
Fluidized-Bed Oil Gasification for Clean Power Generation —
Atmospheric and Pressurized Operation IV-5-1
6. F.L. Robson
Fuel Gasification and Advanced Power Cycles — A Route to Clean Power IV-6-1
7. C.W.Matthews
A Design Basis for Utility Gas from Coal IV-7-1
SESSION V
1. G. Moss and D.E. Tisdall
The Design, Construction, and Operation of the Abingdon Fluidised Bed Gasifier V-l-1
2. M.S. Nutkis and A. Skopp
Design of Fluidized-Bed Miniplant V-2-1
3. D.H. Archer, D.L. Keairns, and E.J. Vidt
A Pressurized Fluidized-Bed Boiler Development Plan V-3-1
4. E.F. Sverdrup, J.R. Hamm, W.E. Young, and R.L. Strong
Gas Turbines for Fluid-Bed Boiler Combined Cycle Power Plant V-4-1
5. H.B. Locke, H.G. Lunn, and A.G. Roberts
Combustion of Residual Fuel Oils in Fluidised Beds V-5-1
6. H. Harboe
Fluidized-Bed Air Heaters for Open and Open/Closed Gas Turbine Cycles V-6-1
SESSION VI
1. D.B. Henschel
Minutes of Panel Discussion and Summary Session VI-1-1
APPENDIX — Attendance List
vii
-------
WELCOMING REMARKS:
Mr. R.P. Hangebrauck, EPA
INTRODUCTORY REMARKS:
Mr. P.P. Turner, EPA
KEYNOTE ADDRESS:
Dr. A.M. Squires, City College of the City University of New York
0-0-1
-------
Welcoming Remarks
1. THIRD INTERNATIONAL
CONFERENCE ON
FLUIDIZED-BED COMBUSTION
R. P. HANGEBRAUCK
Environmental Protection Agency
Control Systems Laboratory
I feel honored this morning to welcome on
behalf of EPA all those attending the Third
International Conference on Fluidized-Bed
Combustion. A special welcome goes to those
of you here from other countries, old friends
and new alike.
Each time at the end of our Conference we
have asked the attendees and ourselves
whether it was worthwhile to assemble as we
have here in the hills of Ohio, and each time I
believe we have come up with a very positive
yes. Aside from the up-to-date information
exchange, it seems that a critical mass of
expertise is achieved at this meeting causing
significant reaction to occur during the
meeting and long after.
The aim in our work on fluidized-bed com-
bustion and in our Conference is to develop
environmentally and economically sound
systems for steam and power generation which
will enable them to meet new source
performance standards and ambient air
quality standards for sulfur oxides, nitrogen
oxides, and particulates, as, provided by the
Clean Air Act. Such systems must be com-
patible with these and other present and
future environmental considerations,
including water, land, heat, and noise
pollution.
A variety of technologies for controlling
pollution from stationary combustion sources
will be forthcoming and should fit together
like pieces of a complex puzzle to solve the
applications problem in a cost and time
optimized fashion. The technology we are
working on here today is most directly
applicable to the power industry, but because
this technology will allow the use of lower cost
dirty fuels and problem fuels, it will free clean
fuels for smaller residential, commercial, and
industrial fuel users constituting area-type
combustion sources.
The projected application of technology
and fuel resources will be such that, initially,
part of the utility clean-fossil-fuel energy gap
will be filled by increased use of low sulfur
coal, physically cleaned coal, low sulfur and
desulfurized oil, and most critically, flue gas
cleaning systems which will act as a counter
balance to prevent the demand pressure for
, clean and/or cleaned fuels from driving fuel
prices too high. From now into the 1980's, the
increased demand for electrical energy gen-
erated from fossil fuel, the shortage of
naturally occurring clean fuels in the users
locations, and improved economics for flue
gas cleaning should cause a great expansion in
the application of these systems. High-Btu gas
and liquids from developing coal conversion
systems will be limited to combustion sources
considerably smaller than utilities, whose
consumer classification is such that it justifies
the inherent, much higher price. These
0-1-1
-------
processes should be commercially significant
in the 1980's in filling the gas and petroleum
gaps, but they will have their own set of
environmental problems.
We realize that even though the needs for
effluent cleaning will be satisfied by such
measures, the ultimate extent of environ-
mental control and lowest cost will not have
been achieved. Considering the size and
growth of the power industry and the multi-
billion dollar annual control costs involved,
the public, government, and industry alike can
no doubt see the potential payout for more
effective, lower cost technology.
Fluid-bed combustion systems hold great
promise for reducing or eliminating the excess
costs created by the technology gap. The
systems unfortunately are not here today, but
are under development and nearing demon-
stration, as will be evidenced by the nature of
our meeting here this week. We feel
reasonably confident the technology will be
available on a commercial basis as we move
into the end of this decade. Aside from the
built-in low-pollution nature of such systems
for several pollutants, their cost effectiveness
should cause a rapidly expanding share of the
U.S. power-generating base to be filled by
straight fluid-bed combustion systems and
advanced power cycle gasification systems
incorporating limestone/dolomite fluid-bed
technology. To this end, EPA has invested
considerable funds over the last five years, an
investment which should be dwarfed, if
successful, by the payout to the public and
industry for more cost-effective use of fuels in
generating power. I trust the progress we will
hear reported this week will bear this out.
EPA program plans rely on an increased
shouldering of the cost of development and
demonstration by industry as the scale and
cost of systems increases in the final stages of
development. Considerable effort has been
made in the EPA sponsored work to
concentrate on the most promising
approaches, and this has been caused in no
little way by the small amount of funding
available. However, it is hoped that each
promising avenue will be explored at least
somewhere in the world. It is also hoped that
all information available can and will be used
by other groups in a way that will prevent large
expenditures of funds on scale up of systems of
questionable environmental merit.
In any event, if we are to avoid getting into
an expensive technology gap as new and more
restrictive emission standards are set
consistent with the health and welfare of this
Nation and the world community, we will have
to continue to look and move ahead with vigor
on more productive and pollution-free
processes.
Once again we welcome you to the
Conference and encourage the fullest partici-
pation possible.
Thank you.
0-1-2
-------
Introductory Remarks
2. EPA-CSL PROGRAM TO CONTROL
POLLUTION FROM STATIONARY SOURCES
P. P. TURNER
Environmental Protection Agency-
Control Systems Laboratory
I would like to take this opportunity to
repeat Mr. Hangebrauck and once again
welcome each of you to the Third Inter-
national Conference on Fluidized-Bed
Combustion.
As was the case with the two preceding
International Conferences, the purpose of this
meeting is to bring together workers in the
field of fluidized-bed combustion, and related
areas, in an atmosphere conducive to co-
operative information exchange. In addition,
we have a number of organizations
represented here which, although not
currently directly involved in the research and
development effort, will become involved in
manufacturing or operating fluidized-bed
boilers and their auxiliaries, in designing
fluidized boiler plants, or in performing some
other necessary function, when this promising
new combustion technique is commercialized
in the future. It is hoped that by getting all of
you together, discussing your individual
projects, and applying your individual
expertise in this area, each of us can go home
better informed of the overall international
effort and with new ideas for direction of our
individual efforts.
A large amount of work has been
completed since the Second International
Conference was held over two years ago. For
our part, EPA has now spent a total of $7.6
million on work related to fluidized-bed
combustion through the end of fiscal year
1972, of which about $5.2 million was
committed since the last Conference here at
Hueston Woods. Also during the past two
years, design has been completed and con-
struction begun on a 630-kW continuous
fluidized-bed combustion pilot plant, capable
of operating over the full range of conditions
of interest, including pressure.
This Miniplant is designed to continuously
regenerate the partially-spent SO 2 control
sorbent generated in the fluidized combustor,
and return the regenerated material to the
combustor for re-use. Studies on the Mini-
plant will enable EPA to obtain answers for a
number of important outstanding questions,
and will provide the continuous combustion-
sorbent regeneration-data required to design
the 20- to 30-MW pilot plant envisioned as the
next stage of the development effort.
Also within the past two years conceptual
designs have been completed for a 30-MW
industrial coal-fired fluidized boiler, and for
300-MW and 600-MW utility-scale coal-fired
fluidized boilers both at atmospheric pressure
and at 10 atm pressure.
EPA's emphasis has turned toward fluid-
dized boilers operating at elevated furnace
pressures, although atmospheric-pressure
systems are also being evaluated. We are
taking a close look at continuous regeneration
of sorbent sulfated in the combustor, but
operation with a once-through sorbent non-
0-2-1
-------
regenerative system has not been ruled out.
During the past two years, a fair amount of
evaluation has been conducted, and data have
been generated, regarding pressurized oper-
ation and sorbent regeneration, in preparation
for, and complementing, the forthcoming
Miniplant program.
Development of the chemically active fluid -
ized-bed (CAFB) atmospheric-pressure add-
on residual oil gasification system, currently in
the 750-kW stage, has advanced to the point
where utility partners are being sought to
provide a site for a 82-MW prototype
installation.
Work is also continuing at an increased
scale toward the development of a high-
temperature fuel-gas desulfurization process
to produce clean low-Btu fuel gas from caking
bituminous coals.
All of this research and development effort
sponsored by EPA will be described during the
course of this meeting by the individual con-
tractors involved. We also look forward, of
course, to the discussions of work being
conducted by organizations other than EPA in
these and related areas.
Expressing the hope that the activities of
this conference during the next three days will
lead to the healthy exchange of information
between all the participating members, and to
numerous individual contacts, I declare this
Third International Conference on Fluidized-
bed Combustion open.
It is now my pleasure to introduce our key-
note speaker who will address the conference
on "The Clean Fuel Technology Gap: Oppor-
tunities for New Fluidization Procedures." He
will review the problems which one sees ahead
in the natural gas and petroleum markets; he
will characterize the substitute technologies
which will be needed; and he will review
advantages of new fluidization procedures for
treating coal and residual oil.
He is indeed well qualified to set the tone
for this conference. Though his formal back-
ground is in chemistry, his interest in
engineering was aroused during World War II
through his association with Dr. Manson
Benedict, whom he assisted in the process
design of the Oak Ridge gaseous diffusion
plant.
After the war he was Director of Process
Development at Hydrocarbon Research, Inc.
until 1959, when he resigned to become an
independent consultant. He joined the faculty
of the Department of Chemical Engineering of
The City College of The City University of New
York in September 1967, and was named
Chairman of that Department in the fall of
1970.
He has published extensively on fluid-
ization, oil and coal gasification, fuel
desulfurization, gas cleaning, and power gen-
eration; he holds 15 U.S. patents in these
fields. He has conducted research at The City
College under EPA grants relating to the
development of new systems for generating
clean power from fossil fuels. His team at The
City College began work last June on the first
18 months of a 5-year effort under a grant
from the National Science Foundation to
support "studies toward improved techniques
for gasifying coal." The objective of these
studies is to convert coal into pipeline gas and
a light aromatic liquid fuel as well as low-Btu
gas for power generation. Gentlemen, I
present to this conference Dr. Arthur M.
Squires.
0-2-2
-------
Keynote Address
3. CLEAN FUEL TECHNOLOGY GAP:
OPPORTUNITIES FOR NEW FLUIDIZATION
PROCEDURES
A. M. SQUIRES
The City College of the City University of New York
Before we take up the specific problems of
interest to this Conference, we may well first
view these problems in the broader context of
the markets for clean and convenient energy.
To do so even briefly creates a sense of urgency
and a demand for boldness. The World's
appetite for clean fuels is sharply rising. That
the World's resources of cheap clean fuels are
finite is an inescapable fact, and economic
consequences of this fact are beginning to be
felt.
Consider Figures 1 and 2, which depict
broadly the supply and demand situation for
oil and natural gas in the United States
between 1955 and 1985. The range of uncer-
tainty in demand beyond the present is
approximately the range among recent pro-
jections for 1985.
Figure 1 implies that America's economic
growth is in jeopardy if we cannot import from
one -half to two-thirds of our oil supplies in
1985. Imports to reach the upper curve
amount to substantially the entire present out-
put from the Middle East. Notice that oil from
the North Slope has small relative effect. We
would need to discover a North Slope each
year between now and 1985 to reach the upper
curve of Figure 1 from domestic supplies.
Gas cannot be imported from overseas as
readily as oil, and Figure 2 implies a sharp
flagging in the growth of the gas market. We
are using a quantity of gas yearly that is
greater than the average annual discoveries of
gas made over the past 20 years and more than
twice the discoveries of last year. No doubt
discoveries can be increased by removing
artificial restrictions on the price of gas at the
wellhead, but no prompt effect on gas supplies
could result. It takes several years to bring a
new field to production, and in the meanwhile
old fields decline. Nuclear stimulation is a
doubtful proposition in light of concern over
spread of radionuclides. All substitute natural
gas from sources now in view will cost at least
about $1 per 106 Btu, and often more. This
includes gas from the Far North. It should be
remarked that not enough is known yet of
North Slope gas reserves and production
problems to project with assurance its supply
to the United States market, yet some gas
from the Arctic can probably be delivered by
about 1980 with vigorous development,
Canada willing.
Since much of the historic growth in
demand for gas has been for fuel to fire
boilers, a projection of the demand for clean
boiler fuel would reveal a much greater gap
between supply and demand than Figure 2
would suggest. The Nation wishes to eliminate
emissions from boilers fired with coal and
untreated residual oil. Satisfactory engineer-
ing solutions to the problem of ridding stack
gases of sulfur'dioxide are not yet in hand, and
the problems of the large coal-fired power
stations projected for the Southwest illustrate
the dislocations of the forced shift from gas to
coal in new plant construction.
An ironic illustration of the sharp change
in the gas market is furnished by the news that
0-3-1
-------
Coastal States Gas Producing Company will
build a plant to manufacture synthetic natural
gas from petroleum feedstocks in Texas.
It is hard to escape the impression that our
energy markets will undergo price upheavals
in the next decade. In view of this, it must
seem astonishing to a layman that our fuel
industries are so little prepared with substitute
technologies. For example, to convert coal into
clean gaseous or liquid fuels, they can
absolutely rely, for immediate construction,
only upon technology introduced nearly 40
years ago to fuel the German war machine.
Moreover, nothing better will be ready to have
much effect in the time span of Figures 1 and
2. Our research and development efforts, both
private and governmental, have been far too
inadequate. The layman must be further
astonished to learn that private efforts are
being reduced.
At least five major oil companies have shut
down laboratories or layed off personnel
engaged in synthetic fuels research. One
architect-engineering firm that caters to the
fuels industries has shut down an historically
important laboratory, and other such firms
appear to be decreasing their research
budgets. Yet even a Manhattan Project on new
fossil fuel technologies could not be expected
to make a major dent in the clean fuel gap
before 1985. It is highly improbable that
synthetic liquid fuels made from coal could
play a significant role before then. Most
substitute natural gas will be made from
petroleum feedstocks, which will themselves
become short in supply. The extent to which
we have been minding the ^ store, in this
particular respect, is ironically illustrated by
the fact that we will license the SNG processes
for petroleum feedstocks from Great Britain,
Japan, and Germany at fees that will total
more than $100 million.
It will be important to understand how the
clean fuel technology gap has arisen. To what
degree has research and development failure
been due to governmental interference with
"normal" economic processes? To what
degree is it due to size and maturity of our
energy industries and concomitant risks and
high costs of research and development for
these industries? Is it due to a general loss of
appetite for risk-taking among the managers
of our technology? Is it due to a general
migration of creative talent, promotional and
managerial as well as technological, into
glamorous activities such as the space effort?
Are there other factors?
Understanding these questions will be
important, not only so that our domestic
arrangements for fuels research and develop-
ment may be overhauled, but also so that we
may better prepare ourselves for an even more
serious clean fuel technology gap which lies in
the not distant future.
Figure 3 gives a gloomy but plausible
scenario for the future course of the world
petroleum market. If the attitude of the World
toward oil parallels that of the United States
toward natural gas, something like the
scenario of Figure 3 will inevitably unfold.
Substitutes for oil will be developed too late;
production of oil will reach the limits of the
World's capability to yield oil before sufficient
experience with substitute technologies has
been acquired; growth of technologies
dependent upon oil will be choked off; and
economic disarray as well as insufficient
experience will prevent rapid growth of
substitutes.
The historic excess in our capacity to
produce gas, seen in Figure 2, was important
to the petrochemical industry in the United
States. Disappearance of the excess capacity,
along with unwise import regulations for light
hydrocarbon feedstocks, is creating serious
difficulties for this industry, which illustrate in
miniature the dampening effect that the dis-
appearance of excess oil production capacity
will have upon invention and development of
new technologies for better use of oil. Existing
equipment and existing technologies will pre-
empt supplies, and opportunities to divert oil
from uses of lower to higher value will be
missed.
0-3-2
-------
Electricity from whatever source (nuclear
fission or fusion, solar, or geothermal) cannot
be readily substituted for clean, portable fuels.
Estimates of the U.S. fuel mix in the year 2000
postulate only about 25 percent nuclear at
best. For the distant future, electricity could
electrolyze water to yield hydrogen. This could
be used directly as a fuel, stored either as
liquid or at high pressure or reversibly
adsorbed upon a solid. Alternatively,
hydrogen and carbon dioxide could be con-
verted to hydrocarbons by Fischer-Tropsch
synthesis, or hydrpgen and nitrogen could be
converted to ammonia. By what steps and over
what kind of time span might such a synthetic
fuel economy be introduced? In an early
transitional stage, electrolytic hydrogen might
add to the heating value of carbon-rich fuels of
natural origin, converting them to lighter
materials. Even earlier, if natural gas is still
available in regions like Venezuela and the
Middle East, hydrogen might be made from
the gas and added to carbon-rich fuels.
To what extent can the world rely upon
synthetics based upon coal? It should be noted
that the Northern Hemisphere is much richer
in coal than the southern. Even in the
Northern Hemisphere, coal deposits are
concentrated in a few favored countries
(notably the United States and Russia,
especially the latter). If SNG from Texas is a
surprise, what about exports of synthetic
liquid hydrocarbons from Wyoming? Our coal
resources seem vast, but would they if
measured against the projected world appetite
for oil in the 21st Century?
What other substitute technologies based
upon electricity can replace liquid fuels? Can
nuclear energy (or solar or geothermal) be
increased beyond present projections before
the year 2000? On the other hand, what
happens if our second try for a liquid-metal-
fast-breeder-reactor is a flop, as was the first?
Can solar-pond algae plausibly contribute
to the gap? Hydrocarbons based upon human
and animal wastes? Cellulose? What about
the "substitute" of making do with less? If
there must be a belt-tightening anyway, it
would be better earlier than later.
Can a combination of these or other
developments, carried out in a timely manner,
produce the more attractive scenario of Figure
4?
Figures 3 and 4 carry no scale for time.
Hubbert has drawn his celebrated "dimple"
for two estimates of "recoverable" world oil
reserves (meaning recoverable at costs not far
advanced from current costs): 1350 and 2100 x
109 barrels. At the lower figure, production
begins to depart from a substantially
exponential rate of growth'at around 1980 and
reaches a peak at about 66 x 106 bbl/day
in about 1990. At Hubbert's higher figure,
he projects a "dimple" with substantially
exponential growth until about 1990 and a
peak production at about 102 x 106 bbl/day
shortly after the year 2000. Current produc-
tion is nearing 50 x 106 bbl/day, and appears
to be ahead of Hubbert's projections.
At Hubbert's lower estimate of reserves, the
crisis in Figure 3 could come as early as about
1985. At the higher estimate, about 2000. This
is assuming at least modest cooperation from
both oil-producing nations and authorities
responsible for leasing off-shore drilling sites.
The key to the scenario of Figure 4 is timely
research and development and timely
construction of full-scale installations in which
to practice the new technologies. It is already
too late to prepare for the world clean fuel
crisis if it should appear as early as 1985.
There is just barely time, perhaps, to get ready
for problems that might reach critical
proportions in 2000. Thirty years is not too
long to develop and test a new technology and
to learn it sufficiently well that it may be
expanded as rapidly as the scenario of Figure
4 will require.
Energy decisions made in the United States
in the next few years can be crucial in the
choice of scenarios. That of Figure 4 may well
imply the lower curves of projected fuel
demand in Figures 1 and 2 as well as a
0-3-3
-------
Manhattan Project to acquire good tech-
nologies for converting coal to gas and oil.
Simultaneously, we must not neglect vigorous
attack on the substitute technologies whose
commercialization must begin in the 1980's if
they are to be ready to forestall a world energy
crisis.
OPPORTUNITIES FOR AGGLOMER-
ATING AND FAST FLUIDIZED BEDS
Our first chapter has raised more questions
than it has provided answers.
My second chapter, turning to the interests
of this Conference must inevitably be a more
particular response to the urgent call for new
fossil fuel technologies.
My colleagues at The City College and I
believe that the time is coming soon when
economics will turn against the practice of
burning chemically-bound hydrogen for large-
scale production of electricity. Instead, the
bound hydrogen in coal or residual oil will be
viewed as too valuable to burn and send as
water vapor up a stack. The hydrogen can
become a part of some clean, convenient fuel
having an economic value higher than coal's
or residual oil's. The hydrogen-rich fuel would
be "creamed off the coal or residual oil
leaving a carbon residue that would be burned
to generate heat or electricity.
The idea of creaming off valuable products
from coal or oil is of course not original with
us. The byproduct coke oven is 100 years old;
many attempts to displace it with improved
coking procedures have been recorded; and
the oil industry has made steady advances in
technologies for reducing the yield of residue
and increasing yields of lighter products. Most
attendees at this Conference will be familiar
with Consolidation Coal Company's efforts, as
well as the recent achievements of the FMC
Corporation.
Any good idea, however, can stand
constant review in light of the appropriate
0-3-4
technological context. We have tried to review
the foregoing idea in light of the ongoing
development to supply the military with better
gas turbines, funded at about $300 million per
year. Progress in engines for civil aircraft and
stationary power has historically followed
military achievements after a lag of only a few
years. The 747 flies today with a temperature
of 1300°C at the inlet to the turbines.
Stationary machines larger than 100 MW are
promised for this temperature before 1980.
Existence of such machines will allow
design of electric power installations of
sharply advanced efficiency and reduced
capital cost. Gas turbines (inherently cheap in
cost) will supply about 50 to 60 percent of the
power, and a steam system (using modest
steam conditions and a cheap boiler) will
scavenge heat from gas turbine exhaust.
The prospect of these designs creates an
imperative to develop better techniques for
gasifying coal and residual oil to provide a
cheap "power gas" to run the gas turbines—
i.e., a low-Btu gas made using air and a little
steam as the gasification medium. (Power gas
was Ludwig Mond's term for producer gas,
which he used to generate electricity in gas
engines. He founded The Power Gas
Company.)
The City College team believes that power
gas technologies will inevitably evolve to allow
some cream-skimming that pulls out fuel
products of value greater than coal or residual
oil. Nevertheless, power gas technologies will
probably arise in the first instance for treat-
ment of raw fuels. It is doubtful that at first
any more than very slight consideration can be
given to the potential of these technologies for
profitable evolution. Let us first consider
power gas - technologies for raw fuels, and
second, how they might evolve.
PRODUCTION OF POWER GAS
Among gas-to-solid contacting procedures,
fluidization will be the strongest candidate for
employment in better technologies for gasi-
fying coal or residual oil. To a large degree,
this is so simply because of the scale of the
-------
power stations to be built in the 1980's and
beyond. Sites for over 4000 MW are already in
service. Many more will be built. Typical sites
will process coal, for example, on the scale of
scores of thousands of tons per day; oil, at
hundreds of thousands of barrels per day.
Only fluidization procedures can readily
provide equipment of capacities that will avoid
the necessity of processing coal or oil in an
unattractively large number of vessels oper-
ating in parallel.
Although the science of the gravitating bed
is far ahead of our knowledge of fluidization,
the former art is difficult to build in large
capacities. The blast furnace, after more than
a century of development, gasifies up to about
4000 tons of coke per day, but coke of course is
a processed and closely-sized solid. More than
20 Lurgi gravitating-bed pressure gasifiers of
the current design would be needed for 1000
MW; scale-up of this approach may prove dif-
ficult and uncertain. Although the gravitating-
bed had significant yield advantages, it did not
win in competition with the fluidized-bed
catalytic cracker because it could not easily
reach capacities appropriate to the scale of oil
processing after 1960.
Gasification of either coal or oil in absence
of a bed of solid appears plagued with a
carbon loss problem which may inevitably
require extra equipment for extinction of
carbon. Either Texaco or Shell "partial oxi-
dation" of residual oil must provide for carbon
recovery and recycle to achieve complete
carbon utilization. Nothing was published
from Texaco's large experiment with a slag-
ging, dilute-phase gasifier at Morgantown in
1957, but experience at Bell and elsewhere
suggests that carbon utilization in such a
gasifier may be poor, especially for a coal with
a refractory ash.
I have not seen any advantage in gasi-
fication of coal in a pool of iron, a procedure
of doubtful integrity at elevated pressure and
doubtful operability at atmospheric pressure.
The City College view is that the strongest
candidate for gasifying coal to obtain power
gas will combine the ash-agglomerating fluid -
ized bed about which we will hear from Godel
and the circulating fluidized bed that Schmidt
will describe. The combination would operate
at about 2000 °F and about 10 ft/sec velocity.
A single vessel could easily handle coal for
1000 MW; at 20 atmospheres, the diameter
would be less than 20 feet. The role of the ash-
agglomerating bed would be to gasify large
particles of coal, up to about 3/4-inch as well
as to agglomerate and separate ash matter
from the carbon-rich bed. At The City College
we have re-dubbed Lurgi's highly expanded
circulating fluidized bed the "fast fluidized
bed;" we have a two-dimensional fast bed of
plexiglas in operation that is exciting to watch.
The fast bed in the gasifier combination would
gasify fines and would provide a zone of high
velocity and intense circulation for intro-
duction of a caking coal. It is reasonable to
hope that a caking coal could be successfully
introduced into the fast bed, fine particles
joining the bed and a large particle being
coked sufficiently on its surface to render it
harmless before it reaches the ash-agglomer-
ating bed below.
Data by Dent, which I have discussed in
"Role of Solid Mixing in Fluidized-Bed
Reaction Kinetics" to appear in an AIChE
Symposium Series volume, strongly suggest
that the kinetic performance of the proposed
combination will be excellent. Because of the
high temperature and good kinetic
performance, flow of steam will be small
relative to air. Table 1 compares Lurgi
gravitating-bed power gas with gas from the
proposed gasifier calculated with the
assumption that steam-carbon equilibrium is
substantially achieved therein, an assumption
suggested by Dent's experience. Table 2 com-
pares electricity-generating efficiencies of
combined-cycle installations (with power
equipment according to United Aircraft's
"second generation" design) using the Lurgi
gasifier with wet gas cleaning, or the ash-
agglomerating-fast-bed gasifier with wet
cleaning, or the latter gasifier with gas
cleaning by the high temperature procedures
that we have under study at The City College.
0-3-5
-------
Table 1. COMPARISON OF CRUDE POWER GAS
FROM THE LURGI GASIFIER AND A CANDI-
DATE FOR DEVELOPMENT
Composition, % by vol
Methane
Carbon monoxide
Hydrogen
Carbon dioxide
Water vapor
Nitrogen
Hydrogen sulfide
Heating value, Btu/ft 3
Lurgi
gravitating-bed
gasifier
4.4
10.7
15.7
10.7
27.8
30.2
0.5
100.0
129
Gasifier combining
ash-agglomerating
and fast
fluidlzed bads
0.5
31.8
15.6
0.5
0.5
50.4
0.7
100.0
157
The high loss of latent heat in the stack
from a combined-cycle plant that depends
upon the Lurgi is inherent for this gasifier.
The loss arises from two sources: air supplied
to the gasification bed is necessarily
accompanied by a high flow of steam in order
to limit the temperature and to keep the ash
free-flowing; and gas leaving the bed must be
quickly reduced in temperature by a water
quench in order to prevent formation of heavy
tars that would lead to deposits of coke.
A fluidized-bed gasifier operating at
2000° F will not make tars or tar-forming
species, and its power gas need not be
quenched.
One cannot be so confident in respect to a
gasifier working at 1700°F. The Wirikler did
not make tars, but secondary oxygen was
introduced into the Winkler above the fluid-
ized bed in order to raise the temperature (to
reduce methane yield? to eliminate tar-
formers?). Late-model Winklers were pear-
shaped and had enormous freeboard regions.
For operation at 1700°F, a carbon-burnup
step must be provided, as Pell will describe,
and there will probably be a price also in
capacity and in loss of latent heat to the stack,
as well as a possible problem with tar-formers.
Nevertheless, until development of a gasifier
exploiting Godel's ash-agglomeration
principle is far enough along, an approach at
lower temperature is well worth carrying as an
alternate.
Experience at Hydrocarbon Research,
Inc.and elsewhere suggests that defluidization
phenomena might plague an attempt to
develop a gasifier to operate between about
1800 and 2000°F, unless perhaps the develop-
ment were to adopt a fluidization velocity so
high as to approach the fast-fluidized state.
The limits would of course vary with
properties of coal ash, and may extend below
1800°F for some important coals.
I have not yet seen advantage for an
approach using an initial step carbonizing
coal at about 1700°F followed by gasification
at 2000°F, the flows of solid and gas between
the two steps being countercurrent. During
this Conference we will perhaps hear evidence
supporting an advantage for this approach.
However, there would appear to be a sub-
stantial risk that tar-formers will appear in gas
from a carbonization step at 1700°F,
requiring a rapid quench of the gas to prevent
coke laydown in transfer lines. To avoid
unnecessary degradation of heat, a quench is
of course best avoided except in a case where a
useful quantity of liquid-fuel byproduct can be
recovered.
I have also not seen advantage for a
separate ash-agglomerating zone such as
Je'quier provided. This was a zone lean in
carbon and was both more dilute and hotter
than Je'quier's primary fluid bed. It afforded
complete combustion of air furnished thereto,
and it served to density ash agglomerates and
to reduce their carbon content. If reduction of
carbon content in ash agglomerates should
become a problem, Godel has provided two
simpler approaches. His grate emerges from
the bed, and agglomerates thereon are
exposed to air in absence of coke particles
external to the agglomerates. Godel has also
demonstrated burnup in a gravitating-bed of
agglomerates from an anthracite of high ash
0-3-6
-------
Table 2. ILLUSTRATIVE ENERGY BALANCES FOR COMPLETE POWER-GENERATION
FLOW SHEETS
Category of energy
loss, %
Electricity sent out
Heating value of sulfur
Loss of sensible heat
in stack gases
Loss of latent heat
(water vapor)
Loss of heat at steam
condenser and elsewhere
Loss of unburned fuel
and heat leakage
Mechanical losses and
power for auxiliaries
Efficiency, allowing
credit for heating
value of sulfur
Conventional
steam power
equipment
without
recovery of
sulfur
39.5
-
5.0
3.8
47.7
2.0
2.0
100.0
39.5
Combined-cycle power equipment
with recovery of sulfur3
Lurgi gasif ier.
gas cleaning at
low temperature
45.0
4.6
14.1
28.4
4.9
2.0
100.0
45.5
Gasif ier combining ash-agglomerating
and f ast f luidized beds
Gas cleaning at
low temperature
49.1
1.1
4.5
5.6
35.7
2.0
2.0
100.0
49.7
Gas cleaning at
high temperature
50.5
1.3
4.7
4.5
35.0
2.0
2.0
100.0
51.2
a United Aircraft "second generation" design parameters.
-------
content, the bed being supplied with air at the
bottom. If carbon burnup resists these
solutions, a separate burnup step exploiting
Battelle's carbon-lean ash-agglomerating bed
for complete combustion would be preferable
to Je'quier's arrangement.
In light of work at Esso Abingdon by Moss
and his colleagues, gasification in a fluidized
bed containing lime emerges as a strong
candidate for production of power gas from
oil. Presence of tar-formers in the Abingdon
gas gives cause for concern, however. Periodic
burnout of coke deposits would not be an
attractive procedure at high pressure. In a
commercial embodiment of the Abingdon
ideas, could the cyclone on the gasification
bed be dispensed with? My thought would be
to introduce secondary air directly above the
gasification vessel at the elevation where the
fuel gas enters the larger cross-section of the
boiler being served. For production of power
gas at high pressure, a comparable idea would
be to introduce secondary air above the fluid-
ized bed, raising the temperature of the power
gas (eliminating tar-formers? as in the
Winkler?). If this does not work, we might be
forced to supply much more air than
Abingdon uses to a gasification bed working
at high pressure, thereby producing a leaner
gas containing less hydrocarbon species. Heat-
removal surface would need to be provided.
EVOLUTION TO THE FUELPLEX
W.C. Schroeder published data in his U.S.
Patent 3,030,297 (April 17, 1962) that seem to
The City College team to provide a strong lead
in respect to a candidate for the first fuel-
treating step in a "Coalplex" producing
substitute natural gas, liquid fuel, and elec-
tricity. Solihull and Bruceton long ago taught
that the treatment of raw coal with hydrogen
at elevated pressure and at temperatures
between about 1500 and 1800°F can result in
attractive yields of methane at high concen-
tration. In Solihull and Bruceton experiments,
the residence time of vapor product generally
ran into the minutes. Schroeder's contribution
has been to call attention to the attractive
yields of benzene, toluene, xylene, and sub-
stantially the same yield of methane,
accompanied by very little heavy tar if the
vapor product residence time is kept short,
preferably around 5 seconds. At the long
residence time of the earlier experiments,
aromatic products polymerized, condensed as
heavy tar upon the coke present in the reaction
zone, cracked to form additional coke, and
vanished. By arranging for a rapid quench of
vapor species to about 700° F, Schroeder
preserved tar-forming aromatic species and
thereby obtained an attractive yield of a light
aromatic liquid.
The City College team sees Schroeder's
chemistry as opening up the possibility for a
Coalplex yielding roughly 25 percent of the
coal's heating value in form of methane and
perhaps 15 to 20 percent as BTX, with the
remainder being converted to electricity at an
efficiency beyond 40 percent. The heat
degradation that results from Schroeder's
quench is tolerable, amply rewarded by ;the
yield of liquid product that the quench
preserves.
For Western coals, the fast fluidized bed is
a candidate device for conducting Schroeder's
chemistry. The fine coke product could be
circulated from the fast bed to a heating step
and returned to sustain the slightly
endothermic Schroeder reactions.
For Schroeder's chemistry with Eastern
coals, the coke-agglomerating fluidized bed
that I discussed here during our Second Inter-
national Conference is a candidate device,
perhaps with a superposed fast fluidized bed
of a fine solid that is circulated to provide
heat.
For either coal, the coke product could be
gasified in the aforementioned combination of
Godel's ash-agglomerating fluidized bed and
a superposed fast bed to deal with fines.
The coke product could also be burned up
in a fluidized bed boiler. Although The City
College team regards production of power gas
0-3-8
-------
to be the main chance, nevertheless we have a
healthy respect for the difficulties of gasifier
development, and the boiler is welcome
competition. Thanks to Godel's 35 Ignifluid
"gasifiers" (to call them "boilers" in this
context is misleading), the race at the moment
appears about even. •
A counterpart of Schroeder's procedure as
applied to residual oil should yield even higher
quantities of methane and BTX and less coke.
Coke from oil treatment is practically certain
to contain negligible sulfur. In form of beads
roughly 1/12 to 1/2-inch in size, the new
petroleum coke, after calcination, should find
lively markets for electrode manufacture and
metallurgy.
Whither Fossil Fuel Development?
Our first chapter set down a gloomy picture
of fossil fuel research and development. Past
efforts, both private and governmental, have
not been responsive to urgent needs, and the
former are contracting.
Our second chapter, although it focussed
upon The City College's view of the future,
nicely illustrates the necessity for sharply
expanded research budgets.
In perhaps no engineering procedure other
than fluidization is art so far ahead of science.
Neither ash-agglomerating, coke-agglom-
erating, fast fluidized beds, nor other com-
peting procedures for processing coal will be
developed through computer modelling or
"optimization" studies. Godel discovered his
ash-agglomeration phenomenon in 1954
during hands-on hardware development of
processes for making activated carbon from
coal. Although he had no experience with
boilers, Godel promptly recognized a new
capability to gasify and render burnable
anthracite slacks from the mine at La Mure,
near his boyhood home at Vif, Isere. With
help from Babcock-Atlantique, he had a small
boiler operating inside a year. Although the
fast fluidized bed arose from Lothar Reh's
work on fluidization in cones for his Dr.-Ing.
Degree, Lurgi had to carry the development
through frustrating difficulties as Reh's team
learned how to achieve the fast-fluidized state
in large-scale, practicable equipment. The
origins of agglomerating beds that make dense
beads of agglomerated material are obscure to
me. Chance observation by Dorr-Oliver during
development of a fluidized roaster for a
"sticky" Katanga ore may have been
important. Dorr-Oliver had a commercial
process for calcining calcium carbonate slimes
before 1957, and a number of pilot operations
were underway at Dorr-Oliver, Fuller, and
Battelle before 1960. Yet in 1972 we have
hardly begun to acquire a scientific knowledge
of these beds.
A number of factors have joined to create
an illusion, shared by public and political
leaders, research managers, and alas far too
many engineers, that a new development
requires only laboratory results and design of
appropriate equipment based simply upon
scientific analysis. One factor has been the
dominance of the physics establishment in
selection of recent national R&D goals. A
second factor has been emphasis among
educators upon engineering science to the
detriment of an appreciation for hands-on
development of new engineering hardware.
The latter has somehow not been respectable
by comparison with fancy mathematical
analysis; it also gets a fellow dirty. A third
factor, of special importance to the chemical
engineering profession, has been the success of
the approach exploited so brilliantly by
Scientific Design Co., Inc. immediately after
World War II; viz., bench development of
fixed-bed catalytic processes followed directly
by commercial-scale equipment. A literature
has even grown up on the theme, "the pilot
plant is obsolete." This situation may exist for
fixed-bed processes in which there is no risk of
minor unwanted products undeteptable at the
bench yet troublesome in the field, but this
special case should not be elevated to a
philosophy. Permit me a bit of personal
history. I got an altogether false idea of
0-3-9
-------
engineering from my first job working on the
design and startup of the first gaseous
diffusion plant at Oak Ridge: this plant was
built from bench data, was "scientifically de-
signable," and was a justly celebrated success.
1 learned the other side of engineering through
my participation in the classic flop of
fluidized-bed process development. The
hydrocarbon synthesis reactors failed at
Brownsville because the development lacked a
timely pilot unit of adequate size. If Dubie
Eastman's 12-in. pilot plant at Montebello
had operated 3 years sooner, we would have
known. Our research budget had been too
small.
Budgets unresponsive to our clean fuel
needs have led to a preference for "safer"
experiments relying more heavily upon earlier
experience and representing a "simpler"
approach. We have learned that the "simple"
is not always so: witness the misfortunes of the
dry-limestone-injection approach and the
troubles of limestone scrubbing. In the mean-
time, for lack of making ourselves ready,
unanticipated opportunities arise that cannot
be seized. Studebaker-Worthington's
Turbodyne and Southern California Edison
are showing how old steam turbines can be
converted to a combined cycle by scrapping
old boilers and adding gas turbines followed
by new waste heat boilers. General Electric
and Westinghouse report brisk sales of
combined-cycle equipment. A market for
power gas is developing right under our eyes,
and only Lurgi is ready for coal and, except
Texaco-Shell, partial oxidation for oil.
In relation to the possibilities and the
urgency of our needs, progress in areas to be
covered during this Conference must seem
disappointing to any visitor at Alexandria 5
years ago who admired the remarkable hands-
on hardware development already accom-
plished by Pope, Evans & Robbins by that
date; or, to a visitor to Leatherhead more than
3 years ago who saw BCURA's feat for burn-
ing coal at elevated pressure at rates
approaching 1000 pounds per hour. Fluidized-
bed boiler development in the United States
0-3-10
needed Esso's "miniplant" 3 years earlier and
10 times bigger, but our hosts at this Confer-
ence simply have not had available to them the
requisite funds. Nearly 3 years ago at the
Christmas AAAS meeting in Boston (and on
many occasions subsequently), I pointed out
the potential value of Godel's ash-agglomera-
tion phenomenon in combination with the fast
bed in a maker of power gas, and I said it was
a shame that we had no commercial exper-
ience here with the phenomenon. It is still a
shame. Our fossil fuel industries have not had
the simple curiosity to buy a small Ignifluid
boiler, a commercial proposition on which
nothing would be lost, to gain firsthand exper-
ience from its operation. Our hosts here have
not had available funds to make good this
omission.
Let me close by quoting from F.C. Dent's
Melchett Lecture of 1965. This great develop-
ment engineer, now enjoying life on his yacht
out of Malta, can reflect with pleasure on the
scores of millions in royalties that the SNG
processes he developed at Solihull will bring
during the next few years to the British
economy. In terms of the quotation to follow,
our own efforts to develop SNG processes for
coal may be said to be just beginning.
It is significant. . .that we usually had
reason to regret any protracted period
of exploratory laboratory investigation.
Small-scale experiments have often
been time-wasting even when large-scale
conditions have been reproduced as
faithfully as possible. . .Operation on a
reasonable scale at an early stage is most
desirable to throw difficulties into their
proper perspective. Laboratory work was
of most value after the problems had
been recognized in this way.
All technologies addressed toward closing
the clean fuel technology gap must include a
major materials-processing step, handling
solids, liquids, or gases on a scale almost with-
out precedent in chemical engineering art.
Serious effort does not begin until this step is
-------
addressed by hands-on hardware development
on a practical scale. Budgets must be big.
Programs must be bold.
APPENDIX
Energy Studies at The Department of
Chemical Engineering of The City College of
The City University of New York
Nine studies are in progress. All except III,
IV, and IX are supported by Grant GI-34286
from the RANN Program (Research Applied
to National Needs) of the National Science
Foundation. Professors Michael Gluckman,
Robert Graff, Robert Pfeffer, Reuel Shinnar,
and Joseph Yerushalmi; Dr. Norman
Holcombe; and Messrs. Samuel Dobner, Kun-
Chieh Lee, Dennis Leppin, Jeffrey Silverstein,
Eugene Yu, Stanley Dobkewitch, and Nurettin
Cankurt are participating in the effort. In the
past, Drs. Leon Paretsky, Melvyn Pell, and
Lawrence Ruth, and Messrs. Richard
Angiullo, Richard Earth, Ralph Levy, Basil
Lewris, Michael Somer, and Lauris Sterns
made substantial contributions. Messrs. John
Bodnaruk, George Dilorio, Michael
Askenazy, and John Spencer have helped with
experimental arrangements.
Research on Power Gas:
I. Study of the Godel Ash-Agglomeration
Phenomenon.
II. Study of Kinetics of Carbon Gasi-
fication in a Fluidized Bed. Our objective is to
test F.J. Dent's hypothesis that the superior
kinetics afforded by a fluidized bed for the
steam-carbon reaction are a consequence of
solid mixing in the bed, bringing about
repeated exposure of carbon to fresh gasi-
fication medium.
III. Study of Kinetics of Removal of Sulfur
Compounds from Power Gas by Action of
Calcined Dolomite.
IV. Study of Removal of Fine Dust from
Power Gas by a Panel Bed Filter. The filter
can be built to clean gas at temperatures
approaching 1800°F. We have achieved clean-
ing efficiencies beyond 99.99 percent for
power station fly ash at a normal stack dust
loading in small-scale tests at atmospheric
temperature.
Research on the Coalplex:
V. Study of Reaction of Coal with Hydro-
gen at High Temperature and Pressure and
Short Residence Time of Vapor Products. We
believe this reaction, disclosed by W.C.
Schroeder, can be the first coal-treating step
in a Coalplex shipping substitute natural gas,
BTX, and electricity at costs below the
combined cost of making each product
separately from coal.
VI. Study of Coke-Agglomerating Fluid-
ized Bed, a candidate device for conducting
Schroeder's chemistry (as in V above) on
Illinois and other Eastern coals.
VII. Study of the Fast Fluidized Bed, a
candidate device for conducting Schroeder's
chemistry (as in V above) on Western coals;
also, for gasifying fine particles of carbon
blown from an ash-agglomerating fluidized
bed.
VIII. Flowsheet Studies for the Coalplex.
Research on the Oilplex:
IX. Flowsheet Studies for an Oilplex in
which oil is first treated by reaction with
hydrogen at high temperature and pressure
and short residence time of vapor products.
0-3-11
-------
10
&
35
30
25
20
15
10
W»|= IMPORTS
^"'' AND SYNTHETICS
CONSUMPTION
DOMESTIC
PRODUCTION
,_ CAPABILITY
"DOMESTIC
PRODUCTION
FIRST OIL
FROM
NORTH
SLOPE
1955
1965 1975
YEAR
Figure 1. United States petroleum
supply and demand.
1985
30
25
S. 20
3
ce.
f 15
10
I DOMESTIC
PRODUCTION
,-«* CAPABILITY^
CONSUMPTION
FIRST GAS
FROM NORTH
SLOPE
I DOMESTIC
N ppnnnrTii
PRODUCTION
I = IMPORTS
AND SYNTHETICS
I
1955
1965 1975
YEAR
Figure 2. United States natural gas
supply and demand.
1985
l:: SYNTHETICS AND
SUBSTITUTE ENERGY
TECHNOLOGIES
WORLD OIL
PRODUCTION
CAPABILITY
CONSUMPTION
WORLD OIL
PRODUCTION
TIME
Figure 3. World petroleum supply and
demand: a gloomy scenario.
CONSUMPTION
(INCLUDING SUBSTITUTES)
WORLD OIL
PRODUCTION
CAPABILITY
WORLD OIL
PRODUCTION
TIME
Figure 4. World petroleum supply and
demand: another scenario.
0-3-12
-------
SESSION I:
Coal Combustion and Additive Regeneration
SESSION CHAIRMAN:
Mr. A.A. Jonke, Argonne National Laboratory
1-0-1
-------
1. BENCH-SCALE DEVELOPMENT
OF COMBUSTION AND ADDITIVE REGENERATION
IN FLUIDIZED BEDS
G. J. VOGEL, E. L. CARLS, J. ACKERMAN, M. HAAS, J. RIHA,
AND A. A. JONKE
Argonne National Laboratory
ABSTRACT
This paper discusses information obtained since the last Hueston Woods Conference on the
combustion of coal and oil with an excess of air and the combustion of coal in a deficiency of air.
The paper is also concerned with the thermodynamics of several proposed regeneration processes
and the regeneration of sulfur-containing additive by the two most promising processes — a one-
step reductive decomposition of CaSO4 and a two-step (reduction-CO2/H2O) procedure.
INTRODUCTION
Research on the combustion of fossil fuels
(particularly coal) in a fluidized bed of solids is
currently under investigation in the United
States and other countries. Most of the U.S.
effort is supported by the U.S. Environmental
Protection Agency (EPA), Office of Research
and Monitoring.
In applications of fluidized-bed combus-
tion, fuel is burned in a fluidized bed of solids
in which boiler tubes are immersed to take
advantage of the high heat-transfer char-
acteristics of fluidized beds. Additive, either
crushed limestone or crushed dolomite, can be
continuously fed to a fluidized-bed combustor
to react with SO2 released during combustion
and provide a means of in situ control of the
emissions of SO2.
Two different combustion modes are possi-
ble, one with complete and the other with par-
tial combustion of the fuel in the fluidized
bed. In the complete-combustion mode (also
called one-stage or oxygen-excess combus-
tion), oxygen in excess of the stoichiometric
amount required to burn the fuel to CO2 and
H2O is added to the fluidized bed. In the
second mode (called two-stage or oxygen-defi-
cient combustion), a stoichiometric deficiency
of air is added to the fluidized bed, and the
resulting H2, CO, and hydrocarbons are com-
busted to CO2 and H2O by providing
additional oxygen (air) either in the region
above the bed or in a separate combustor.
At Argonne National Laboratory (ANL),
data have been obtained on combustion of
coal and oil with an excess of oxygen and on
coal with a deficiency of oxygen. All experi-
ments have been made at atmospheric
pressure. The objectives of these experiments
have been as follows:
1. To determine how sulfur retention is
affected by independent fluidized bed
operating variables such as bed temper-
ature, gas velocity, oxygen concentration,
bed height, calcium to sulfur ratio, type of
additive and coal, and additive and coal
particle size.
1-1-1
-------
2. To determine the level of NO in the flue
gas at different operating conditions.
3. To obtain information on combustion effi-
ciency, combustion products, limestone
utilization, extent of calcination, decrepi-
tation rates.
4. To obtain information on the mechanism
of the lime sulfation reaction.
Efficient removal of SO2 from the gas
phase requires moderately large quantities of
limestone (compared to the quantity of coal
ash). When coal is burned, it will be desirable
to regenerate the partially sulfated lime. Ther-
modynamic calculations and experimental
data are presented on the two most promising
reactions; i.e., high temperature (~2100°F)
reductive decomposition of CaSO4 and a two-
step process, low temperature (~1600°F)
reduction of CaSO4 followed by reaction of the
CaS with CO2/H2O.
1-1-2
-------
BENCH.SCALE ATMOSPHERIC
COMBUSTION EXPERIMENTS
Materials, Bench-Scale Equipment, and
Procedure
Figure 1 is a schematic diagram of the
bench-scale fluidized-bed combustor system.
The combustor is a 6-in. diameter stainless
steel vessel. The fluidizing air enters the
combustor through a bubble-cap air distribu-
tor mounted on the bottom flange. Feeding
and metering of coal and additive is done by
variable-drive volumetric screw feeders
mounted on scales. The solids are fed pneu-
matically (entrained in a transport air stream)
into the fluidized bed at a point just above the
gas distributor.
Solids are removed from the off-gas by two
high-efficiency cyclone separators in series
and a glass fiber final filter. Downstream from
the cyclones, approximately 20 percent of the
total flue gas is diverted to a gas-analysis
system and its water content is reduced to
3000 ppm (by condensation and refrigeration)
to prevent moisture interfering with gas
analysis. Continuous analyses of the dried gas
for NO, SO2, CO, CH4, and O2 are conducted
with infrared analyzers and a paramagnetic
oxygen analyzer. Gas chromatography
provides intermittent analyses for CO2-
Periodically during a run, the bed and the
overhead solids are sampled to permit
chemical analysis and to obtain material
balances. All instrument signals, pneumatic
and electrical, are routed to a data logger
which produces a paper tape record for
further data processing and a typed output of
the signal values.
In a startup, the fluidized bed of
particulate solids is preheated to ~1000°F by
passing heated air through the bed and using
heaters mounted on the reactor wall. Coal is
then introduced into the bed and ignited,
increasing the bed temperature to the desired
operating temperature (e.g., ~1600°F). The
bed is maintained at a selected temperature by
passing air or an air-H^O mixture through
annular chambers on the exterior of the
combustor wall.
In one-stage combustion experiments (with
an oxidizing atmosphere in the bed), all
combustion air is introduced at or near the
bottom of the fluidized bed.
In two-stage combustion, a stoichiometric
deficiency of air is introduced at the bottom of
the bed to partially burn the fuel. All or most
of the oxygen in the air fed to the first stage is
consumed, and reducing conditions prevail in
the bed. In some ANL experiments, additional
air was introduced into the freeboard above
the bed through a tube located about 6 inches
above the fluidized bed; oxygen in this air feed
reacted with the CO, hydrocarbons, and the
unburned carbon elutriated from the bed.
The coals used in the various series of
experiments were : (1) Illinois coal from Seam
6, Peabody Coal Co. Mine 10, Christian
County, Illinois (furnished by Commonwealth
Edison); and (2) Pittsburgh Seam Coal from
the Humphrey Preparation Plant, Osage,
West Virginia. Sulfur contents of the coals (on
an as-received basis) were 3.7 and 2.4 weight
percent respectively. The as-received coal was
crushed to pass a -14 mesh sieve whereupon
more than 80 percent of the coal was in the
-14, + 170 mesh fraction. Limited additional
size reduction occurred as the coal passed
through the screw feeder.
A residual crude oil was obtained from
Esso Research and Engineering Co. Its sulfur
content was 1.9 weight percent, viscosity
(Seconds Saybolt Furol) was 162.5, and the
flash point was 178°C.
The natural.gas (obtained from Northern
Illinois Gas Company) had a heat content of
1035 Btu/ft3.
The additive materials studied include: (1)
limestone No. 1359, Stephen City, Virginia
(97.8 wt % CaCO3, 1.3 wt % MgCO3); (2)
limestone No. 1360, Monmouth, Illinois (69.8
wt % CaCO3,19.2 wt % MgCO3); (3) dolomite
1-1-3
-------
No. 1337, Gibsonburg, Ohio (53.4 wt %
CaCO3, 46.5 wt % MgCOs); and (4)
Tymochtee dolomite, Huntsville, Ohio (49.3 wt
% CaCOa, 36.6 wt % MgCO3).
Results and Discussion
1. Air Excess, Coal Combustion
Experiments
The first experiments discussed here were
made with sufficient fluidizing air introduced
at the bottom of the reactor so that the gas
leaving the fluidized bed contained unreacted
oxygen. A large-particle-size limestone was
fed, little of which was elutriated from the bed
(which consisted of calcined, partially sulfated
lime).
a. Effect of operating variables on SOi
retention — For this mode of operation, the
operating variables having significant effects
on sulfur retention were the Ca/S mole ratio in
the feed streams (the ratio of moles of calcium
in the additive to moles of sulfur in the coal),
the fluidized-bed temperature, and the super-
ficial gas velocity. (Sulfur retention is defined
as the percentage of the sulfur associated with
the coal feed that is not contained in the off-
gas as SO2-) Less significant variables were the
types of coal and additive, the particle sizes of
coal and additive, the height of the fluidized
bed, and the amount of excess air fed to the
fluidized bed. Three variables having no
demonstrable effect on sulfur retention were:
(1) premixing of coal and additive before they
were fed to the combustor (instead of feeding
separate streams of coal and additive), (2)
temperature of the gas in the freeboard above
the fluidized bed, and (3) addition of small
quantities of water to the fluidized-bed.
Ca/S mole ratio. Figure 2 shows the effect
of Ca/S mole ratio on sulfur retention at
1450°F for Pittsburgh coal and at 1550 and
1600°F for Illinois coal. Sulfur retention
increased as additive (Ca) feed rate was
increased in relation to the coal (S) feed rate.
Relatively good removals were attained at
Ca/S ratios above 3.
Fluidized-bed temperature. Experimental
results indicate that there is a temperature at
which sulfur retention is at a maximum over
the range of bed temperatures studied (Figure
3). With Illinois coal, a Ca/S mole ratio of
~2.5, and limestone No. 1359 additive, the
optimum bed temperature apparently was
1500-1550°F. With Pittsburgh coal, a Ca/S
mole ratio of~4.0, and the same limestone
additive, 1450-1470° F appears to be the
optimum temperature. Those results were
obtained in experiments using a gas velocity of
3 ft/sec and 3 percent excess O2 in the flue
gas.
The difference in optimum temperatures
may be associated with different properties of
the coal or alternatively, the temperature for
optimum sulfur retention may have been
influenced by Ca/S mole ratio. In any case, an
operating temperature of 1500°F would be
near optimum.
Superficial gas velocity. Sulfur retention
.was observed to increase with decreased
superficial gas velocity (in the range of 3.5 to
7.4 ft/sec) at a coal combustion temperature of
1550°F and with addition of limestone
No. 1359 (> 1000 fxm average particle size) and
Illinois coal at a Ca/S mole feed ratio of ~4
(Figure 4). The relatively coarse additive was
selected to ensure that additive particles would
be retained in the fluidized bed at high gas
velocities. At gas velocities of 3.5, 5.5, and 7.4
ft/sec, the average SO2 concentrations in the
flue gas were 770, 1250, and 1500 ppm,
corresponding to retentions of 83, 73, and 66
percent of the sulfur fed to the reactor. These
data may be correlated with the equation
where:
R = 101.79 e-
R = SO 2 retention, %
v = superficial gas velocity, ft/ sec
(1)
Results of British experiments (1) using
Welbeck coal, 440-nm British limestone, Ca/S
mole ratios of 1 and 2, and a coal-ash fluidized
1-1-4
-------
bed show that sulfur retention is greater at a
gas velocity of 2 ft/sec than at 3 ft/sec (Figure
4). The slopes decrease as the Ca/S mole ratio
increases; thus, at sufficiently high Ca/S
ratios, sulfur retention may be essentially
independent of superficial gas velocity.
Excess air. The oxygen level in the off-gas
was varied by adding pure oxygen at several
rates to the fluidizing air before it entered the
preheater. At 1550°F, ~3 ft/sec gas velocity,
and Ca/S mole ratio of ~3, the sulfur reten-
tions were 67, 71, and 75 percent, respectively,
at 0.7, 2.4, and 5.6 percent C*2 in the flue gas.
Apparently, oxygen concentration affects
slightly the reaction of SO2 with limestone,
and sulfur retention can be expected to in-
crease when oxygen concentration in the flue
gas is increased.
Fluidized-bed height. Runs were per-
formed at 1550°F, 3 ft/sec gas velocity, and
Ca/S ofM with three different bed heights —
14,24, and 46 inches (length to diameter (L/D)
ratios of 2.3, 4.0, and 7.7). The sulfur
retentions were 78, 80, and 83 percent,
respectively, indicating that bed height has a
small but real effect.
Type of coal. The effect of type of coal on
sulfur retention could not be evaluated
because of insufficient data in ANL ex-
periments. However, qualitatively, sulfur
retentions for Illinois and Pittsburgh coals
differed little. Work by the British and by the
U.S. Bureau of Mines has evaluated this
variable in greater detail.
Type of additive. Sulfur retentions for runs
performed at a Ca/S mole ratio of 2.5 and
fluidized-bed temperatures of 1550 or
1600°F with several types of additives may be
compared in Figure 2. Sulfur retention varied
about 10 percent indicating that differences in
these additive types had only small effects at
these operating conditions.
Size of coal particles. To determine the
effect of sulfur retention of the particle size of
the coal feed, two experiments were completed
with -12 +50 mesh and -50 mesh (a -12 +50
fraction ground to -50 mesh) Illinois coal at a
Ca/S mole feed ratio of 2.4 and a fluidized-
bed temperature of 1550°F. Sulfur retentions
by No. 1359 limestone additive were 81 and 75
percent, respectively, for the -12 +50 mesh
and -50 mesh feeds. An experiment in which
Illinois -14 mesh coal was burned under
operating conditions nearly identical to those
used in the above experiments yielded a 78
percent sulfur retention. Thus, similar sulfur
retentions were obtained by burning coal of
three particle size distributions, but the
coarsest coal feed appears to yield the best
results.
Size of additive particles. Sulfur retention
was calculated by interpolation to be ~87
percent for larger additive particles (1000 /xm
average) and 93 percent for smaller particles
(630 Mm average) in runs with a Ca/S mole
ratio of 4.0, a temperature of 1550°F, and a
gas velocity of 3 ft/sec. This suggests that at
least in the region of high sulfur retention,
additive particle size in this range has only a
moderate effect on sulfur retention.
b. NO levels in the flue gas — Nitrogen
oxides, principally nitric oxide (NO), are
formed during the combustion of fossil fuels
and are an important contributor to air
pollution. Although the quantities of NO
observed in the flue gas from high-
temperature conventional combustors may be
accounted for by the equilibrium of the
nitrogen fixation reaction, this is not the case
for low-temperature fluidized-bed coal
combustion in which NO concentrations far in
excess of those expected on the basis of the
equilibrium have been observed. At 1600° F, a
common temperature for fluidized-bed
combustion, the equilibrium concentration of
NO from the fixation of atmospheric nitrogen
ranges from 50 to 200 ppm, depending on the
oxygen concentration. (The oxygen con-
centration, in turn, depends on the level of
excess air employed.) However, in actual fluid-
bed coal combustion, nitric oxide levels of 400
to 800 ppm in the flue gas have been measured
with oxygen concentrations of ~3 volume
1-1-5
-------
percent in the off-gas (15 to 20 percent excess
air).
In previous ANL work, the major source of
nitric oxide during the combustion of coal was
determined to be the nitrogenous content of
the coal (1 to 1.5 weight percent in U.S. coals).
In more recent work, the effect of moisture
content of the coal on NO level was studied by
adding water at several rates to the fluidizing
air. Pittsburgh coal and limestone No. 1359
were the feed materials. The fluidized-bed
temperature was 1450° F and the Ca/S mole
ratio was 1. The concentration of NO
decreased from 530 ppm to 510 ppm when 10
cm3/min water was added (equivalent to 26
weight percent water in the coal), and to 380
ppm when the rate of water addition was
further increased to 30 cmVmin (equivalent to
51 weight percent water in the coal). These
decreases in NO concentration may be due to
chemical reduction of NO by hydrogen pro-
duced by the water-gas-shift reaction, but the
effect is not great enough to warrant further
attention.
c. Calcium utilization — The relative
extent of conversion of CaO to CaSO4 for bed
and elutriated materials was calculated from
calcium and sulfur concentrations determined
by wet chemical analysis. In several experi-
ments with No. 1359 limestone of an average
particle size of 490 jum, at temperatures
ranging from 1400 to 1600°F, and gas
velocities from 2.5 to 2.8 ft/sec, the conversion
of calcium oxide to calcium sulfate was
highest (2/3 converted) for particles collected
on the final filter. These particles have a high
surface to volume ratio and would be expected
to react rapidly even though their residence
time in the bed is relatively short. Next highest
conversion, ~2/5, was obtained in bed par-
ticles which have a relatively long residence
time in the bed. Lowest conversion, ~l/4, was
obtained with solids removed from cyclones.
These particles are larger than final filter
particles, smaller than the bed particles, and
are present in the bed for a relatively short
time.
2. Air Deficient, Coal Combustion Ex-
periments
The concept of two-stage combustion
provides for a substoichiometric quantity of
air (that is, less air than is required to burn the
coal completely to COa and H2O) introduced
into the first stage of the fluidized bed to
which coal is fed. Additional air may be in-
jected into the disengaging section above the
fluidized bed (the second stage) to burn
gaseous hydrocarbons, Ha, and CO in the gas
stream from the first stage.
Two-stage combustion, experiments of an
exploratory nature were conducted to
determine if this combustion mode might have
benefits, as compared with single-stage
fluidized-bed combustion. To simulate the
conditions of combustion in the first stage
only, experiments were performed in which a
substoichiometric quantity of air was intro-
duced into the bottom of the fluidized bed, but
no secondary air was fed. In other experi-
ments, secondary air was introduced above the
fluidized bed. The bed consisted of coarse lime
particles in all of these experiments.
The experimental results include in-
formation on the concentrations of SOa, HaS,
NO, and CO in the off-gases at various air feed
rates and bed temperatures, as well as data on
the sulfur content of solid products when a
substoichiometric quantity of air was fed to
the first stage only.
a. Effect of decreasing air input on ratio of
HaS to HaS + SO2 in flue gas — The concen-
tration of Ha S in the off-gases was measured
to determine which operating conditions affect
the formation of this sulfur compound. In
experiments in which air was introduced to the
first stage only, the amounts of H2S and SOa
in the off-gas were compared. The percentage
of sulfur in the off-gas as HaS was sensitive to
the amount of air introduced into the fluidized
bed, increasing drastically when the air feed
rate was reduced below a value corresponding
to ~70 percent of the stoichiometric quantity
of air necessary to react with the coal feed (see
Figure 5). (Although the parameter, air feed
1-1-6
-------
rate as a percent of stoichiometric, was based
on feed rates of coal and air, it is recognized
that the quantity of coal actually oxidized
varies with other parameters (i.e., tempera-
ture, etc.). For certain correlations, it may well
be more suitable to use the parameter, stoichi-
ometric air feed rate based on the coal actually
oxidized.) At an air feed rate equivalent to
~50 percent of the stoichiometric quantity,
the concentration of H2S (611 ppm) was
nearly equivalent to the concentration of SO2
(660 ppm). At air inputs of 70 to 80 percent of
the stoichiometric quantity, the relative
amount of sulfur as H2S fell to about 2 percent
of the total sulfur in the gas.
In those experiments in which secondary
air was introduced above the fluidized bed
(Figure 5), the E^S level in the off-gas was low
— corresponding to less than 1 percent of the
total sulfur in the gas. This suggests that any
H2S in the gas leaving the first stage is
oxidized to SOa by air fed to the second stage.
No consistent relationship was apparent
between H2S level and either the temperature
of the fluidized bed (1450-1650°F) or the
temperature of the off-gas in the freeboard
above the bed (1100-1800°F).
b. Sulfur retention — Sulfur retention is
defined as the percentage of the sulfur
associated with the coal feed but not contained
in the off-gas as SO2 or H2S. (Because a
fraction of the carbon was not burned in these
experiments, the sulfur retention values given
are probably higher than would be realized if
all of the carbon were burned.) Experiments
were performed with no introduction of
secondary air (Figure 6) to determine sulfur
retention as a function of the Ca/S mole ratio
in the feed at 1450, 1550, and 1650°F. Also
shown in Figure 6 (to allow comparison) are
data for experiments carried out earlier under
one-stage oxygen-excess conditions at 1450-
1470° F.
The data presented for the sub-
stoichiometric air experiments show a large
amount of scatter principally due to variation
in the quantity of air fed to the fluidized bed,
which was not the same in all experiments.
Stoichiometric air added in each experiment
ranged from 51 to 91 percent. The best
retentions were observed at stoichiometric
additions of less than 60 percent.
For experiments carried out at a Ca/S ratio
of about 2 and temperatures of 1450-1650° F
(Figure 7), no simple relationship between the
amount of air introduced into the bed and
sulfur retention was evident; however, a line
has been fitted to the points as shown. At an
air feed rate of 100 percent of stoichiometric,
sulfur retention is about 65 percent. As the air
feed rate is decreased, sulfur retention first
decreases to about 45 percent as the air rate is
decreased to 75 percent of the calculated
stoichiometric requirement and then increases
rapidly as the air rate is decreased further.
This suggests that in an oxygen-deficient
region (75 to 95 percent of calculated stoichio-
metry), removal of sulfur by lime in the form
of SO2 is poor, but that at lower air flow rates
sulfur is in the form of H2S and is efficiently
removed. This would be expected because
oxidizing conditions are required for the
retention of SO2 by lime (to convert a CaSC>3
intermediate to CaSC>4), whereas reducing
conditions are required for the retention of
H2S by lime.
The introduction of secondary air above the
bed ^resulted in erratic but generally lower
sulfur retentions. Decreases were about 10
percent at 1450°F, 20 percent at 1550°F, and
40 percent at 1650°F. The increased sulfur
content of the off-gas after secondary air was
introduced was probably caused by burning of
entrained coal particles in the second stage to
produce additional SCh.
c. NO levels in the flue gas — When coal
was burned with a deficiency of air fed to the
first stage, concentrations of NO in the off-gas
from the first stage as a function of the
amount of air introduced into the bottom of
the fluidized bed were as shown in Figure 8.
To obtain these data, only the first stage was
operated. The NO concentrations were
1-1-7
-------
generally < 250 ppm and apparently were
affected by both the amount of air introduced
and the temperature of the fluidized bed. At
the lower air feed rates, the NO levels were
generally lower. At a given air feed rate, NO
levels were higher at lower bed temperatures.
Since earlier work at ANL showed that
nitrogenous compounds in coal are oxidized to
NO during fluidized-bed combustion, it can
be postulated that the lower levels of NO
observed at higher temperature are due to
more rapid decomposition of NO. This
decomposition may be promoted by the
presence of CO; another possibility is that
nitrogenous compounds other than NO may
be formed in the highly reducing atmosphere
of the bed.
The data presented in Figure 8 are not
corrected to an equivalent off-gas volume
basis. However, if this correction were made,
the dependence of NO emissions on air feed
rate would be even more pronounced,
assuming that feed rates of coal were
equivalent.
Upon the introduction of secondary air
above the fluidized bed, NO levels in the off-
gas varied eratically — usually increasing.
Possible explanations for this behavior are: (1)
any reduction of NO by CO in the zone above
the bed would be suppressed by introducing
secondary air or (2) if a nitrogen compound
such as ammonia were present in the gas, it
may be oxidized to NO by the secondary air.
d. Sulfur species in the bed — Results show
that the sulfide content of bed material
decreased as air flow was increased. At air
inputs corresponding to 50 percent of
stoichiometric, as much as 100 percent of the
sulfur in the bed was sulfide. In most experi-
ments in which the air input exceeded 65
percent of stoichiometric, sulfide content
dropped off rapidly to less than 1 percent. No
relationship was found between sulfide
content and bed temperature.
The sulfite content of bed samples was
erratic, ranging between 6.2 and < 0.1 weight
percent. No correlation of sulfite content with
either bed temperature or air feed rate could
be found.
e. Carbon balance — Carbon balances were
made for three experiments. For three other
experiments, all data for making the balances
except the COa level in the flue gas are
available (see Table 1). The small quantity of
carbon not accounted for is represented by
hydrocarbons (other than CH4) for which no
analyses are made and by a small loss of fine
carbon participate from the combustion
system. The data show that as the volume of
air added to the bed decreases (experiments
14-1A, -IB, -2) at the same temperature, the
CO content of the flue gas increases markedly,
the CH4 content increases slightly, and the
quantity of carbon elutriated to the first and
second cyclone separators from the fluidized
bed increases.
The carbon content of the bed under low
stoichiometric air additions (55 percent) was
as high as 31 percent (experiment 14-3B).
Under these conditions the amount of carbon
elutriated was about 15 percent of that fed.
f. Preliminary evaluation of the concept —
Although the work conducted on two-stage
combustion was exploratory in nature, a
preliminary evaluation of the concept can be
made. The principal advantages of the two-
stage combustion concept over one-stage
combustion are: (1) lower NO emissions; (2)
retention of sulfur in the form of calcium
sulfide (rather than sulfate), allowing for
potentially easier regeneration of the additive;
and (3) production of a combustible gas that
could be used in conjunction with a gas
turbine.
The principal disadvantages are: (1) greater
elutriation of carbon, (2) possible compli-
cations in additive regeneration owing to the
high carbon content of the bed, and (3)
necessity for removal of heat from the bed
under conditions that might be corrosive to
immersed steam tubes.
1-1-8
-------
Table 1. CARBON BALANCES AND CARBON CONTENTS OF SOLIDS IN
• SUBSTOICHIOMETRIC AIR-COMBUSTION EXPERIMENTS
Carbon in coal, g/hr
Carbon out, g/hr
First cyclone
Second cyclone
Flue gas
CH4
CO
CO 2
Total carbon out, g/hr
Carbon concentration
in solids streams, wt%
Bed
First cyclone
Second cyclone
Run conditions
Temperature, °F
Coalfeed,lb/hr
Air, % of stoichiometric
14-1A
1428
71
22
>24
42
1230
>1470
<1
29
52
1450
5.0
90
14-1 B
1428
136
11
32
350
NDa
ND
<1
39
53
1450
5.0
86
14-2
1684
174
24
39
441
840
1632
6
46
55
1450
6.3
71
14-1C
1799
167
7
42
521
ND
ND
<1
39
56
1550
5.9
54
14-3A
1485
235
7
28
330
ND
ND
20
47
35
1600
5.2
64
14-3B
1713
233
16
32
455
800
1682
3.1
62
53
1600
6.0
55
No data available.
1-1-9
-------
Sulfur retention appears to be roughly
equal for the two concepts. It is notable, also,
that no problems of coal caking were en-
countered even at high bed carbon contents.
Further work might be warranted at lower
air addition rates and higher bed tem-
peratures to avoid heat generation in the bed.
Under such conditions the process would
become gasification rather than combustion.
3. Air Excess, Oil Combustion Experiments
To assess the removal of SOi from com-
bustion gases when residual fuel oil is burned
in a fluidized bed of sulfated lime with con-
tinuous feeding of limestone additive, experi-
ments were performed in the 6-in. diameter
fluidized-bed combustor at a variety of oper-
ating conditions. Residual fuel oil was burned
in an excess of oxygen at bed temperatures
ranging from 1450 to 1650°F, Ca/S mole
ratios up to 11.9, a gas velocity of ~3 ft/sec
(except for one experiment at 5.5 ft/sec), and
with 3 volume percent oxygen in the flue gas
(except in one experiment with 1 volume per-
cent oxygen in the flue gas).
The following results were obtained in
these experiments.
1. The effect of temperature on sulfur
retention is similar to that observed in coal
combustion experiments, in which there is a
temperature yielding maximum sulfur
retention. In the oil-combustion experiments,
maximum sulfur retention was at 1500-
1550°F.
2. The shape of the curve for sulfur
retention as a function of Ca/S_ mole ratio is
similar to that obtained in coal combustion
experiments. Sulfur retention in the oil
combustion runs increases as Ca/S mole ratio
increases to about 5, then levels off at a 90
percent sulfur retention level as the Ca/S ratio
is increased further. The slope of the curve for
sulfur retention as a function of Ca/S ratio is
less steep than the slope for Illinois coal (3.7
weight percent sulfur) at similar operating
conditions.
When oil was combusted, the NO levels in
the flue gas ranged from 110 to 150 ppm for
the experiments with 3 percent O2 in the flue
gas. This may be compared with the 400 to 800
ppm range observed when coal was burned.
However, the nitrogenous content of residual
oil is also less than that of coal. No correlation
of NO level with bed temperature or Ca/S
mole ratio was observed. Combustion ef-
ficiency in these experiments is discussed in
the section of combustion efficiencies below.
4. Miscellaneous
a. Additive decrepitation rates during coal
combustion experiments — Decrepitation ^and
attrition of several additives during coal
combustion experiments has been estimated
from the calcium content of elutriated fines.
(The fraction of additive carried over can only
be estimated because the particle matter
elutriated during the combustion of coal in a
fluidized bed is a mixture of solids of different
origins and compositions.)
In most experiments a gas velocity of ~2i6
ft/sec was used; at this velocity all of the flyash
and additive particles having diameters of
<177 nm are.expected to elutriate from the
fluidized bed.
The expected elutriation for each of several
series of experiments calculated in this
manner is shown in Table 2. The actual elutri-
ation was determined from calcium material
balances (with an allowance made for the
calcium content of the flyash, which was also
expected to elutriate). The difference between
actual and expected elutriation gave an
estimate for the decrepitation of larger ad-
ditive particles (Table 2).
The indicated decrepitation of BCR-13.59
limestone was ~8 percent, but no decrepi-
tation of a British limestone was evident.
Decrepitation of limestone BCR-1360 and
dolomite BCR-1337 was more severe—40 arid
85 percent, respectively. These results indicate
that decrepitation of BCR-1359 and British
limestone is low and that limestones of this
1-1-10
-------
Table 2. ESTIMATED DECREPITATION OF ADDITIVE MATERIALS
FROM FLUIDIZED-BED COMBUSTOR
Experiments
Amer-1 ,-3,-4
BC-6,-7,-8
AR-1,-2,-4,-5,-6 ,
BC-9
BC-10
Brit-1,-2,-3,-3A
and Amer-Brit
Additive type
BCR-1359
BCR-1359
BCR-1359
BCR-1360
BCR-1337
British limestone
Distribution of total calcium
in combustor system, wt %
Expected
elutriation3
12
2
13
2
<2
37
Actual
elutriation b
19
11
21
42
~87
33
Estimated
decrepitation
7
9
8
40
85 c
Calcium contained in particles < 177/4 m diameter in the additive feed to the system.
These particles are expected to elutriate at a superficial velocity of 2.6 ft/sec.
Fine particles in the starting fluidized bed are not included.
bCalcium fed with the coal was deducted from the total calcium found in the elutriated material.
c Derived from both calcium and magnesium material balances.
type are desirable materials for use in a full-
scale fluidized-bed combustor with regen-
eration and recycle of additive. Higher decrep-
itation rates for BCR-1360 and BCR-1337 may
make these materials less promising for regen-
eration and recycle. Data are for one cycle of
use only.
b. Cyclone collection efficiencies during
coal combustion experiments — Data on the
particle removal efficiency of the ANL cyclone
separators have been compiled as a basis for
estimating the dust loading and filter area of a
cartridge filter for a pressurized combustion
bench-scale plant now being designed. In the
present atmospheric pressure system, flue gas
passes through two cyclones in series and a
final filter. (The diameter of the first in-line
cyclone is 6-5/8 inches; the diameter of the
second is 4-1/2 inches.) It is planned to use the
same two cyclones with the pressurized
combustor.
To determine the adequacy of the glass
fiber mat filters used in the atmospheric plant,
collection efficiencies (defined as ratio of the
weight of particles removed to the weight of
particles entering the cyclone) were compiled
for 26 earlier ANL experiments. In these one-
and two-stage runs, the flue gas flowrates
ranged from 8 to 14 ft3/min, the coal feed
rates from 4 to 7.3 Ib/hr, the additive feed
rates from 1.1 to 2.3 Ib/hr, and the dust
loadings at the -combustor exit from 0.16 to
1.78 g/ft3. Combined efficiency of the two
cyclones was above 80 percent in 24 of the 26
experiments and above 90 percent in 21 of the
experiments. The dust loading in the flue gas
leaving the second cyclone averaged 0.06 g/ft3
for the 26 runs; the maximum loading was
0.22 g/ft3.
1-1-11
-------
c. Combustion efficiencies for coal, oil and
gas with excess air — The combustion ef-
ficiency for experiments performed in the
combustor was determined as the ratio of
carbon burned to carbon fed, multiplied by
100. The carbon loss is calculated by deter-
mining unburnt carbon leaving the system by
three routes: (1) carbon associated with the
elutriated solids; (2) incompletely burned
gases (e.g., carbon monoxide and hydro-
carbons); and (3) carbon associated with
fluidized-bed material taken from the system.
All experiments were conducted without
recycle of fines.
Combustion efficiencies in ten experiments
with coal ranged from 93 to 96 percent. In all
experiments, carbon losses in the bed material
were negligible. Only about 10 to 20 percent of
the carbon loss was due to the formation of
carbon monoxide and hydrocarbons. The
major carbon loss (80 to 90 percent) occurred
as a result of elutriation of fine particles in the
exhaust gases before they were completely
combusted. Combustion efficiency can be
increased by recycling the elutriated ash-
carbon mixture to the fluidized bed or to a
carbon burnup cell. Oxygen concentration in
the flue gas in these experiments was ap-
proximately 3 percent.
Combustion efficiency in oil combustion
experiments was similar to that observed for
coal combustion experiments under similar
conditions, ranging from 94 to 96 percent for
experiments with 3 percent O2 in the flue gas.
However, sources of carbon losses for the two
fuels differed. In coal combustion, most of the
carbon loss is represented by the carbon
content of solids elutriated to the cyclones; in
oil combustion, inefficiency results from
incomplete burning of the CO and hydro-
carbons formed during combustion. Efficiency
can probably be improved by operating the
combustor with a deeper bed or by increasing
the freeboard temperature. Lower combustion
efficiencies were observed with less excess
oxygen in the flue gas and at higher gas
velocities.
During the combustion of natural gas, the
elutriation of carbon-bearing fine particles is
negligible, and the major loss of unburnt
carbon is in the carbon monoxide and total
hydrocarbons in the flue gas. Combustion
efficiencies, calculated from analyses of
samples of the flue gas, ranged from 94.1 to
99.2 percent at 1600°F with 3 percent excess
oxygen. At 1800°F, combustion efficiencies of
94.1 to 98.8 percent were observed. These
results are similar to data reported by the
USSR on combustion of gas in a fluidized bed.
Although the USSR data indicate that
combustion efficiency is principally affected
by bed temperature, combustion efficiency is
also likely to be a function of bed depth, gas
velocity, and excess oxygen concentration. For
example, in experiment NG-3 at 1800°F,
combustion efficiency was decreased to 91
percent when the combustion was in-
tentionally forced toward more reducing
conditions; i.e., 1.5 volume percent O2 in the
flue gas rather than 1.8-4.5 volume percent
02.
REGENERATION OF SULFUR CON-
TAINING ADDITIVES
When coal is burned in a fluidized bed
containing limestone or dolomite, sulfur-
containing gases from the combustion of
sulfur-containing substances in the coal react
with the bed material and are retained in the
bed. The reaction product is calcium sulfate if
combustion is carried out under oxidizing
conditions, or calcium sulfide if combustion is
carried out with a deficiency of air.
Several regeneration processes are under
consideration. These are:
1. Reductive decomposition of calcium
sulfate.
2. Roasting of calcium sulfide in air or
oxygen.
3. Reaction of calcium sulfide with water
and carbon dioxide.
Processes 2 and 3 can be used to regenerate
not only calcium sulfide, but also material
containing calcium sulfate if the sulfate is first
reduced to the sulfide.
1-1-12
-------
Thermodynamic Analyses
Valuable information can be gained by
considering the thermodynamics of the
process. The yields of gaseous sulfur-
containing products, the composition of solid
phases, and the variations of these yields and
compositions with temperature, pressure, and
gas composition for a system at equilibrium
can all be obtained. Optimum reactant feed
ratios and gas compositions can also be
calculated easily when product concen-
trations, compositions, and pressures are
specified.
All of the following predictions and con-
clusions are based on the supposition that
chemical equilibrium is achieved among the
various phases. This implies that the rates of
all relevant chemical reactions are large on the
time scales being used, which scales are
determined by mass transport rates within the
system. The maximum rates at which this
supposition is valid vary with temperature and
must be determined in the laboratory and in
the pilot plant. It is further assumed that the
system is not stoichiometrically limited. There
must always be at least small amounts of the
appropriate solid phases present for the
results of these calculations to be valid. One
must not assume that all actual processes will
be operated with all these solid phases present,
however.
The assumption is also made that solid
solutions do not form to any great extent.
Exploratory experiments to date support this
assumption.
1. Reductive Decomposition of CaSO4 with
CO/CO2
Before the relative amounts of the species
in an equilibrium mixture from the reduction
of calcium sulfate with carbon monoxide-
carbon dioxide mixtures can be calculated, the
solid phases present at the various conditions
of temperature and carbon monoxide/carbon
dioxide ratio must be determined. The
possible sulfur-containing solids are con-
sidered to be calcium sulfate, calcium sulfite,
and calcium sulfide.
a. Conditions for the presence of calcium
sulfate and calcium sulfide — The solid
phases present at equilibrium with a
PCO /Pco2 °f 0.005-0.055 and temperatures
of 1600 to 2400°F are shown in Figure 9 (in
which temperature is the ordinate and
Pco /PCO2 the abscissa). Examination of the
expression for Kp in reaction 2
1/4 CaSO4 + CO - 1/4 CaS + CO2
C02
CO
(2)
shows that for any temperature, there is but
one ratio of carbon monoxide to carbon
dioxide at which calcium sulfate and calcium
sulfide can coexist at equilibrium. The co-
existence conditions appear as the line run-
ning from the lower left to the upper right part
of Figure 9 and represent CO/CO2 ratios at
which CaS and CaSO4 can both be present at
equilibrium. In the area to the right of this
line, the gas mixture is so rich in carbon
monoxide that calcium sulfate is completely
reduced to calcium sulfide. To the left of the
line, the gas mixture is so rich in carbon
dioxide that calcium sulfide is completely
oxidized to calcium sulfate. This line is called
the coexistence line for calcium sulfate and
calcium sulfide.
b. Conditions for presence of calcium
sulfite — Calcium sulfite is not stable in the
presence of CO/CO 2 mixtures at any tem-
perature from 1500 to 2400°F. This has been
established by plotting "coexistence" lines for
calcium sulfite with calcium sulfate and for
calcium sulfite with calcium sulfide. These are
analogous to the calcium sulfate-calcium sul-
fite coexistence line described above and are
determined in the same way from the equilib-
rium constants for reactions 3 and 4.
CaSO4 + CO - CaSO3 + CO2
C02
(3)
1-1-13
-------
1/3 CaS + CO 2-1/3 CaSO3 + CO
(4)
c. Conditions for the presence of calcium
carbonate and calcium oxide — When
calcium sulfate is reduced or calcium sulfide is
oxidized by a mixture of CO and CO2,
calcium oxide is formed. However, in the
presence of carbon dioxide at sufficient
pressure, calcium oxide is converted to
calcium carbonate.
A coexistence line for the carbonate and
the oxide is determined by the equilibrium
dissociation pressure of calcium carbonate
(reaction 5).
CaC03 ^ CaO + CO2
Kp =PC02
(5)
It also appears as a nearly horizontal line at
about 1950°F in Figure 9. This line represents
the temperature at which the partial pressure
of COa in the equilibrium mixture just equals
the equilibrium pressure of CO2 over calcium
carbonate.
The partial pressure of CO 2 in the
equilibrium mixture is obtained by assuming a
total pressure of 10 atm and subtracting the
pressures of SO2 and CO. Clearly, if the total
pressure is lowered or if an inert gas is added,
the pressure of CO2 will be lower and the
horizontal line will be at a lower temperature.
It is also clear that the calcium carbonate does
not exist above the horizontal line and that
calcium oxide does not exist below it.
d. Sulfur dioxide pressure — The pressure
of sulfur dioxide in the equilibrium mixture
can be calculated from the CO/CO 2 ratio and
the equilibrium constant of the reaction
appropriate to the part of Figure 9 under
consideration; however, in the areas labeled C
and D, one must generate independent in-
formation about the CO2 pressure by making
assumptions exactly analogous to those made
in the above discussion of calcium carbonate.
In area A of Figure 9, SO2 is generated by
reaction 6.
CaS04 +CO ^ CaO + SO2 + CO2
Kp =-
SO2 • CO2
pco
(6)
The pressure of SO2 is shown as a family of
isobars slanted down toward the right. In area
B, SO2 is generated by the oxidation of
calcium sulfide in accordance with reaction 7.
1/3 CaS + CO2 ^ 1/3 CaO + CO + 1/3 SO2
Lco
C02
(7)
The isobars of constant SO2 pressure in area
B curve down to the left, meeting those of area
A at the calcium sulfide-calcium sulfate co-
existence line. At any temperature, the SO2
pressure is at a maximum at this junction. For
example, at 2000° F, the maximum attainable
equilibrium pressure of SO2 is 0.46 atm at a
CO/CO2 ratio of 0.020. This maximum in the
SO2 pressure may be understood by
examination of the appropriate equilibrium
constants. For example, the expression for the
equilibrium constant for reaction 6 predicts
that the pressure of SO2 is directly propor-
tional to the CO/CO2 ratio. Thus, the
pressure of SO2 must increase as the CO/CO2
ratio increases as long as reaction 6 obtains.
From the expression for the equilibrium
constant in reaction 7, it may be seen that the
SO2 pressure is inversely proportional to the
cube of the CO/CO2 ratio. Thus, the SO2
pressure increases with decreasing CO/CO2
ratio as long as reaction 7 obtains. Reactions 6
and 7 occur simultaneously only along the co-
existence line. Thus, as one moves away from
the coexistence line, the SO2 pressure must
decrease.
1-1-14
-------
In area C, reactions applies.
CaSO4 + CO=^CaCO3 + SO2
P
KP =
SO-
CO
(8)
In this area, the pressure of SO2 is dependent
only on Kp and carbon monoxide pressure,
and the SO2 isobars are nearly vertical. The
main effect results from the variation of Kp
with the temperature. In area D, the SO2
pressure is once again a strong function of the
CO/CO 2 ratio as may be seen from reaction 9.
CaS + 4 CO2 ^CaCO3 + SO2 + 3 CO
'SO.
(9)
The slope of the isobars in area D differs
only slightly from the slope in area B as a
result of increased dependence on CO2
pressure in area D.
e. Sulfur pressure — The pressure of sulfur
vapor is quite low (<10~2 atm) over the
temperature range 1700 to 2300 °F. The
pressure of sulfur in area B was calculated
from reaction 10.
CaS + CO2 -1/2 S2 + CO + CaO
(10)
Since the formation of sulfur is entirely
analogous to the formation of SO2, sulfur
concentration may be expected to exhibit the
same sort of maximum at the calcium sulfide-
calcium sulfate coexistence line.
f. Carbonyl sulfide pressure — The car-
bonyl sulfide pressure was calculated with
reaction 11,
CaS + CO 2- CaO + COS
K«
COS
rCO,
(11)
Since P^OS *s dependent on the pressure of
CO2, assumptions made in calculating COS
pressure are similar to those made in the
discussion of calcium carbonate.
(~10"3atm) in this system.
g. Solid-solid reaction of calcium sulfide
with calcium sulfate — SO2 is generated by
the reaction of calcium sulfide with calcium
sulfate, as is shown in reaction 12.
1/3 CaS + CaSO4 ^4/3 SO2 + 4/3 CaO
4/3
KP = (PSO J
(12)
Since reaction 12 is exactly equivalent to the
sum of reactions 6 and 7, the SO 2 pressure
calculated from reaction 12 must be just that
calculated from reaction 6 or reaction 7 using
the CO/CO2 ratio at the coexistence line.
Another way of saying this is that the presence
of both calcium sulfide and calcium sulfate
determines an oxidizing potential for the
atmosphere with which it is in equilibrium;
this oxidizing potential determines the
CO/CO 2 ratio of the atmosphere. If carbon
monoxide and carbon dioxide are present in
the gas phase over a mixture of calcium sulfide
and calcium sulfate, they serve as a facile
route to the production of SO2 so that rapid
reaction rates for mixtures of the two solids
are possible.
h. Reduction of calcium sulfate with
H2/H2O mixtures — The system calcium
sulfate-calcium sulfide-H2-H2O is exactly
analogous to the system calcium sulfate-
calcium sulfide-CO-CO2. This means that all
the features of the CO-CO 2 system are present
in the H2-H2O system. The sulfur-containing
solid phases once again are calcium sulfate
and calcium sulfide, but calcium oxide and
1-1-15
-------
calcium hydroxide are the non- sulfur-
containing solid phases. The pressure of SO 2
at a given temperature has a maximum value
where the calcium sulfate and calcium sulfide
are in equilibrium with the H2-H2O mixture.
The instability of calcium sulfite can be shown
in the same way as in the CO-CO2 system. In
fact, the major differences between the
H2-H2O system and the CO-CO2 system are
that carbonyl sulfide is replaced with H2S and
that at any given temperature, the numerical
value for the H2/H2O ratio differs from that
of the CO/CO2 ratio.
An additional difference between the
systems is that in the H2-H2O system, calcium
hydroxide can form at lower temperatures and
higher pressures of H2O (just as CaCOa can
form in the CO-CO2 system). However, at 10-
atm H2O pressure, Ca(OH)2 is not stable
above 1200° F.
For any temperature and SCh pressure, the
H2/H2O ratio can be calculated from the
equivalent CO/CO2 ratio via reaction 13.
H2 + CO2 ^
P / P
*CO CO,
AJ2 "^ (13)
This is the familiar water-gas shift reaction.
The principle involved here is that equal SO 2
pressures are obtained in the Hj -H2 O system
and the CO-CO2 system when the oxidizing
potentials of the atmospheres are the same;
i.e., when the two atmospheres are in equil-
ibrium with each other.
An important conclusion is that the
maximum pressure of SO2 from any system in
which calcium sulfate is reduced or calcium
sulfide is oxidized is the pressure of SO 2
observed along either the H2 -H2O coexistence
line or the CO-CO2 coexistence line. The same
is true for 82 pressures. The basis for these
rather far-reaching conclusions is that in any
process involving a reduction of calcium
sulfate, the SO2 and 82 pressures will increase
with increasing reducing ability of the atmos-
phere until calcium sulfide is formed. At that
point, increasing the reducing ability of the
atmosphere no longer increases the amount of
SO2 or 82 formed, but rather causes the
calcium sulfate to be transformed into calcium
sulfide. Similarly, in a process in which
calcium sulfide is oxidized, the SO2 and S2
pressures increase with increasing oxidizing
ability of the atmosphere until calcium sulfate
is formed. This point is again a limit, and SO2
and 82 pressures cannot be increased further.
2. Roasting of Calcium Sulfide
In the roasting process, calcium sulfide is
oxidized with oxygen or air according to
reaction 14.
CaS + 3/2 O2 - SO2 + CaO
(
3/2
(14)
As may be seen by examining the equilibrium
constant for reaction 14, the pressure of SO 2
at any temperature increases with increasing
pressure of oxygen. However, it follows from
the arguments presented above that above
some definite oxygen pressure (given at any
temperature by Kp of reaction 15),
1/2 CaS + O 2 -*l/2 CaSO4
(15)
calcium sulfide is no longer stable, but is
converted to calcium sulfate. At this particular
oxygen pressure, the SO2 pressure is that
observed along the coexistence line in the CO-
CO2 system or in the H2-H2O system. Thus,
what appear to be two very different processes,
the reductive decomposition of calcium sulfate
and the roasting of calcium sulfide, are in fact
very similar. Both processes give rise to
identical maximum SO 2 pressures at a given
temperature.
1-1-16
-------
3. Pressure Effects in the Above Processes
In all regeneration processes discussed
above, the pressure of SO2 is a function of the
temperature or of the oxidizing ability of the
atmosphere (in the case of the roasting and
reductive decomposition processes). The pres-
sure of SO 2 as a function of the total system
pressure has not been discussed because SO2
pressure is independent of the total system
pressure in these processes. However, percent
of SO2 in the gas mixture is an inverse func-
tion of the total system pressure since the
pressure of SO2 is fixed at any temperature.
The pressure of SO2 is also independent of the
presence of inert gaseous diluents if sufficient
oxidizing or reducing gas is present. As stated
above, in the case of reductive decomposition
with CO-CO2 mixtures, the presence of inert
gases may affect the CO2 pressure enough to
change the reaction product from calcium
carbonate to calcium oxide.
4. Acid-Base Reaction of Calcium Sulfide
with H2O and CO2
Reaction 16 has been proposed as a
regeneration reaction for calcium sulfide.
CaS + H2O + CO2 «k CaCO3 + H2S
H2S
KP ~ P
C02' PH2O
(16)
Calcium sulfide is formed in the additive by:
(1) burning coal in a fluidized bed of limestone
or dolomite with a deficiency of air or by (2)
reducing the CaSO4 in the additive from a run
in which combustion was with an excess of air.
Unlike all other regeneration reactions dis-
cussed here, this reaction is pressure-sensitive.
The percentage and the pressure of H2S
increase with increasing total system pressure.
The pressure of H2S is also sensitive to the
presence of inert-gas diluents, in contrast to
the previously mentioned regeneration
schemes. The equilibrium constant for this
exothermic reaction becomes smaller as the
temperature is increased (Figure 10). This is in
direct contrast to the other (endothermic)
regeneration schemes mentioned.
Maximum H2S yield is obtained when the
H2O/CO2 ratio in the feed gas is 1 to 1, as
may be seen by an examination of the equil-
ibrium constant expression. Figure 11 shows
the pressure of H2S as a function of tem-
perature, assuming 10-atm total pressure and
an inlet gas stream composed only of H2O and
CO2 at various ratios. However, H2O may be
readily removed from the product gas stream
by condensation. Thus, higher values of H2S
concentration in a dried gas stream may be
obtained by operating with an excess of H2O
in the inlet gas stream. For an inlet gas
composition of 50 percent water and 50
percent CO2 and temperatures of 1000-
1400° F, Table 3 gives the percentage of H2 S
in the gas effluent from the reactor and the
percentage of H2S in the same effluent after it
has been dried.
TableS. H2S CONCENTRATION9 IN DRIED AND
UNDRIED PRODUCT GAS STREAM AT EQUIL-
IBRIUM
Temperature,0 F
1000
1100
1200
1300
1400
% H2S in
undried gas
23.0
11.4
4.7
2.7
1.4
% H2S in
dried gas
37.4
20.5
9.9
5.1
2.8
Assumptions are 10-atm total pressure with 50
percent H20 and 50 percent C02 inlet gas.
B. Experimental Studies
4. Reductive Decomposition of CaSO4
To test the accuracy of the equilibrium
compositions calculated for the reduction of
CaSO4 with CO/CO2 mixtures, experiments
have been performed in a static system. The
1-1-17
-------
apparatus consists of a horizontal tube reactor
fabricated from recrystallized alumina. The
tube is 36 inches long and has an ID of 3
inches with 1/4-in. walls. One end of the
alumina reactor is closed and the opposite end
is capped with a stainless steel 0-ring flange.
The flanged end is outside the furnace.
The experiments are performed in the
following manner. A 3-gram sample of CaSO4
(Drierite) is placed in an alumina boat and
loaded into the reactor. The system is closed
and leak-checked. The CaSC>4 is dried at
500°F under vacuum for 15 to 20 hours. While
the system is still at 500°F and isolated from
the vacuum pump, a predetermined pressure
(
-------
b. Reaction of CaS with CO2/H2O — The
product of each of the reduction experiments
was carbonated at 10 atm to simulate a
product that would be obtained in an actual
10 atm combustion-reduction experiment.
This material was then reacted batchwise with
an equimolar mixture of CCh/HbO at
temperatures ranging from 900 to 1100°F, at a
gas velocity of approximately 1 ft/sec and at
10 atm pressure in the 2-in. diameter
fluidized-bed reactor. The H^S concentration
in the outlet gas was monitored using a
quadrupole mass spectrometer.
The results to date have shown that:
1. The reaction producing H2S is initially
rapid, but the rate decreases after a short
time. Typically, the reaction rate drops to near
zero after several minutes.
2. The peak concentration of H2S in the
outlet gas is high and near the expected
equilibrium value.
3. Typically, half or less of the CaS reacts.
In continuing work, the effects of process
variables are being studied in an attempt to
increase the quantity of CaS that is reacted.
C. Sulfation-Regeneration Cyclic
Experiments
Since it will be desirable to reuse the
additive material several times in commercial
application's, a cyclic experiment has been
performed to obtain data on the pickup and
removal of sulfur from additive particles and
to determine decrepitation and attrition of
additive particles during sulfation-
regeneration cycles. Six cycles of simulated
combustion and two-stage regeneration were
performed with a single bed of additive. The
starting material (1.2 kg) was obtained from a
coal combustion experiment in which dolo-
mite No. 1337 had been used as additive. The
initial sulfur content of the bed was 15.4
weight percent. The experiment was per-
formed batchwise in the 2-in. diameter fluid-
ized-bed reactor:
The sulfation portion of cycle 1 was omitted
since the additive already contained sulfur.
For the remaining cycles, the constituents of
the sulfating gas were N2, CC»2, H2O, Ch, CO,
and SO?. The sulfation reaction was allowed
to proceed until the bed material had
essentially ceased further pickup of SO2. After
the bed had been sulfated, the CaSCM was
converted to CaS, using H2 or CO as
reductant at 1550 to 1600°F and 10 atm. The
bed was then reacted with a CO2/H2O gas
mixture at 1000°F and 10 atm to convert the
CaS to CaCOa. A sample of the bed material
was taken after each step in the cycle and
analyzed for sulfur and sulfide content.
The effluent gas stream was analyzed for
H2S concentration, using the quadrupole
mass spectrometer. A plot of H2S concen-
tration versus reaction time in the six cycles is
presented in Figure 15.
The results indicated that the conversion of
CaSO4 to CaS in the reduction step was
ineffective. Only in cycles 1 and 4 was the
conversion to CaS greater than 50 percent. A
possible cause could be the interaction of
CaSO4 and CaS to form nonporous surfaces;
the formation of easily sinterable cakes has
been reported when these materials are
present.
The data for the regeneration step
(presented in Figure 15) showed that the peak
concentrations of H2S in the effluent gas
decreased from 13 volume percent (dry basis)
for cycle 1 to 0.5 volume percent (dry basis) for
cycles 5 and 6. The percent calcium sulfide
converted to CaCOs decreased to a very low
indeterminate value after several cycles. It
appears from these data that a layer of
material of low permeability is built up on or
within pores of the additive particles,
inhibiting the removal of sulfur. The high sul-
fur loading of the bed particles in these
experiments may be a factor in the poor
conversion.
In continuing work, it is planned to
investigate the use of a high reduction
temperature (e.g., 1800°F) to promote more
complete reduction and to remove part of the
1-1-19
-------
sulfur as SO 2. The remaining sulfur would
then be removed as H2S by reaction with CO2 -
H2O. The two gas streams could be combined
for conversion to elemental sulfur.
PRESSURIZED COMBUSTION AND RE-
GENERATION—PILOT PLANT DESCRIP-
TION
Equipment has been installed for
combusting coal at pressures up to 10 atm and
for continuously regenerating sulfated lime for
reuse. A simplified equipment schematic is
shown in Figure 16. The regenerator and the
fluidized-bed combustor have a common off-
gas system (cyclones, filters, gas-sampling
equipment, pressure let-down valve, and
scrubber) and will not be operated
simultaneously. Either the combustor or
regenerator will be disconnected from the off-
gas line and flanged off when the other unit is
in operation.
The combustion unit consists of a 6-in.
schedule 40 pipe (Type 316 SS) approximately
11 ft long, with an outer shell consisting of 12-
in. schedule 10 pipe (Type 304 SS) over nearly
the entire length. A bellows expansion joint is
incorporated into the outer shell to
accommodate the differential thermal
expansion of the inner and outer vessels.
The unit is of a balanced pressure design;
i.e., the annular chamber between the two
pipes is maintained under pressure so that a
differential pressure does not exist across the
hot inner pipe wall. The balancing pressure
for the shell is supplied by a bank of nitrogen
cylinders.
A bubble-cap-type gas distributor is
flanged to the bottom end of the inner vessel;
thermocouples, solids feed lines, and solids
take-off lines extend through the gas
distributor. The outer wall of the 6-in. pipe is
wrapped alternately with sixteen 3000-W
tubular resistance heaters and 3/8-in. OD
cooling coils that are spray-metal-bonded.
Internal cooling coils of 3/8-in. pipe extend
down into the interior of the 6-in. vessel from
the flanged top to provide additional heat
transfer area. Water flow to the cooling coils is
regulated using flow indicators and is adjusted
on the basis of the temperatures of the fluid -
ized bed and reactor wall.
Both the annular pressure chamber and the
reactor itself are equipped with rupture disc
assemblies and pressure relief valves vented to
the room ventilation exhaust ducts.
The regenerator has a 3-in. ID surrounded
by 2-1/2 inches of Plibrico castable refractory
and encased in an 8-in. schedule 40 pipe (316
SS). This entire assembly is enclosed by a
pressure shell made of 12-in. schedule 20
carbon steel pipe. Differential thermal
expansion between the inner and outer pipes is
accommodated by the use of packing glands
on the lines entering the bottom flange of the
unit. The unit is of a balanced pressure
design; i.e., the annular chamber between the
two pipes is maintained under pressure so that
a large differential pressure does not exist
across the hot inner pipe wall. The balancing
gas is nitrogen. Since the annular space is not
gas tight with respect to the regenerator inner
vessel, the pressure in the annular space will
be maintained slightly higher than the
regenerator pressure to prevent process gases
from entering the annulus. A pressure alarm
gauge will monitor the pressure in the annular
space and will be set to warn of both high and
low pressure.
A bubble-cap-type gas distributor is
connected to the bottom of the inner vessel via
a slip fit and held in place with retaining
screws. Thermocouples, solids feed lines, and
solids take-off lines pass through the gas
distributor and then through packing glands
on the bottom flange of the outer pressure
vessel. The wall of the inner vessel is wrapped
alternately with 3000-W tubular resistance
heaters and 3/8-in. OD tubing coils. Both the
annular chamber and the regenerator itself
are equipped with rupture disc assemblies and
pressure relief valves vented to the room
ventilation exhaust ducts.
1-1-20
-------
The primary filter cartridges are suitable ASME code requirements . The design rating
for temperatures up to 350°F (epoxy- of the unit is 150 psig at 1500°F. Air (or gas)
impregnated cellulose-base material with glass passes through an annulus, reverses direction,
fiber substrate). The secondary filter and passes through a heated section.
cartridges are Rigimesh (woven metal wire).'
The gas preheater is of a balanced pressure The feeders are of the rotary pocket type
design and was designed in accordance with equipped with hoppers.
1-1-21
-------
TO
GAS-ANALYSIS
SYSTEM
PREHEATER
TO GLASS FIBER
FINAL FILTER AND
VENTILATION EXHAUST
Figure 1. Bench-scale combustor system.
1-1-22
-------
100
90L_•-
70
60
o
5
LU
ce
oc
PITTSBURGH
COAL
ILLINOIS
COAL
SYMBOL
A
A
•
O
9
RUN NO.
HUWP-1A,
-ID, -2A, -2B
AR-4, -5
AR-1E
BC4
BC-9
BC-10
TEMP., ADDITIVE
COAL TYPE T NO.
PITTSBURGH 1450 1359
ILLINOIS
ILLINOIS
ILLINOIS
ILLINOIS
ILLINOIS
1550 1359
1600 1359
1600 1359
1600 1360
1600 1377
AVERAGE PARTICLE SIZE RANGE FOR ADDITIVE: 490-630 jjm
GAS VELOCITY IN OOMBUSTOR: 2.6 TO 2.8 ft/sec
2 3 4 5
Ca/S MOLE RATIO
Figure 2. Effect of Ca/S mole ratio on.sulfur retention.
M-23
-------
UJ
LLJ
Cd
100
90
80
70
60
50
40
30
20
O ILLINOIS COAL, Ca/S~2.5
• PITTSBURGH COAL, Ca/SM.O
1300
1400
GAS VELOCITY, 3 ft/sec
EXCESS OXYGEN, 3%
LIMESTONE NO. 1359
J
1600
1500
TEMPERATURE, °F
Figure 3. Effect of fluid! zed-bed temperature on sulfur retention .
1700
1-1-24
-------
100
90
80
70
60
50
40
Ca/S=l
30
20
10
CRE DATA (1470 T)
WELBECK COAL, BRITISH LIMESTONE (440
ANL DATA (1550 "F)
ILLINOIS COAL, LIMESTONE NO. 1359 (>1000 pi)
A ILLINOIS COAL, POINT TAKEN FROM CURVE (FIGURE 2)
FOR LIMESTONES AND DOLOMITE (*630jim), Ca/S = 4
2 4 6
SUPERFICIAL GAS VELOCITY, ft/sec
Figure 4. Effect of superficial gas velocity on sulfur retention.
1-1-25
-------
70
60
50
CO
+
CM
^
"3.
in
40
30
20
10
0
BED TEMPERATURE
P1450°F
O1550°F
1600 °F
SECONDARY AIR INTRODUCED
LIMESTONE ADDITION
NO LIMESTONE ADDITION
H
0
Figure
50
60 70 80
AIR, % of stoichiometric (based on feed rates)
90
5. Effect of air feed rate on percent of sulfur in off-gas as H2S.
1-1-26
-------
o
I-
LU
a:
cc
O 63
O 58
O 53
O 60
NUMBER NEAR POINT IS THE
PERCENT OF STOICHIOWIETRIC
AIR ADDED
SINGLE-STAGE OXIDIZING
EXPERIMENTS
(1450-1470 °F)
BED TEMPERATURE, SUBSTOICHIOMETRIC
AIR EXPERIMENTS
[] 1450°F
O1550°F
A1650°F
Ca/S MOLE RATIO
Figure 6. Sulfur retention in oxygen-excess and oxygen-deficient experiments.
1-1-27
-------
100
90
80
70
60
50
40
30
20
10
50
o
BED TEMPERATURE
D 1450 °F
O 1550°F
A 1650°F
60 70 80
AIR, % of stoichiometric
90
100
110
Figure 7. Effect on sulfur retention of air-feed rate to first stage.
1-1-28
-------
300
250
200
150
100
50
40
1450 °F
1650°F
1550"F
40
50
60 70 80
AIR, % of stoichiometric (based on feed rates)
90
100
Figure 8. Effect of air feed rate and fluidized-bed temperature on NO concentration in off-gas
from the first stage during combustion of coal.
1-1-29
-------
2400
2320
6.0 5.0 4.0 3.40
7.0 I 5.5 | 4.5 j 3.70 J 3-°0
1600
0.001
0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 0.050 0.055
PCO/PC02
Figure 9. Pressure of S02 in equilibrium with CO/C02 mixtures as a function of
temperature (10 atm total pressure).
1-1-30
-------
FOR CaS +H20 +C02^H2S + CaC03
1000
1100
1200 1300
TEMPERATURE,°F
Figure 10. Equilibrium constant as function of temperature, CC>2 +1^0 + CaS-*CaCC>3 •+H^S-
1-1-31
-------
2.0
H20 +C02 + CaS;±CaC03 +H2S
1100 1200
TEMPERATURE, °F
Figure 11. Pressure of H2S in undried gas
stream as function of temperature at
10-atm Ptotal-
1-1-32
-------
0.18f
0.17
0.16
0.15
0.14
0.13
0.12
0.11
E
eg
£ 0.10
^ 0.09
;
_i
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
TEMPERATURE: 1900 °F
TOTAL PRESSURE: 1 aim
• A-14
1-3/4 hr
A-9,1 hr
A-13,2 hr
A-12,2 hr
, A-14,20-3/4 hr
A-15
Ihr
C0/C02= 0.034'
A-8, 1 hr
A-7, 1.5 hr
• A-9,19-1/4 hr
CALCULATED EQUILIBRIUM PARTIAL
PRESSURE OF S02 FOR
CaS04-j.CO » CaO+C02 +S02
• A-15,18-1/4 hr
\ I I
0.002
0.004
0.006
0.008
0.010
0.012
0.014
Figure 12. Partial pressures of S02 over a range of PCO/PC02 ratios.
1-1-33
-------
CONDITIONS
o
UJ
CQ
UJ
20
15
10
o
o
g 5
CATS-16 100% H2, 1 atm, 1600 °F
CATS-18 100% H2, 1 atm, 1450 °F
CATS-19 100% H2, 1 atm, 1350 °F
100
200
TIME, min
300
Figure 13. Sulfide content of bed during re-
duction of partially sulfated dolomite with
hydrogen at various temperatures.
100
90
80
ss
g 70
u_
_j
3 60
£ 50
LL.
O
40
30
£
LU
o
o
20
10
CATS-17
CATS-190-
REDUCTANT
o HYDROGEN
•CO
1200 1300 1400 1500 1600 1700 1800 1900
TEMPERATURE, °F
Figure 14. Effect of temperature on reduction
of CaS04 (dolomite), 4.5 hours of reduction
time.
1-1-34
-------
s
CO
•=C
CO
UJ
QCYCLEl
D CYCLE 2
• CYCLE 3
A CYCLE 4
• CYCLE 5
A CYCLE 6
to 7 —
0 2 4
TIME, min
30
Figure 15. H2S concentration in effluent gas stream in regeneration step.
1-1-35
-------
REHEATER
H. P. STEAM
PRESSURE
LET-DOWN
VALVE
ROOM VENTILATION
AIR
H2 CO C2H2
Figure 16. Simplified schematic of combustion-regeneration equipment.
-------
2. A REGENERATIVE LIMESTONE PROCESS
FOR FLUIDIZED-BED COAL
COMBUSTION AND DESULFURIZATION
R. C.HOKE, H. SHAW, AND A. SKOPP
Esso Research and Engineering Company
ABSTRACT
The factors influencing NOx emissions from a fluidized limestone bed coal combustor were
studied. NOx emissions were decreased by a decrease in temperature and a decrease in excess air.
Sulfated lime depressed NO x emissions compared to an inert alundum bed. An apparent cause
of these effects is reduction of NOX by CO. The NO/CO reaction was then studied further in fixed
bed units. CaSO4 catalyzed the reaction slightly compared to alundum in a dry system. CaO
promoted the reaction significantly, giving over 90 percent conversion in the absence of CO2- CO 2
was found to inhibit the rate, possibly due to a kinetic limitation caused by the presence of the
COa. Study of the reaction of SO 2 and NO indicated that the reaction is catalyzed by partially
sulfated lime, but not by CaSO4 or alundum. Temperature was found to have a negative effect on
the reaction, apparently due to the thermal instability of an intermediate,CaSO3. Two-stage
combustion of coal was studied to promote the CO/NO reaction and reduce NO emissions further.
NO emissions were reduced by two-stage combustion, and the reduction was enhanced by
operating the first stage at lower air levels.
Regeneration of CaSO 4 to CaO and SO2 by CO and H2 was studied at pressures up to 9.5 atm.
SO2 levels in the off-gas as high as 7.5 percent were measured at pressures up to 6 atm. The
maximum concentration measured to date at 9.5 atm is 2.2 percent. The measured levels are 50-60
percent of the levels calculated from equilibrium considerations.
INTRODUCTION
Esso Research and Engineering Company
is conducting an experimental program for the
Environmental Protection Agency under con-
tract CPA 70-19 to develop a regenerative
limestone process for fluidized-bed coal com-
bustion and desulfurization. This is a part of
EPA's overall program to examine fluidized-
bed combustion as a possible new power
generation technique. The potential of fluid-
ized-bed combustors for air pollution control
is good because the intimate gas-solid con-
tacting in a fluidized bed promotes high SO2
removal efficiency on suitable materials such
as limestone or dolomite.
A schematic diagram of the process is
shown in Figure 1. In the combustor, the sul-
fur in the coal is burned to SO2 which then
reacts with the lime to form CaSO4. The
system being studied by Esso involves trans-
ferring the partially sulfated lime from the
combustor to a separate regeneration vessel
where the sulfated lime is regenerated accord-
ing to the reaction
CaSO4+CO -> CaO+SO2+CO2
H2 H20. (1)
1-2-1
-------
The regenerated stone (CaO) can then be regenerator has a high SO2 concentration and
returned to the combustor for further use, can be used as feed to a by-product sulfur or
thereby substantially reducing the fresh lime- sulfuric acid plant.
stone requirement. The off-gas from the
1-2-2
-------
Previous Studies
Various laboratories including Esso
Research have studied fluidized-bed coal
combustion over the past few years. The
results of the studies are summarized in Table
1 and have shown that coal can be burned
Table 1. PREVIOUS FBC FINDINGS
Coal combustion efficiency high
Over 90% removal of S02
NOX emissions reduced
Sulfated lime can be regenerated
CaS04 + CO -* CaO + S02 + C02
Activity maintenance of recycled lime satisfactory
after 7 cycles
Pressurized FBC system more attractive
efficiently with over 90 percent removal of SO 2
and with reduced NOx emissions. Regener-
ation of sulfated limestone has been studied
using a number of regeneration methods. The
method studied at Esso Research consisting of
the one step reduction of CaSO4 to CaO and
SO2 gives 6-10 percent SO2 in the product gas
when carried out at 1 atm and about 2000°F.
The recycled lime was also shown to maintain
a reasonably high level of activity after seven
combustion/regeneration cycles.
Economic studies were carried out by
Westinghouse Research Laboratories under
contract to EPA.1 These studies indicated that
operation of the combustor and regenerator at
higher pressures, approximately 10 atm,
would be significantly more economical than
atmospheric pressure operation. As a result,
the current studies are being made at higher
pressures.
Objectives
Objectives of Esso Research's current
experimental program are summarized in
Table 2 and consist of (1) investigating the
factors influencing the reduction of NOX
emissions in fluidized-bed combustion, and (2)
studying the regeneration of sulfated lime at
pressures up to 10 atm. The latter objective
Table2. OBJECTIVES OF CURRENT EXPERI-
MENTALWORK
1. Investigate factors influencing reduction of
NOX emissions
Effect of temperature, excess air, bed
materials
Reaction of NO with CO and S02
2. Study regeneration of sulfated lime
S02 levels attainable at higher pressures
Kinetics of regeneration and stone activity
maintenance
has required construction of higher pressure
experimental equipment.
EXPERIMENTAL EQUIPMENT
A number of experimental units were used
in the current program. A flow diagram of the
atmospheric pressure fluidized-bed combus-
tion is shown in Figure 2. The reactor is a 3-
in.-ID Incoloy tube. Four continuous flue gas
analyzers are used including IR SO2 and CO
analyzers and polarographic NOx and O2
analyzers. A high pressure regeneration unit
capable of operating up to 10 atm was recently
built and is shown in Figure 3. The reactor
consists of a 3-in.-ID alumina tube contained
in a 12-in. carbon steel vessel. The reactor is
15 feet long. The interior of the steel vessel is
lined with 4-1/2 inches of castable refractory
insulation. The regeneration feed gas is
produced by combustion of propane. N2 and
CO2 can also be added to the burner to adjust
the composition of the regeneration gas. Most
of the input heat is provided by combustion of
propane, but additional heat input is provided
through electrical heaters adjacent to the
alumina tube and an air preheater. The unit is
heated by burning the propane under excess
air conditions. When the operating tempera-
ture has been reached, the air/fuel ratio is
instantaneously changed to substoichiometric
conditions and N2/CO2 flow is started,
thereby assuring a rapid change from heat-up
to operating conditions.
Two small fixed bed units were also used in
these studies. Three different reactors were
used: a 2-1/2-in. alumina tube operating at 1
atm, a 1-in. stainless steel tube operating at
1-2-3
-------
pressures up to 10 atm, and a specially
constructed regenerator. The regenerator
consisted of a 1-in. alumina tube contained in
a 3-in. pipe with fiber insulation between the
pipe and tube. This unit was capable of
operating at pressures up to 9.5 atm at tem-
peratures up to 2000 °F. All fixed bed units
were electrically heated.
EXPERIMENTAL RESULTS
Factors Affecting NO x Emissions
It was determined previously that NOX
emissions measured at the low temperatures
occurring in fluidized-bed combustion are
formed by oxidation of nitrogen compounds in
the coal. Oxidation of atmospheric N2 occurs
only at higher temperatures. In this study, the
effects of temperature, excess air, and fluid-
ized bed material on NO emissions were
measured. The effect of temperature using a
bed of CaSO4 in the combustor is shown in
Figure 4. As temperature decreased, NO emis-
sions dropped rather sharply below 1500°F.
The effect of excess air using a bed of CaSO4 is
shown in Figure 5. Actual NO emissions
decreased as excess air (percent O2) was
increased. However, when the emissions were
normalized to a constant gas volume (at 3
percent ©2), the NO emissions increased as
the excess air increased. The NO formation
rate was thus increased by the higher average
oxygen concentration in the bed. The effect of
bed material is shown in Figure 6. CaSO4 gave
lower emissions than alundum. With a CaO
bed the emissions were high initially, but as
the bed sulfated the emission level approached
that of CaSO4.
One consistent explanation for these results
is the reaction of NO with CO. Carbon
monoxide emissions are higher at the lower
temperatures and at lower excess air condi-
tions. The higher CO levels then give lower NO
emissions. The effect of bed materials appears
to be a catalytic effect.
Reactions of NO and CO
The reaction of CO and NO was studied
1-2-4
further in fixed-bed units. The effects of bed
material, temperature and feed gas composi-
tion were studied. In a dry system, CaSO4
catalyzed the reaction slightly arid showed a
small effect of temperature, but alumina and
an empty bed gave essentially no reaction.
This is shown in Table 3. However, the
Table 3. NO-CO REACTIONS-EFFECTS OF
TEMPERATURE AND BED MATERIAL
Bed material
Bed temperature, "F
Inlet gas composition
NO, ppm
CO, ppm
NO conversion, %
CaSO4
1500
1990
4000
4
CaSO4
1700
2025
4000
6
Alumina
1500
2010
4000
1
None
1500
2035
4000
0.5
addition of water enhanced the reaction and
gave the same NO conversion regardless of the
presence of the bed material. This is shown in
Table 4. But when CaO was used as the bed
Table 4. NO-CO REACTIONS-EFFECT OF
WATER VAPOR
Bed material
Inlet gas composition
NO, ppm
CO, ppm
H20, %
NO conversion, %
CaS04
1990
4000
0 7.2
4 12
Alumina
2010
4000
0 7.2
1 13
None
2035
4000
0 7.2
0.5 15
Temperature, 1500"F
material in a dry system, a very rapid reaction
occurred which gave over 90 percent conver-
sion of the limiting reactant as shown in Table
5. The reaction proceeded in 1:1 mole ratio of
CO and NO suggesting the reaction
2 CO + 2 NO
2CO
(2)
Carbon dioxide was then added to the feed
and reduced the conversion significantly over
-------
both calcined limestone and calcined
dolomite. This is shown in Table 6,
Table 5. NO-CO REACTIONS-EFFECT OF
CaO
Bed source
Inlet gas composition
NO, ppm
CO, ppm
Outlet gas composition
NO, ppm
CO, ppm
Conversion, %
Lime #1359
1400
940
400
10
99
1800
1870
20
160
99
Dolomite #1337
1400
900
350
20
98
1990
2080
240
100
95
Temperature, 1600"F
Residence time, 0.3 sec
Table 6. NO-CO REACTIONS-EFFECT OF
CO 2
Bed source
Inlet gas composition
NO, ppm
CO, ppm
CO2, %
Conversion, %
Lime #1359
1400
940
0
99
860
990
17
26
Dolomite #1337
1400
900
0
98
840
980
16
19
Temperature, 1600°F
Residence time, 0.3 sec
Table 7. NO-CO REACTIONS-EFFECT OF
PRESSURE AND RESIDENCE TIME
Bed source
Pressure, atm
Residence time, sec
Inlet gas composition
NO, ppm
CO, ppm
CO,, %
Conversion, %
Lime #1359
1
0.3
860
990
17
26
10
3
890
980
18
82
10
0.3
1150
1240
13
16
Dolomite #1337
1
0.3
840
980
16
19
10
3
840
980
16
88
Temperature, 1600"F
TableS. NO-CO REACTIONS-EFFECT OF O-
Bed source
Inlet gas composition
NO, ppm
CO, ppm
CO-., %
02, %
Conversion, %
Lime #1359
1150
1240
13
0
84
970
1060
16
2.4
95
Dolomite #1337
840
980
16
0
88
980
1080
15
2.3
91
Temperature, 1600°F
Pressure, 10 atm
were considered as possible explanations, but
were ruled out after closer examination.
The effect of pressure was studied and
although increasing the pressure to 10 atm in
the presence of CO 2 apparently increased the
conversion, the increase was probably due to
increased residence time. At equivalent
residence times, increasing pressure appeared
to decrease the conversion slightly. This is
shown in Table 7. Changing the background
gas from N2 to argon appeared to increase the
conversion very slightly, but the effect may not
be significant. Oxygen was added to the feed
and appeared to increase conversion slightly.
This is shown in Table 8.
The most likely explanation for these
effects is a kinetic limitation caused by the
presence of the CO2. Formation of CaCOa
and inhibition caused by chemical reversibility
Reactions of NO and SO2
Further studies of the reaction of NO and
SO 2 were made in a fixed-bed reactor. The
effects of bed material and temperature were
studied. The effect of bed material is shown in
Table 9. The results show that NO and SO2
did not react in the vapor phase or over
alundum or CaSO/^. However, a reaction did
occur over partially sulfated lime and
appeared to be dependent on SO2
concentration. Further rate studies indicated a
0.5 order dependence on the NO concentra-
tion. Temperature had a negative effect on the
rate, decreasing the rate with increasing tem-
perature, as shown in Figure 7. A proposed
mechanism for the reaction involves the
reversible formation of CaSOs intermediate
from CaO and SO 2- The sulfite then reacts
1-2-5
-------
Table 9. NO-SO2 REACTIONS-EFFECT OF
BED MATERIAL AT 1600°F
Bed material
Gas phase
Partially sulfated
limestone
Alundum
CaSO
NO concentration, ppm
Before SO 2
introduced
900
840
830
820
860
After SO 2
introduced
900
440
180
820
860
SOZ concentration, ppm
Inlet '
1290
780
1510
1000
670
Outlet
1290
300
480
1000
670
with NO to form N2 and CaSO4. However, it is
known that the sulfite becomes unstable in the
temperature range where the SCb/NO
reaction rate drops; this instability is the
probable explanation for the negative temper-
ature effect.
Two-Stage Combustion
The reactions of NO with CO suggest the
possible lowering of NO emissions by
operating a staged combustion system. Air
would be injected at two points in the
combustor giving an 02 lean section at the
bed inlet which should promote NO reduction
because of the relatively high CO levels. The
second step would then complete combustion.
The fluid-bed combustor was modified to
operate in a staged fashion by injecting second
stage air 6 inches above the grid. The results of
two runs are shown in Figure 8. As the ratio of
the second stage air to the first stage air was
increased, the NO emissions dropped. Nitric
oxide emissions were lowered to 200 ppm.
Although these conditions may not be feasible
in commercial operation, the principle of
stage combustion appears attractive.
Regeneration of Sulfated Limestone
Regeneration studies were carried out in
fixed and fluidized beds using CaSO4 at
pressures up to 9.5 atm. The results are shown
in Table 10.
Concentrations of SO2 in the off-gas as
high as 7.5 percent have been measured at
pressures up to 6 atm. At 10 atm, the highest
SO 2 concentration measured to date was a
little over 2 percent. Comparisons were also
made with SO 2 levels estimated from
equilibrium calculations made by Argonne
National Laboratory.2 The equilibrium SO2
partial pressure is determined by the tempera-
ture and the CO/CO 2 ratio in the gas in
equilibrium with the solids. However, at each
temperature, there is a CO/CO2 ratio which
gives the maximum attainable SO2 partial
pressure for the temperature in question. In
the fixed-bed runs, the off-gases were not
analyzed for CO and CO 2, and the measured
SO2 concentrations had to be compared to the
maximum equilibrium SO2 concentration. In
the fluidized-bed runs, comparisons were
made at the actual CO/CO 2 ratio measured,
although these ratios were probably in error
due to oxidation of CO in the exit lines from
the reactor. The comparison of measured and
calculated SO2 levels is given in Table 10. The
SO2 levels measured in the fixed-bed unit
were less than 50 percent of the maximum
attainable at the temperatures of the runs.
The results from the fluidized bed were closer
to the equilibrium concentrations calculated
for the CO/CO2 ratio measured for each run.
In general, the measured SO 2 concentrations
were 50-60 percent of the equilibrium levels;
Table 10. REGENERATION OF SULFATED LIMESTONE, CaSO4 AS BED MATERIAL
Unit
Rxed
Rxed
Fluidized
Fluidized
Fluidized
Fluidized
Pressure,
atm
3
9.5
3.2
6.2
6.0
6.0
Temperature,
°F
2000
2000
1990
2100
1950
1870
SO2 concentration,%
MeasursH Calculated
7.2
2.2
5.2
7.5
2.0
1.8
15.0
4.8
9.8
11.3
3.7
1.8
SO 2 ratio
measured/calculated
0.48
046
0.53
0.66
0.54
1.0
1-2-6
-------
the last run met the calculated equilibrium
concentrations. Further work is planned in the
fluidized-bed regeneration unit to determine
the SO 2 levels attainable at pressures up to 10
atm as a function of temperature, regener-
ation gas composition and flow rate, particle
size, and sulfated lime source. A new
pressurized combustor unit is being built
which will be used with the regenerator to
measure cyclic activity maintenance of various
stones.
BIBLIOGRAPHY
1. Archer, D. H., D. L. Keairns, J. R. Hamm,
R. A. Newby, W. C. Yang, L. M.
Handman, and L. Elikan. Evaluation of
the Fluidized-Bed Combustion Process.
Westinghouse Research Laboratories,
Pittsburgh, Pa. Prepared for the Environ-
mental Protection Agency, Research
Triangle Park, N. C. under Contract
Number CPA 70-9. November 1971.
2. Jonke, A.A., G.J. Vogel, J. Ackerman, M.
Haas, J. Riha, C.B. Schoffstoll, J. Hepperly,
R. Green, and E.L. Carls. Reduction of
Atmospheric Pollution by the Application
of Fluidized-Bed Combustion. Argonne
National Laboratory, Argonne, 111. Pre-
pared for the Environmental Protection
Agency, Research Triangle Park, N.C.
under agreement EPA-IAG-0020. Monthly
Progress Report Number 38, December
1971.
1-2-7
-------
FLUE GAS AND
COAL FLY ASH
FLUID BED
COMBUSTOR
FRESH
SORBENT
&COAL
SULFATED
SORBENT
REGENERATED
SORBENT
HIGH S02 GAS
TO BY-PRODUCT
PLANT
FLUID BED
REGENERATOR
DISCARDED
SORBENT
FLUIDIZING
AIR
REDUCING
GAS
-»CaO+S02+C02
H2 H20
Figure 1. Esso proposed fluidized-bed combustion-lime regeneration system.
1-2-8
-------
CYCLONE
AND
FILTER
FEEDER
SCALE
nnnnnik
WTM
REFRIGERATOR
VENT-
INSTRUMENT CALIBRATION BY-PASS
COAL
HOPPER
AND FEEDER
N2 CO $02
02 NO AIR
GAS ROTANIETERS
CONDENSATE
Figure 2. Esso fluidized-bed combustion unit.
IRS02
ANALYZER
IR CO
ANALYZER
WTM
NOX ANALYZER
1
-WTM
POLAROGRAPHIC
02 ANALYZER
T
I
J_
INTERMITTENT
GAS SANIPLER
1-2-9
-------
VENT
^ COOLER
Q-TT-!
VENT
DRY TEST
METER
ANALYZERS
KNOCKOUT
PREHEATER
Figure 3. Fluidized-bed regeneration unit.
800
700
600
500
400
I 300
200
100
U=6 ft/sec
Ho=6 in.
02=4%
-\
1300
1400
1700
1-2-10
1500 1600
BED TEMPERATURE, °F
Figure 4. NO emissions as a function of bed temperature.
1800
-------
900
U=6 ft/sec
T-1600°F
BED MATERIAL'CaSOi (-1000»)
(dp)COAb300jul: "
700
600
500
400
NORMALIZED
TO 3% 02
ACTUAL
1000
800
2
8
4 6
02 IN FLUE GAS, %
Figure 5. Effect of 02 in flue gas on NO emissions (CaSO4 bed).
CaO
10
ALUNDUM
CaS04
600
400
200
T»1600 °F
U = 6 ft/sec
02 (FLUEGAS)=4%
H=6 in.
0.5
1.0
1.5
2.0
RUN TIME, hr
Figure 6. NO emissions using different bed materials.
2.5
1-2-11
-------
60]
50
40
35
30
,20
oT 15
"""o
z
^ 10
* 9
8
7
6
5
4
INLET GAS COMPOSITION
N0 = 485ppm
S02 =700ppm
N2 = BAL.
GAS RESIDENCE TIME=0.04 sec
BED MATERIAL-16.6% SULFATED LIME
* RATE MEASURED 15
MINUTES AFTER GAS
FLOW STARTED TO
REACTOR
"I I I I I I I I I I I I I I I I I I I I I I I I I
0.86 0.88 0.90 0.92 0.94 0.96 0.98
1000/t, °K
Figure 7. Temperature dependence of NO-S02-CaO reaction system.
1.0
800
700
600
E 500
3-
V)
B 400
CO
CO
§
o 300
200
100
0
I 1 I I
BED MATERIAL = CaS04
TOTAL AIR =110% of STOICHIOWIETRIC
T=1600°F
TOTAL AIR =130% of STOICHIOMETRIC
T=1750 °F
0.5
1.0
1.5
1-2-12
SECOND STAGE AIR
FIRST STAGE AIR
Figure 8. Staged FBC results.
-------
3. COMBUSTION OF COALS
IN FLUIDIZED BEDS
OF LIMESTONE
R. L. RICE AND N. H. COAXES
Morgantown Energy Research Center
Bureau of Mines
U.S. Department of the Interior
INTRODUCTION
The experimental work reported here was
performed by the Bureau of Mines, Morgan-
town Energy Research Center, under contract
to the Control Systems Division, Office of
Research and Monitoring, Environmental
Protection Agency. The phase of work
assigned to the Bureau was to test various
coals as fuel in fluidized beds of limestone, to
compare sulfur retention, and to measure heat
transfer with tubes immersed in the bed. The
program involved testing five types of bitumi-
nous coal from high volatile A to low volatile,
which varied in ash content from 8 to 24 per-
cent and in sulfur content from 2 to 4 percent.
1-3-1
-------
EQUIPMENT AND PROCEDURE
Figure 1 shows the 8-ft high combustor
used in the tests. The bed was supported by a
cone-shaped plate, perforated by 1/8-in. holes
fitted with welded stainless steel 90° elbows to
inject the fluidizing air axially and parallel to
the cone surface. Water passing through a
heat exchanger made of 3/4-in. pipe extracted
heat from the combustion bed. Tests were
performed at superficial fluidizing velocities of
3 ft/sec and 6 ft/sec. In the 6-ft/sec tests, an
additional heat exchanger was installed to
control the temperature of the gases leaving
the combustor.
Figure 2 is a flow diagram of the system.
Coal was metered by a screw conveyor, and
then fed pneumatically near the base of the
bed. Limestone was fed through the side of the
combustor just above the bed by a screw
conveyor. The bed level was maintained by
periodically removing material from the
bottom with a 3-in. screw conveyor.
Combustion products were passed through
two centrifugal separators for removal of most
of the entrained solids, and then to a bag filter
for final cleaning. Solids from the first cyclone
could be reinjected into the combustion zone.
Combustion gases were monitored
continuously for O2, CO2, CO, and SOa by
.infrared analyzers except for O2 which
utilized a paramagnetic system. After each
test the residue was removed from the bottom
of the combustor by the screw conveyor.
Startup was accomplished in about two
hours by burning natural gas in the combustor
and then injecting coal mixed with limestone
into the combustion chamber. A 2-ft bed was
established at about 1200°F, after which the
gas was shut off, and coal and limestone were
fed at a rate that is compatible with the super-
ficial air velocity and the designated run
conditions.
Combustion Tests
Five types of coal were burned in beds of
limestone to determine its effectiveness for
retaining sulfur. The limestone was the type
1-3-2
designated BCR 1359 (97 percent CaCC-3,
Northern Virginia); various sizes of limestone
were used in tests at 3 ft/sec but in tests at 6
ft/sec the limestone was sized to 1/4- by 3/16-
in. The coals were crushed by a hammer mill
to the range of sizes shown in Table 1. Typical
Table 1. TYPICAL SIZE RANGE OF COALS
BURNED IN FLUID-BED COMBUSTION
TESTS
Screen size,
mesh(USS)
-1/4-inch + 20
-20 + 40
-40 + 100
100 + 150
-150 + 200
-200
Weight percent
41-53
16-24
15-20
2-6
2-5
4-10
analyses of the various types of coal are given
in Table 2. These analyses varied somewhat
throughout the test series because batches of
the same coals were purchased at different
times.
Combustion tests, generally of over 70-hr
duration, were made at fluidizing velocities of
3 and 6 ft/sec. At 3 ft/sec, several tests were
made with each coal; at 6 ft/sec, only one test
was made with each coal. Results of the tests
are given in Table 3. Figure 3 shows the effect
of Ca/S mole ratio on sulfur retention in the
bed. It should be noted (from Table 3) that for
most of the tests at 3 ft/sec, material from the
primary cyclone was recycled to the bed. At 6
ft/sec, recycle was possible only with one coal,
hvbb, due to cooling of the bed by reinjection
of the large volume of solids.
The data of Figure 3 generally show that
for the coals burned at 3 ft/sec there appears
to be a trend' in which S removal increases
rapidly to 90 percent as Ca/S is increased to
approximately 2. Two of the tests.however,
appear to deviate from this general pattern:
hvab (A-5-L), Ca/S = 1.8, S removal = 73
percent; Ivb (G-5-L), Ca/S = 2.0, S removal =
68 percent. The test with hvab (A-5-L) was one
of the first tests made and was more cyclic in
-------
Table2. TYPICAL ANALYSES OF COALS
BURNED IN FLUID-BED COMBUSTION TESTS
Proximate analysis, wt %
Moisture
Volatile matter
Fixed carbon
Ash
Ultimate analysis, wt %
Carbon
Hydrogen
Nitrogen
Sulfur
Oxygen
Heating value, Btu/lb
hvcb
(III. #6)
2.4
33.8
55.6
8.2
72.4
5.1
1.2
2.6
8.I
13,045
hvbb
(Ind. #5)
6.0
35.5
50.6
7.9
69.6
5.2
1.2
2.9
7.2
12,530
hvab
(Ohio)
2.3
39.8
49.7
8.2
72.3
5.5
1.5
3.9
6.3
12,930
hvab
(W. Va.)
1.1
33.1
58.0
7.8
76!5
4.8
1.4
3.0
5.4
13,820
mvb
1.1
19.8
62.4
16.7
71.2
4.4
1.0
3.3
2.3
13,050
Ivb
1.1
17.3
69.9
11.7
76.0
4.0
1.2
2.3
3.7
13,120
nature than the later tests. In the test with Ivb
(G-5-L), the SOa meter functioned only a part
of the time, so the average SO 2 concentration
is suspect. Thus, there is reason to believe that
the average results reported for tests A-5-L
and G-5-L are not representative.
The results from tests at 6 ft/sec show a
slightly different pattern, but duplicate tests
would have to be made to confirm this. In only
one of these six tests, that with hvbb, material
from the primary cyclone was recycled to the
bed. Four of the remaining five tests without
recycle indicated the S retention increases as
Ca/S is increased, but not as rapidly as in the
3 ft/sec tests, and that Ca/S of 3 or more is
required to retain approximately 90 percent of
the sulfur in the bed. The test at 6 ft/sec with
Ivb does not fit the pattern, but no explanation
can be offered. In the one test at 6 ft/sec when
recycle was used, S removal was 89 percent at
Ca/S = 2.1, closely approximating results
from tests at 3 ft/sec when recycle was
employed. Therefore, based on the few tests at
6 ft/sec, it is difficult to determine whether the
reduced S removal was caused by the increase
in fluidizing velocity or the absence of recycle.
Increasing the fluidizing velocity has the
following effects:
1, Gas residence time in the bed is reduced.
2. Bed density is decreased which lessens gas-
solids contact.
3. Higher rates of limestone are required
resulting in a reduction of solids residence
time.
4. Larger gas bubbles are formed which per-
mit more bypassing of the SO2.
All of the above effects of increasing the
fluidizing velocity would tend to decrease S
retention. In addition, since the entrained
solids leaving the combustor likely contain
some unreacted limestone, recycling of this
material would be expected to improve S
retention. Thus, the lower S retention at 6
ft/sec is probably caused by both the higher
velocity and the absence of recycle.
1-3-3
-------
Table 3. RESULTS OF FLUID-BED COMBUSTION OF COALS IN BEDS OF LIMESTONE, BCR-1359
Duration, hr
Bed temperature, °F
Superficial velocity,
ft/ sec
Coal rate, Ib/hr
Air/coal, scf/lb
Limestone rate.
Ib/hr
Sulfur in coal, wt %
Ca/S mole ratio
SO2 in POC, ppm
Sulfur removal, wt %
Recycle in use
Carbon utilization, %
Type of coal/run number
hvcb
E-4
84
1500
3.0
52.9
95.5
15.0
3.7
2.5
163
95.8
Yes
-
E-5
79
1510
3.0
51.0
98.8
12.7
3.7
2.2
416
90.8
Yes
89.8
E-6
72
1505
3.0
49.5
101.5
13.3
3.7
2.4
461
89.5
Yes
92.7
E-7
50
1520
5.9
72.7
140.2
17.4
4.1
1.8
1094
69.8
No .
75.6
hvbb
Enos Mine
C-4
87
1560
3.1
36.9
137.1
4.0
4.0
0.85
993
71.4
Yes
98.8
C-5
84
1510
3.0
33.8
148.0
5.0
4.0
1.2
520
83.8
Yes
97.1
Blackfoot Mine
C-6
73
1510
3.0
34.0
148.2
12.7
3.1
3.8
305
87.8
Yes
97,3
C-7
70
1520
3.0
31.7
161.6
13.4
2.9
4.4
189
91.7
Yes
92.5
C-8
84
1545
6.2
67.4
154.4
13.2
2.9
2.1
273
88.7
Yes
96.3
Ivb
G-5
84.5
1535
2.7
31.1
143.5
5.6
2.8
2.0
800
67.7
Yes
85.3
G-6
60
1470
3.0
36.3
138.8
5.7
2.3
1.8
192
92.5
Partial
88.6
G-8
46
1530
6.0
87.6
112.4
11.1
2.6
1.5
121
96.1
No
65.0
-------
Table 3 (continued). RESULTS OF FLUID-BED COMBUSTION OF COALS IN BEDS OF LIMESTONE, BCR-1359
Duration, hr
Bed temperature, °F
Superficial velocity,
ft/ sec
Coal rate, Ib/hr
Air/coal, scf/lb
Limestone rate.
Ib/hr
Sulfur in coal, wt %
Ca/S mole ratio
SO2 in POC, ppm
Sulfur removal, wt %
Recycle in use
Carbon utilization, %
Type of coal /run number
hvab (W. Va.)
Humphrey
Mine
A-5
84
1525
3.1
34.4
144.8
4.4
2.2
1.8
496
73.1
Yes
98.0
A-6
71
1525
3.0
31.8
157.4
3.9
2.2
1.7
63
96.3
Yes
95.1
Love-
ridge
Mine
A-7
84
1525
3.0
31.4
160.3
9.7
3.0
3.4
134
93.6
Yes
96.6
Ire-
land
Mine
1-1
84
1535
6.3
66.6
157.5
13.8
4.2
1.5
848
73.9
No
91.1
hvab (Ohio)
B-4
75
1580
3.0
34.3
144.1
4.3
3.9
1.0
1003
69.4
Yes
-
B-5
84
1520
3.0
34.8
149.0
5.9
3.9
1.4
194
94.1
Yes
93.2
B-6
75
1560
6.0
58.9
170.7
15.2
4.1
2.0
594
80.0
No
85.3
mvb
F-4
60.5
1505
2.8
30.1
155.2
4.6
5.2
0.9
2310
42.7
No
-
F-5a
84
1490
3.0
41.8
118.0
8.0
3.3
1.6
445
88.5
No
82.1
F-6
36
1435
2.6
52.2
86.3
10.0
5.2
1.2
2036
72.4
No
81 .4
F-7a
84
1495
3.0
36.1
140.0
9.3
3.3
2.2
196
93.9
Partial
89.9
F-8a
80
1575
6.0
65.3
154.7
15.2
2.5
2.9
293
86.3
No
74.1
Coal was air-table cleaned.
en
-------
Results in Table 3 also show that when
material from the primary cyclone is not
recycled to the bed, carbon burnup decreases.
This is even true at the lower velocity of 3
ft/sec. In a commercial boiler, if recycle was
not used, the boiler would have to incorporate
a method for increased carbon utilization such
as the "carbon burnup cell" proposed by
Pope, Evans and Robbins.1 Results from tests
at 3 ft/sec show carbon utilizations with
recycle to range from about 90 to 99 percent.
To consistently achieve acceptable burnup,
i.e., more than 99 percent, recycle might not
obviate the need for a burnup cell.
Heat Transfer
Heat transfer in a fluid-bed boiler is
important in establishing the commercial
potential of this combustion technique and
would also be important in the design of fluid-
bed boilers. In the combustor previously
described, which contained a series of water-
cooled U-tubes immersed in the bed, data
were taken on one U-tube during the combus-
tion tests. Values of Ui were calculated from
the data and values for the water coefficient
(hi) were calculated using the Dittus-Boelter
relationship. Values for the bed-to-tube
coefficient (he) were then calculated by the
relationship
D0hB
(1)
= overall heat transfer
coefficient based on inside
area of pipe
hj = inside film coefficient,
steam
hg = outside film coefficient,
fluid bed
Di,Dav,D0 = inside, average, outside
pipe diameters
x = pipe wall thickness
k = thermal conductivity of
pipe.
Results are given in Table 4 for 18 tests at 3
ft/sec and six tests at 6 ft/sec. The results from
tests at 3 ft/sec are generally consistent, except
for two tests (B-5-L, C-7-L). Neglecting those
two tests, the average bed-to-tube film
coefficient is 67.5 Btu/hr-ft2-°F. At 6 ft/sec,
the average bed-to-tube coefficient was 32.3
Btu/hr-ft2-°F.
Heat transfer also was investigated in
another 18-in. diameter combustor that was
operated to evaluate the performance of
various alloy tubes. This combustor contained
a steam-cooled tube bundle which passed
horizontally through the fluid bed. Figure 4
shows the layout of the tube bundle. During
three tests of approximately 500 hours each in
duration, one at 3 ft/s6c and .two at 6 ft/sec,
heat transfer to the various tubes was
measured.
Heat transfer data from this steam-cooled
tube bundle were examined via graphical
interpretation of overall heat transfer coeffi-
cients. The overall heat transfer coefficient
based on the inside area of the pipe is given by
preceding equation. Since conditions in the
fluid bed were essentially constant in each
long-duration -test, h B should be essentially
constant. Neglecting thermal expansion of the
tubes, Di, Do, D av., and x are constants for the
tubes; k is constant for the various alloys over
the temperature range of the alloys (k is some-
what higher for carbon steel, but the
resistance term for the metal wall is so small
that it is insignificant). Hence, the resistance
terms for the metal wall and the outside film
can be combined into one constant R ]
(2)
Thus, U. =
+ R,
(3)
1-3-6
-------
Table 4. HEAT TRANSFER RESULTS FROM
WATER-COOLED U-TUBE
Test
A5L
A6L
A7L
B4L
B5L
C4L
C5L
C6L
C7L
E4L
E5L
E6L
F4L
F5L
F6L
F7L
G5L
G6L
B6L
C8L
E7L
F8L
G8L
I1L
Limestone size
Uj
Btu/hr-ft2-°F
Fluidizing Velocity = 3 ft/sec
1/4 in. x 10 mesh
3/8 in. x 3/16 in.
3/ 1 6 in. x 30 mesh
3/8 in. x 3/16 in.
8 x 28 mesh
1/4 in. x 10 mesh
3/8 in. x 3/16 in.
8 x 28 mesh
1/4in.x3/16in.
8x28 mesh
8x28 mesh
3/16 in. x 30 mesh
3/8 in. x 3/16 in.
3/8 in x 3/16 in.
3/8 in. x 3/16 in.
8 x 28 mesh
3/8 in. x 3/16 in.
8 x 28 mesh
71.2
68.9
55.4
64.0
36.I
61.3
60.8
62.4
31.2
69.6
69.1
66.0
68.5
79.1
67.7
65.1
70.8
68.8
Fluidizing Velocity = 6 ft/sec
1/4 in. x 3/16 in.
1/4 in. x 3/16 in.
1/4 in. x 3/16 in.
1/4 in. x 3/16 in.
1/4 in. x 3/16 in.
1/4 in. x 3/16 in.
36.9
39.9
37.6
39.9
34.4
34.9
hBed
Btu/hr-ft2-°F
75.5
70.1
52.0
60.9
31.9
61.8
60.9
61.3
26.7
67.7
68.3
64.3
70.1
83.5
70.7
66.0
73.4
72.7
Avg. 67.5a
32.0
34.9
32.3
34.7
29.5
30.1
a Neglecting two abnormally low values.
Avg. 32.3
or
-I- R,
nG°-8
D.0.2
(4)
where:
(5)
For a gas or vapor flowing inside smooth,
circular pipe in turbulent flow, the film
coefficient hj is given by a number of
relationships of the type:
n = a constant that depends on
the physical properties of the
fluid
G = mass velocity of the fluid
D; = inside diameter of the pipe.
1-3-7
-------
For steam over the temperatures in these
tests, n can be considered constant, and D{ is
constant for all the tubes. Therefore,
and
where:
hj - CGU°
1 1
Ti- CG"'®
i
C is a constant.
+ R,
Data from each tube were used to calculate
Q, the rate of heat transfer. Values of Q were
then substituted into the formula Q = UiAAt
to obtain values for Uj. A plot of 1/Ui (as
ordinate) versus 1/G°'8 gives a straight line
with a slope of 1/C. The vertical intercept of
this line, b, represents Rj, the sum of the resis-
tance of the metal wall and the outside (fluid
bed) film.
Figures 5 and 6 show the graphical
interpretations of the overall coefficients of
heat transfer. The linear correlation for the
first test (Figure 5) has a vertical intercept of
0.0127, and a calculated fluid-bed film
coefficient (hs) of 63.4 Btu/hr-ft2-°F. The
correlations for second and third tests (Figure
6) were combined since both tests were made
at 6 ft/sec and with -1/4 in. + 3/16 in.
limestone. In this latter case, the vertical
intercept is 0.0185, and a calculated fluid bed
film coefficient (hB) of 43.3 Btu/hr-ft2-°F.
The graphical interpretations appear
reasonable. Results from the test at 3 ft/sec
with -8 + 30 mesh stone gave a bed coefficient
of 63; results from the two tests at 6 ft/sec with
beds of -1/4 in. + 3/16 in. stone gave a bed
coefficient of 43. The differences between the
two values were caused by differences in
particle size and fluidizing velocity. The
location of data points from the second test
suggests there is a difference in heat transfer
between the top and bottom rows; data from
the first and third tests do not appear to
support this.
At the same superficial fluidizing velocity,
results from the steam-cooled and water-
cooled exchangers' were expected to be
comparable. At 3 ft/sec, the agreement was
quite good,with bed coefficients of 63.4 for the
steam-cooled tubes and 67.5 for the water-
cooled U-tube. At 6 ft/sec, the agreement was
not as good: 43.3 for the steam-cooled tubes
and 32.3 for the water-cooled U-tube.
CONCLUSIONS
At a fluidizing velocity of 3 ft/sec, and
when fines from the primary cyclone are
recycled to the bed, S retention by a limestone
bed increases rapidly to 90 percent as Ca/S is
increased to approximately 2. At 6 ft/sec and
without recycle, it appears that Ca/S must be
at least 3 to retain 90 percent of the S in the
bed.
Carbon burnup was too low for commercial
boiler operation when recycle was not
employed, regardless of the fluidizing velocity.
Even when recycle was used at the lower
velocity of 3 ft/sec, carbon burnup might not
be commercially acceptable so that a separate
burnup cell would be required. .
At 3 ft/sec fluidizing velocity, using beds
ranging in size from 8 x 28 mesh to 3/8- x
3/16-in., heat transfer coefficient from the bed
to a tube immersed in the bed was 60 to 70
Btu/hr-ft2 -°F. At a velocity of 6 ft/see and
with beds of 1/4- x 3/16-in. particles, the
coefficients to steam-cooled and water-cooled
tubes were 43 and 32, respectively.
REFERENCE
1. Bishop, J.W., E.B. Robinson, E. Erlich,
A.K. Jain, and P.M. Chen. Status of the
Direct Contact Heat Transferring Fluid-
ized Bed Boiler. (Presented at Winter
Annual Meeting, ASME, New York. Paper
Number 68-WA/FU-4. December 1-5,
1968.)
1-3-8
-------
PRODUCTS OF
COMBUSTION
CASTABLE REFRACTORY
CARBON STEEL
SHELL, X in.
INSULATING
FIREBRICK
CASTABLE
REFRACTORY
SIGHT GLASS
LIMESTONE
WATER
IGNITER PORT
AIR
NATURAL GAS
AIR DISTRIBUTOR
COAL-AIR MIXTURE
Figure 1. Fluid-bed combustor.
1-3-9
-------
PRODUCTS OF COMBUSTION
TO ANALYZERS FOR 02 , NO, S02, CO, AND COj
CYCLONE
SEPARATOR
LIMESTONE
AIR
NATURAL GAS
RESIDUE
FLUID-BED
COWBUSTOR
Figure 2. Flowsheet for fluid-bed combustion system.
1-3-10
-------
1UU
90
** 80
"5
_T
o
u5 70
cc
u_
_J
c/>
60
50
40
ra O A
* « 191 D
D AA A
6
A
A ?
— •
6
O
A | 6 OA5L
— ^ * A
w 6 13 G5L
O hvab (W.Va)
«
> hvab (Ohio)
A hvbb
A hvcb
L"] mvb (washed)
•
— E
| mvb (unwashed)
0 Ivb —
6 test at 6 ft/sec
•
others ,3 ft/sec
!
1 2 3
CALCIUM /SULFUR, mole ratio
Figure 3. Effect of Ca/S mole ratio on sulfur removal.
- 5/8-in. TUBING, 0.065-in. WALL
1-1/4 in.
H/4i,
TUBE
A
B
C
D
E
F
G
MATERIAL
SEAMLESS CARBON STEEL
WELDED 340 STAINLESS STEEL
SEAMLESS 316 STAINLESS STEEL
SEAMLESS 410 STAINLESS STEEL
SEAMLESS 446 STAINLESS STEEL
SEAMLESS 304 STAINLESS STEEL
WELDED 316 STAINLESS STEEL
GAS FLOW
Figure 4. Arrangement of steam-cooled tube bundle.
i-3-ii
-------
§
• BOTTOM TUBE ROW
OMIDDLE TUBE ROW
• TOP TUBE ROW
10
12
Figure 5. Graphical analysis of overall heat transfer coefficients: 3 ft/sec fluidizing
velocity, -8+30 mesh limestone.
1-3-12
-------
BOTTOM TU BE ROW, 2nd TEST
O MIDDLE TUBE ROW, 2nd TEST
TOP TUBE ROW, 2nd TEST
A ALL TUBES, 3rd TEST
104/GO-8
Figure 6. Graphical analysis of overall heat transfer coefficients: 6 ft/sec fluidizing
velocity, - 1/4 in. +3/16 in. limestone.
1-3-13
-------
4. THE REDUCTION OF EMISSIONS
OF SULPHUR OXIDES AND NITROGEN OXIDES
BY ADDITIONS OF LIMESTONE OR DOLOMITE
DURING THE COMBUSTION OF
COAL IN FLUIDISED BEDS
S. J. WRIGHT
National Coal Board., London, England
INTRODUCTION
It has been known for a number of years, at
least since the First International Conference
on Fluid ised-Bed Combustion held in the
autumn of 1968, that additions of limestone or
dolomite to fluidised-bed combustors could
materially reduce the proportion of the sul-
phur which was emitted in the flue gases as
sulphur dioxide.
As research work progressed, both in the
U.S.A. and the U.K., it became apparent that
there were significant and unexplained differ-
ences between results obtained at different
establishments under apparently similar con-
ditions. At that time the National Coal Board
(NCB) had in operation the most comprehen-
sive range of fluidised-bed combustors avail-
able. In May 1970, therefore, the National
Coal Board and the Environmental Protection
Agency agreed to jointly finance a consider-
able experimental programme designed (1) to
establish the causes of some of the anomalies
in the extant data, (2) to establish, within the
range of the rigs available, the effects of scale,
and (3) to systematically investigate some of
the many variables effecting sulphur retention
in fluidised combustion beds.
The programme was scheduled to cover a
period of 12 months and involved the follow-
ing work at the NCB's Leatherhead and
Cheltenham laboratories.
1. Experiments on a number of pilot-scale
combustors to measure the effect on emis-
sion of sulphur, nitrogen oxides, and
particulates of a selected range of process
conditions; e.g. coal type; the quantity,
size and type of additive (limestone/dolo-
mite) fed to the combustor to retain sul-
phur; combustion conditions as regards
temperature, pressure, and fluidising
velocity; plant scale; and design features
such as bed depth and the recycling of
incompletely reacted fuel .and additive.
2. Experiments on selected pilot-scale com-
bustors to assess the extent to which the
addition of limestone or dolomite to coal
in a fluidised bed in a large test rig
influences the corrosion, erosion, and
deposit formation on specimens
representative of typical evaporator,
superheater, and reheater tube metals.
3. Laboratory scale experiments to charac-
terise the coals and additives used.
4. Development of a mathematical model to
assist in correlating the factors which
influence the pollution control charac-
teristics of a fluidised combustion system.
1-4-1
-------
THE SCOPE OF THE RESEARCH
PROGRAMME
The main objectives of the research
programme were:
1. To assess the effectiveness of the fluidised-
bed combustion process, with and without
the addition of limestone or dolomite,
towards the reduction of SO2 emission; to
show those operating parameters that
significantly affect the attainment of the
immediate target emission (300 ppm v/v);
and to indicate how these data may affect
plant design.
2. To gather data, over the same range of
operating conditions, on the levels of NOX
emission that occur during the combus-
tion process when the SO 2 is partially
absorbed by added limestone or dolomite.
3. To measure the particulates elutriated
from the fluidised-bed combustor in order
to provide data for the design of a particu-
lates removal system which will reduce
atmospheric emissions to an acceptable
level.
4. To contribute towards an understanding
of the way in which the porous properties
of limestone or dolomite affect SO 2
retention under the conditions prevailing
in a fluid-bed combustor, and to develop a
simple method of classifying limestones
and dolomites according to their utility for
SO? retention in the fluidised-bed
combustion process.
5. To develop a mathematical model of the
retention of SOa in a fluid-bed combus-
tion system to allow the performance of
new plant, with respect to SO 2 emission,
to be predicted at the design stage from
the design and other basic data.
6. To study corrosion of typical steels used in
boiler construction when immersed in a
fluid-bed burning coal, both with and
without the addition of limestone/dolo-
mite.
The research programme to meet these
objectives was organised into eight tasks, at
various plants or locations as follows:
1. To compare the performance of the 36-in.
rig with that of the 6-in. rigs at C.R.E. and
Argonne, and to extend the range of
operating conditions for which experi-
mental data are available (36-in.
combustor, CRE).
2. To obtain, for operation under pressure,
data on the emission of sulphur and nitro-
gen oxides and on corrosion/deposition of
boiler metal and turbine blade specimens
(48- x 24-in. pressurised combustor,
BCURA).
3. To carry out long-term tests to assess the
effect of limestone addition on corrosion
of evaporator, superheater, and reheater
metals immersed in the fluid bed (27-in.
combustor, BCURA).
4. To obtain data on corrosion of evaporator,
superheater, and reheater materials for
lower fluidising velocities (12-in. combus-
tor, CRE).
5. To obtain data on sulphur retention for a
range of coals and limestones, in particu-
lar to allow comparison to be made with
the 6-in. rig at Argonne (6-in. combustor,
CRE).
6. To complete the development of a
mathematical model of sulphur retention,
to compare its predictions with the results
of laboratory and rig experiments, and to
up-date it as appropriate (mathematical
work, BCURA).
7. To investigate the distribution of sulphur
in a range of coals and in the residue from
the rigs (laboratory work, CRE/BCURA).
8. To investigate the pore structure and
related factors that affect sulphur reten-
tion by lime (laboratory work, BCURA).
The features of the combustors used, the
range of operating conditions explored, and
the aggregate number of test hours
1-4-2
-------
accomplished are shown in Tables 1 and 2. It
will be noted that experiments were carried
out at combustion pressures up to 5 atmos-
pheres absolute, fluidising velocities up to 11
ft/sec, bed temperatures up to 1680°F, using
four different coals, three limestones, and two
dolomites; the test running time totalled 5300
hours.
MATERIALS USED
The experimental work in the programme
was carried out using four coals, three lime-
stones, and two dolomites. Typical analyses of
the materials are given in Tables 3 and 4.
DISCUSSION OF THE RESULTS
There were considerable differences, both
in size and geometry, between the various rigs
used in the programme. For instance:
1. Rigs were either circular or rectangular in
cross section and ranged in area from 0.2 to
8.0ft2 .
2. The geometries of cooling surfaces within
the bed ranged from deep banks of closely
spaced 1-in. diameter tubes to relatively
shallow banks of widely spaced 2.4-in. di-
ameter tubes.
3. The area of the cross section served by a
single coal feed point varied from 0.2 to 4.5
ft2.
4. Some combustors had only internal fines
recycle systems, some had both internal
and external recycle systems, and others
had only external recycle systems. Results
referred to as being without recycle are
from rigs with only external, and hence
controllable, recycle systems.
5. In some combustors the walls of the bed
and freeboard were uncooled; in others
they were cooled throughout.
Despite those differences it was found that
geometry as such was not a variable; the whole
body of the results could be discussed in terms
of the process variables.
Table 1. MAIN FEATURES OF THE PILOT-SCALE COMBUSTORS
Feature
Designation
Bed cross
section
Bed depth, ft
Operating
pressure, atm abs.
Fluidising
velocity, ft/sec
Coal rate,
Ib/hr
Total
running hours
Location
BCURA, Leatherhead
27 in.
27 in. dia
1.5-2.0
1
6-11
200-300
2150
48 in. x 24 in.
(pressurised)
48 in. x 24 in.
3.5-4.0
up to 5
2
300-500
430
CRE, Cheltenham
36 in.
36 in. x 18 in.
2-7
1
2-8
75-300
1000
12 in.
(corrosion)
12 in. x 12 in.
2
1
3
20 - 25
1100
6 in.
6 in. dia
2-3
1
2-3
4-6
600
1-4-3
-------
Table 2. THE VARIABLES EXPLORED
Operating
variables
Coal
Ash content, %
Volatile matter, %
Sulphur, %
Chlorine, %
H20asfed,%
Ash fusion, °F
Size
Bed depth, ft
Temperature, °F
Fluidising velocity, ft/sec
Excess air, %
Recycle of cylone
fines
Additive
Ca/S mole ratio
Pilot-scale combustor(s)
Non-pressurised
U.S. Pittsburgh
U.S. Illinois
U.K.Welbeck
U.K. Park Hjll
12-18
37-47
1 .3 - 4.4
0.1 -0.6
1-10
1800-2600
-1/8in.and-1/16in.
1.5-7
1420-1680
2-11
-12 to +29
Zero, partial, full
U.S. Limestone 18
U.S. Limestone 1359
U.K. Limestone
U.S. Dolomite 1337
0-6
Pressurised
U.S. Pittsburgh
U.K.Welbeck
13-18
30-41
1.3-3.1
0.1-0.6
1-6
2100-2600
-1/16in.
3.5-4
1470
2
1 1 to 33
Partial
U.S. Limestone 18
U.S. Dolomite 1337
U.K. Dolomite
0-3
The order in which the operating variables
are commented upon takes into account both
their relative importance and some of the
interactions; e.g., through their effect on gas
and solids residence times.
Pittsburgh coal (3 percent sulphur) was
used with either Limestone 18 or Dolomite
1337 in the majority of experiments. While it
is believed that most of the comments in the
following statement of the main findings apply
to other coals and limestones or dolomite, they
refer primarily to these materials unless other-
wise stated.
Ca/S Mole Ratio: The SO2 is reduced
asymptotically to zero as the feed rate of addi-
tive to the fluidised bed is increased. The
percentage SO 2 reduction obtained at a given
operating condition is a function of the mole
ratio of added calcium to sulphur in coal; it is
almost independent of the sulphur content of
the coal, since the reaction is approximately
first order with respect to SO 2 concentration.
Clearly, in order to obtain a specified concen-
tration of SO2 in the off-gas when burning a
coal of high sulphur content, it is necessary to
achieve a higher percentage SO 2 reduction by
using a higher Ca/S mole ratio.
For a given coal the lowest values of the
Ca/S mole ratio were required at (1) low
fluid ising velocities (i.e. 2 to 3 ft/sec), (2) a bed
1-4-4
-------
Tables. TYPICAL ANALYSES OF COALS USED
Analysis
Proximate analysis
Total moisture, % a.r.
Ash, % a.r.
Volatile matter, % a.r.
Ultimate analysis
Carbon, % d.b.
Hydrogen, % d.b.
Nitrogen, % d.b.
Sulphur, % d.b.
Oxygen + errors, % d.b.
Chlorine, % d.b.
Calorific value (d.a.f .), Btu/lb
Swelling number
Gray King coke type
Ash analysis
CaO, %
MgO, %
NaaO, %
K2O, %
S!O2,%
Size
(as received)
Coal
Illinois
9.8
11.8
46.6
67.8
4.5
1.3
4.4
8.5
0.2
14,300
4-1/2
D
10.1
1.0
1.7
1.8
40.8
-1/4 in.
Pittsburgh
1.6
13.5
41.1
71.7
4.5
1.4
2.8
4.4
0.1
15,100
8
G9
8.0
1.3
0.7
1.6
45.8
-1/4 in.
Park Hill
2.1
16.5
39.2
68.2
4.4
1.3
2.5
5.3
0.1
14,750
1
D
2.2
1.7
0.8
3.6
46.0
-1-1/2 in.
Welbeck
4.2
18.2
38.3
67.5
4.3
1.5
1.3
5.1
0.6
14,400
1
C
1.8
1.4
1.8
3.2
57.5
-1-1/2 in.
Table4. TYPICAL ANALYSES OF LIMESTONES AND DOLOMITES
Component
CaO
MgO
H20 + C02
Si02
Fe203
S03
Total
Dolomite
1337
28.9
22.9
47.4
0.5
0.2
-
99.9
U.K. Dolomite
29.3
21.5
46.3
-
0.1
97.2
Composition, %
Limestone
18
,45.7
1.4
36.6
13.6
0.3
-
97.6
Limestone
1359
55.7
0.3
43.6
0.5
0.1
-
100.2
U.K. Limestone
55.4
0.3
43.5
0.7
0.1
-
99.8
1-4-5
-------
temperature of around 1500°F, and (3) when
most of the fines larger than about 10 Mm were
recycled. See Figure 1.
Bed Temperature: The optimum bed tem-
perature was between 1400° F and 1600°F.
The level of SC)2 emission and the change in
emission with change of temperature on each
side of the optimum appeared to depend on
the type of additive used and to some extent on
the Ca/S ratio employed. The increase in
emission on either side of the minimum
tended to be greater at low than at high Ca/S
ratios; i.e. under conditions where the fraction
of calcium sulphated was higher. The opti-
mum temperature was found to be 1500-
1550°F for limestone additive and 1400°F-
1500°F for the dolomite. The data suggest
that maintaining the same level of sulphur
emission (e.g., 85 percent sulphur retention)
at, for example, 100° F above the optimum,
would involve increasing the Ca/S ratio by a
factor of about two. The effect of changing
bed temperature was not investigated on the
pressure combustor.
The rapid increase in sulphur emission at
bed temperatures above about 1550°F is
unexpected from laboratory measurements of
the reaction rate between CaO and SC»2. Since
the effect of temperature appears to be rever-
sible (i.e., the SC>2 reduction reverts to a high
value as soon as the bed temperature is
reduced) it cannot be accounted for by irrever-
sible factors such as sintering or slag forma-
tion at the particle surface. One tentative
explanation postulates that an oxygen-
containing species (e.g., hydroxide ions
derived from traces of water, which are known
to be difficult to remove) is involved in the
conversion of CaSOs to CaSCM. It is possible
that above the optimum temperature the
hydroxide ions become more mobile and
hence less able to participate in the reaction.
At low temperatures (i.e., below 1350°F)
sulphur retention with dolomite was higher
than with limestone, because of the lower
calcination temperature of the MgCOs in
dolomite which leads to the development of
1-4-6
pore structure below the temperature of
calcination of the CaCOs. See Figure 2.
Fluidising Velocity: Increase in fluidising
velocity resulted in an increase in sulphur
emission. An empirical correlation was
derived and is reported later in the summary.
Increase in velocity without any compensating
action results in reduction in both gas and
solids residence times. To maintain the same
sulphur retention (e.g., 85 percent, at 8 ft/sec
as at 2 ft/sec fluidising velocity) the Ca/S mole
ratio (at a bed temperature of 1500°F and
without recycle) would have to be increased
from about 2 to about 4 (Figure 3).
/
Bed Height: Increase in bed height usually
resulted in a reduction in SCh emission. An
empirical correlation is reported later in the
summary. In principle it should be possible to
counteract the adverse effect of increasing
velocity by a proportionate increase in bed
height. At atmospheric pressure the attendant
increase in pressure loss for other than a small
increase in bed height could be prohibitive,. In
addition, because the tube bank required
would occupy only a part of the bed height, the
effectiveness of increasing bed height may be
reduced by the formation of large gas bubbles.
The effect of bed height in super-charged
boilers is potentially of greater significance.
Here the deep banks of close packed tubes
may assist in breaking up large gas bubbles
and hence may improve the contact between
gas and solids. Further, the increase in
pressure loss due to increasing bed height is
less important under pressure. See Figure 4.
Fines Recycle: A high proportion of the addi-
tive is elutriated from the bed before being
fully utilised. By efficient recycle of fines
larger than 10 urn, to the bed, SOi reduction
was increased significantly; e.g., from 73 to 99
percent at a fluidising velocity of 2 ft/sec and a
Ca/S mole ratio of 1.6.
Operating Pressure: The effect of operating
pressure on SO2 reduction was negligible
when dolomite was used as an additive. This is
to be expected with a reaction which is first
order with respect to the partial pressure of
-------
SO2. With limestone as an additive, the
reduction obtained at 5 atm was appreciably
lower than with dolomite or with limestone at
atmospheric pressure. This was also to be
expected, since at 1470 °F, calcination of lime-
stone to give a porous structure would not
occur at operating pressures above 2 atm.
Penetration of the particle by SO 2 would
therefore be difficult and only a surface layer
of sulphate would form. It was found.however,
that the performance of the limestone was
better than this reasoning would imply; it
suggests that the exposure of fresh surface by
attrition plays a significant role. Nevertheless,
from the point of view of both the Ca/S mole
ratio and the total quantity of additive
required to attain a target level of sulphur
retention, dolomite was superior to limestone.
To retain 85 percent of the sulphur, for
example, the estimated Ca/S mole ratios for
dolomite and limestone were 1.1 and 3.25
respectively; the estimated quantities of
additive were 7.6 Ib and 12 Ib per Ib of sulphur
removed, respectively.
Particle Size: For coarsely crushed limestone
the percentage SOi reduction increased when
the particle size of limestone was reduced; this
effect was probably due to the consequent
increase in available reaction surface. On the
other hand, with dolomite there was no effect
of particle size, suggesting that access to
internal surface is not a limiting factor for
dolomite.
With additive ground to -125 pm or -150
urn, it was found that the fluidising velocity
had a profound effect on SO 2 reduction.
Whereas at low velocity (3 ft/sec) the fine
additive improved SOa reduction; the reverse
was true at high velocity (8 ft/sec). The data of
Pope, Evans and Robbins suggest that, with a
Ca/S mole ratio of 2.6 in beds 10 in. deep
fluidised at 12 ft/sec, limestone 1359 gave 80
percent SO2 reduction when ground to 44 nm,
and 60 percent reduction when ground to -74
Mm. Evidently the SOj reduction is very
sensitive to the size of finely ground particles,
so the Pope, Evans and Robbins data are at
least qualitatively consistent with those for
-150 vm limestone from the present study.
Limestone ground to -150 urn may have too
short a residence time at high velocity to
achieve a high degree of sulphation; superfine
material will become highly sulphated, since
its residence time will not be markedly less
than that of -150 Mm material. However,
superfine material may cause serious gas-
cleaning problems.
Type of Additive: The type and source of
additive affects the reduction in SO 2 that can
be achieved. At atmospheric pressure Lime-
stone 18 was the most effective additive on
both molar and weight bases; the least effec-
tive on a mole basis was Limestone 1359, and
on a weight basis Dolomite 1337. To achieve
the same level of retention with the poorer
Limestone 1359 as with the. Limestone 18
would require an increase of up to 100 percent
in the Ca/S mole ratio. As mentioned
previously, for operation under pressure both
the dolomites were superior to Limestone 18
on weight and molar bases.
An important finding from the point of
view of simplifying prediction of suitability of
stones was that measurements made at room
temperature and at combustor temperatures
showed the same accessibility of the structure
to gases of similar molecular size to SO 2.
Temperature cycling (as may occur in some
plant designs when elutriated particles are
recirculated) does not affect the pore structure
significantly from the point of view of SOi
uptake. An empirical reactivity test was
considered to be the most economic method
for classifying stones. For limestones the
results of laboratory experiments give pessi-
mistic predictions of plant performance.
These results are thought to be because the
tests do not take into account the beneficial
effect of attrition in the combustor which
results in removal of the sulphated surface
layer. The effect of attrition was particularly
important for limestones in the pressurised
combustor and for Limestone 1359 at atmos-
pheric pressure. For dolomite, access to the
1-4-7
-------
internal surface of the particles does not
appear to be a limiting factor.
Whereas thermal losses are incurred from
the sensible heat requirement and the heat of
calcination of an additive, the heat of sulpha-
tion represents a thermal gain. Up to a Ca/S
mole ratio of about 2 using limestone
(sufficient to retain 85 percent of the sulphur
of a 3 percent sulphur coal under good
operating conditions), it is estimated that the
heat of sulphation will counterbalance the
sensible heat requirements and heat of calcin-
ation. With dolomite, however, the net
thermal loss would be about 1-1/2 percent of
the coal heat input. Under pressure calcina-
tion of CaCOs is inhibited, and the thermal
loss incurred by using dolomite would be
negligible for Ca/S mole ratios up to 2.
Type of Coal: The most important coal
property in this context is the sulphur content,
which determines not only the quantity of
additive required for a given Ca/S mole ratio,
but also the percentage SOa reduction (and
hence Ca/S mole ratio needed) to meet set
limits of SOi emission. Since the SOi
absorption reaction is first order with respect
to SO 2 concentration, it could be expected
that the same relationship between percentage
reduction and Ca/S mole ratio would hold for
all coals irrespective of the sulphur content.
However, the experimental results showed
that, for a given Ca/S mole ratio, similar SCh
reductions were obtained for three of the
coals, but the reductions were up to 15 percent
higher with Welbeck coal. Differences in the
rate of sulphur release have been found
between coals and might partly account for
differences in performance. A more likely
explanation of the higher SO 2 reduction with
Welbeck coal is its low sulphur content, which
had the consequence that additive was fed at a
lower rate and hence had a longer residence
time. This could have resulted in the higher
degree of sulphation, particularly if particle
attrition was an important effect.
Plant Design: As mentioned earlier, it was
concluded that despite the difference of scale
1-4-8
and design over the range of combustors used
there was no significant difference between the
SO2 reductions obtained in different combus-
tors with the same operating conditions.
Nevertheless, direct application of the present
results to combustors of commercial size
requires some caution. Some factors which
might alter the SO2 emission, such as the
depth of the tube bank, have already been
mentioned. Further, observation of a radial
distribution of SO 2 concentration in the free-
board of one of the larger pilot plants suggests
that coal feed spacing, if greater than that
used in the pilot-plants, may assume signifi-
cance in commercial boilers.
Mathematical Model and Correlation of Data:
The mathematical model has been developed
to give fairly satisfactory prediction of the con-
sequence of changing some operating
conditions. Additional development is needed:
(1) to take further account of attrition of addi-
tive and (2) to extrapolate the results to
combustors that differ significantly from the
present pilot plants. The model in its present
form has not been useful in correlating the
experimental data. However a number of
empirical correlations have been derived as
follows.
There is an approximately exponential
relationship between the SO 2 reduction and
the Ca/S mole ratio of the form
R = 100[l-exp(-MC)] (1)
where: R = percentage SO 2 reduction
C = Ca/S mole ratio
M = empirical constant depending
on the coal, limestone, and
operating conditions.
The effect of fluidising velocity on SO 2
reduction may be approximately correlated by
A = X,/V (2)
where: A = absorption ratio, defined as
R/UOO-R)
V = fluidising velocity
X] = empirical constant depending
on the Ca/S mole ratio and
other operating conditions.
-------
The effect of bed height on SO2 reduction
may be approximately correlated by
= X2H
(3)
where: H
X,
bed height
empirical constant depending
on the Ca/S mole ratio and
other operating conditions.
Emission of Nitrogen Oxides
Emission of NOX from the pressure
combustor (50-200 ppm) was significantly
lower than from the non-pressurised combus-
tors (300-600 ppm). All the combustors
produced less NOx pollution, both with and
without the use of additives, than is common
with conventional plant. The reason for the
superior performance is uncertain. NOx
emission could not be correlated with SO2
emission, although on some occasions a
decrease in the SO2 emission (due to feeding
limestone or dolomite) was accompanied by an
increase in the NOx emission. It was
concluded that more information was needed
on the mechanism of NOx formation before a
contribution could be made towards reducing
emission.
Emission of Alkalis and Chlorine
As expected, the low combustic«i tempera-
tures in fluid-bed combustors resulted in low
alkali emissions. The combustion gases from
the pressure combustor contained about 2
ppm of Na; i.e., about one tenth of the lowest
concentration reported for the gases from
conventional plant. The concentration of K
was less than 0.5 ppm. Higher emissions were
measured when limestone was added to the
pressurised combustor instead of dolomite (5
ppm of Naand 1.5 ppm of K) and from one of
the non-pressurised combustors that was
being operated at the higher bed temperature
of 1560°F (6 ppm of Na and 3 ppm of K). As
expected most of the chlorine of the coal was
released into the combustion gases.
Emission of Particulates
Particulate matter elutriated from
fluidised-bed combustors comprises 5 to 15
percent of the carbon and 80 to 100 percent of
the ash and additive. By using primary and
secondary cyclones having collection efficien-
cies of 90 percent at about 10 Mm it was
possible to collect 95 - 98 percent of this
material to give dust emission of 0.2 - 0.6
gr/scf. Within this range the emission was
approximately proportional to the feed rate of
ash plus additive. Increasing the fluid ising
velocity increased elutriation from the bed,
but because of more efficient cyclone opera-
tion with higher gas flow rates there was little
effect on emission. Fines recycle in the 36-in.
combustor increased the dust emission to 1.4
gr/scf. The pressurised combustor had an
internal recirculation cyclone in addition to
primary and secondary cyclones; dust
emissions in the range 0.05-0.1 gr/scf were
obtained.
Based on these results it is unlikely that
there would be any problem in meeting
projected statutory limitations on particulate
emission.
Corrosion and Deposition
The addition of limestone or dolomite had
no significant effect on corrosion or deposition
of tubes in the bed or in the gas space under
the range of operating conditions likely to be
experienced in a commercial plant.
The amount of material settling on the
turbine blade cascade at the outlet of the
pressure combustor was slight and was judged
to be unlikely to affect turbine performance.
There were no signs of sintered deposits or
erosion.
CONCLUSIONS
The main conclusions reached from the
work are:
(1) With fluidised combustion and the addi-
tion of limestone (or dolomite) the
1-4-9
-------
emission of sulphur oxides from coal
burning power plant can readily be con-
trolled to meet the very rigorous restric-
tions (100 ppm v/v SO2) planned for
certain densely populated areas in the
U.S. For a power plant burning a 3 per-
cent sulphur coal this would involve feed-
ing sufficient additive to retain 95 percent
of the sulphur. Under the best combina-
tions of operating conditions about 1.8
times the stoichiometric quantity of addi-
tive would be required; for a 100-MW
plant this would involve supplying 160
ton/day of limestone or 280 ton/day of
dolomite. The less stringent restrictions
that have been proposed for built-up areas
(300 ppm), and for power stations
generally in the U.S. (700 ppm), would
require sulphur retentions of 85 percent
and 67 percent respectively, for 3 percent
sulphur coals. These limits can be met
under a wider range of operating con-
ditions and/or at less expense for
additives.
(2) Emission of oxides of nitrogen from
fluidised combustion systems can be
expected to be at least 60 percent less than
from conventional combustion systems
but additional measures would be needed
over and above those used for SO2
reduction to meet the very stringent
restrictions envisaged for the latter part of
the century (i.e., 100-200 ppm).
(3) The particulates emitted from fluidised-
bed systems are unlikely to cause prob-
lems in meeting current or possible future
restrictions.
(4) The use of limestone/dolomite additive to
restrict sulphur emission is unlikely to
affect adversely the exemplary behaviour
of the fluid-bed combustion system from
the point of view of (a) fouling and
corrosion of tubes immersed in the bed
and (b) deposition or erosion of turbine
blade materials exposed to the com-
bustion gases.
In terms of sulphur retention the most
important variable is the Ca/S mole ratio. The
most stringent requirement for SOa emissibns
yet proposed can readily be met if sufficient
calcium is present in the bed. If economic
factors require it, the usage of
limestone/dolomite can be minimised, by
reducing the design fluidising velocity. This
will make the boiler bigger and hence more
expensive.
In terms of boiler operation and control the
most important variable is bed temperature. It
has been shown that under some conditions
the efficiency of sulphur retention is very
sensitive to bed temperature; for ease of boiler
start-up and flexibility during load following it
is useful to be able to let the bed temperature
vary through the maximum allowable range.
For American coals this range is probably
from about 1460 to 1800°F.
ACKNOWLEDGMENTS
The author wishes to thank the National
Coal Board for permission to publish this
paper. Any views expressed are his own and
not necessarily those of the Board.
This paper is no more than a brief
summary of research work which was fully
written up in a Main Report and nine
Appendices involving the efforts of more than
20 people.
The author, therefore, gratefully
acknowledges the efforts of all the following:
The NCB Contract Manager was D. H.
Broadbent, assisted by S. J. Wright.
The research programme was directed by A.D.
Dainton (CRE) and H.R. Hoy (BCURA). The
project co-ordinator was D.J. Loveridge. The
pilot plant experimental work at CRE was
administered by J. McLaren. The following
personnel were involved in the tasks into
which the programme was divided.
1-4-10
-------
Task I The project leaders were D. C.
Davidson and D. F. Williams; A.
A. Randell was responsible for
operation of the plant; D. G. Cox
and J. Highley carried out the
data processing and assessment
of results.
Task II The project leader was A. G.
Roberts; D. M. Wilkins was
responsible for operation of the
plant; J. E. Stantan carried out
the data processing and assess-
ment of results.
Task III The project leader was D. J.
Loveridge; M. H. Barker was
responsible for operation of the
plant; D. M. Wilkins carried out
the final data processing and
assessment of results.
Task IV The project leader was M. J.
Cooke; B. J. Bowles was
responsible for operation of the
plant; E. A. Rogers carried out
the corrosion studies.
Task V The project leader was D. C.
Davidson; A. W. Smale was
responsible for operation of the
plant; D. G. Cox, J. Highley and
J. Holder carried out the data
processing and assessment of
results.
Task VI The project leader was D. W.
Gill; he was assisted by F. V.
Bethell and B. B. Morgan.
Task VII The project leader was D. C.
Davidson; the experimental work
was carried out by R. F.
Littlejohn.
Task VIII The project leader was D. H. T.
Spencer; he was assisted by A. A.
Herod and B. A. Napier.
Additional work to obtain data
for the mathematical model was
carried out by F. V. Bethell and
G. McDonald.
Work on emission of NO x was carried out by
J. T. Shaw.
The monthly and quarterly progress
reports were prepared by D. C. Davidson and
D. J. Loveridge. Assessment of the experi-
mental results for the Main Report was
carried out by J. E. Stantan, J. Highley, and A.
G. Roberts. The Appendices to the Main
Report were edited by J. Highley and W. K.
Joy.
The OAP representative in the U. K. was E.
L. Carls. His valuable contribution both in the
experimental work and in the preparation of
the progress reports and the final report is
acknowledged.
GLOSSARY OF TERMS
Absorption ratio; SOj reduction divided by
100 minus SO2 reduction.
Ca/S mole ratio: moles of calcium in additive
divided by moles of sulphur in coal.
Excess air: air input minus stoichiometric air
tor coal input divided by stoichiometric air
for coal input, times 100 percent.
Fluidising velocity: volume flow rate of gas at
combustion temperature and pressure
divided by cross section of combustor
(neglecting tubes).
SO 2 reduction: SO2 emission without additive
minus SO 2 emission with additive divided
by SO2 emission without additive, times
100 percent.
Sulphur retention: sulphur in coal minus sul-
phur in gas divided by sulphur in coal,
times 100 percent.
Unburnt carbon loss: unburnt solid carbon
divided by carbon in coal input, times 100
percent.
Utilisation of additive: moles of sulphur
retained by additive divided by moles of
calcium in additive, times 100 percent.
1-4-11
-------
100
60
CM
O
o
o
40
PITTSBURGH COAL
LIMESTONE 18 (-1680 urn)
BED DEPTH: 2ft
FLUIDISING VELOCITY: 3 ft/sec
BED TEMPERATURE: 1470 °F
NO RECYCLE
O 6-in COMBUSTOR
• 36-in COMBUSTOR
20
1 2 3
Ca/S MOLE RATIO
Figure 1. Comparison of S02 reduction in 6-in. and 36-in. combustors.
1-4-12
-------
100
v
O
80
60
UJ
O£.
PITTSBURGH COAL
LIMESTONE 18 (-1680 urn)
BED DEPTH: 2 ft
FLUIDISING VELOCITY: 4 ft/sec
NO RECYCLE
36-in. COMBUSTOR
Ca/S MOLE RATIO: 2.2
40
PITTSBURGH COAL
DOLOIVIITE1337 (-1680 iim)
BED DEPTH: 2 ft
FLUIDISING VELOCITY: 4 ft/sec
NO RECYCLE
36-in. COMBUSTOR
Ca/S MOLE RATIO: 2.7
1400 1500 1600
BED TEMPERATURE, °F
1400 1500 1600
BED TEMPERATURE, °F
100
80
1700
I 60
LU
s
40
PITTSBURGH COAL
LIMESTONE 18 (-3175 Jim)
BED DEPTH: 2 ft
FLUIDISING VELOCITY: 8 ft/sec
WITH RECYCLE
27-in. COMBUSTOR
Ca/S MOLE RATIO: 2.8
PITTSBURGH COAL
DOLOMITE 1337 (-1587 pi)
BED DEPTH: 2 ft
FLUIDISING VELOCITY: 8 ft/sec
WITH RECYCLE
27-in. COMBUSTOR
1400
1500
1600
1400
1500
1600
BED TEMPERATURE, °F BED TEMPERATURE, °F
Figure 2. Effect of bed temperature on S02 reduction.
1700
1-4-13
-------
VELOCITY, ft/sec
5 3
D/
/
r1
—A-
85
0*
a*
"I
E
60 £
50
0.1
0.2 0.3 0.4
1/VELOCITY, ft'Vsec
0.5
SYMBOL
•
o
A
V
•
D
COAL
PITTSBURGH
11
WELBECK
11
It
PITTSBURGH
ADDITIVE
18
11
U.K. L'STONE
11
11
18
SIZE,
jum
3175
1680
11
11
3175
11
BED DEPTH,
ft
2
2
2
2
2
2
BED TEMP.,
°F
1560
1470
?!
11
1560
TT
Ca/S
MOLE RATIO
1.7
2.2
0.8
1.8
2.8
2.7
36-, 27-, AND 6- in. COMBUSTORS
Figure 3. Empirical relation between SO2 reduction and fluidising velocity.
1-4-14
-------
oe
I
1
SYMBOL
•
0
A
V
COAL
WELBECK
PITTSBURGH
ILLINOIS
1)
ADDITIVE
U.K. L 'STONE
1337
1359
ii
SIZE,
jim
3175
11
1680
Tf
VELOCITY,
ft/sec
8
TT
3
1)
BED TEMP.,
°F
1560
»»
1470
11
Ca/S
MOLE RATIO
2.8
5.0
1.1
2.2
36- AND 6-in. COMBUSTORS
Figure 4. Empirical relation between 802 reduction and bed depth.
1-4-15
-------
5. SELECTIVE EXTRACTION OF CLINKER
AT THE BOTTOM OF A
DEEP SELF-AGGLOMERATING FLUIDIZED BED
A. A. GODEL
Societe Anonyme Activit
In our former Conferences, very few
communications have dealt with self-agglom-
erating fluidized beds. To my knowledge, the
only case of the subject which has been studied
— at least in the field of industrial
achievements — was the use of agglomerating
fluidized beds for coal combustion via the
"Ignifluid Process" which has been the sub-
ject of several reports given here: in 1968, by
myself on behalf of the Activit Company, and
in 1970, by Mr. Svoboda, Manager of the
Babcock-Atlantique Company, and by Mr.
Demmy, Vice-President of the U.G.I.
Corporation.
Professor Squires has honored me by
crediting me with having brought to light the
process in which self-agglomeration in a
heavily turbulent fluidized bed results from
preferential bonding of slag particles when
they reach their sintering temperature (about
1100°C for most coal ash); on the contrary,
when slag particles—even adherent ones—
encounter coal particles, there is little chance
of their agglomeration. This phenomenon
stems from laws governing the probability of
encounter of particles.
The result of our study, quite fortunately,
is the possibility of forming slag agglomerates
in a state of quasi-purity. Slag agglomerates
fall to the bottom of the bed when they acquire
sufficient weight and must be eliminated
promptly in order to avoid blocking fluidi-
zation. Extracting ash by slag agglomeration
is doubly interesting, for it also permits rein-
jecting the dust carried over into the fluidized
bed since it has no excess of cinders.
In most cases, this leads us to prefer using
agglomerating fluidized beds which require no
temperature limit.
As I stated here in 1968, the Ignifluid
combustion process achieves extraction quite
simply by using an inclined upward-moving
fluidization grate. This grate supports the
fluidized bed with clinker deposited at the
bottom; it passes through the surface of the
bed, operating thereby like a clinker extractor.
Such grates are relatively narrow and may
be easily placed under rectangular boilers.
This position has enabled their successful
industrial development in equipping boilers of
various steam output for more than 16 years
by the Babcock-Atlantique Company (exclu-
sive license-holder for the process). A 60-MW
Ignifluid power plant comprising two boilers
has been in successful operation for the past
three years. Equipping larger boilers raises no
problems; several projects have been designed
with such equipment and exported from
France since the use of coal is constantly
decreasing in Europe.
Having stated this, I should like to discuss
the essential reason for my communication—a
new process for extracting slag at the base of a
deep fluidized bed. This new process is quite
different from the former in both the means
1-5-1
-------
used and the aims envisaged, although both
the Ignifluid Process and the new one operate
in self-agglomerating fluidized beds.
Our goal in the new process is principally to
achieve certain chemical reactions in deep
fluidized beds, sometimes under pressure;
e.g., the processing of mineral ore, the
processing of chalky marl for producing
cement clinker, etc.
It should be noted that the new process
may be adapted to the use of beds for fuel
gasification in conjunction with appropriate
complementary treatments, such as
desulphurization and, in certain cases, cataly-
tic reactions.
In my invention, I have been guided by the
obvious principle that it is easier to sort slag in
a shallow bed than in a deep one. I have
therefore combined the use of two fluidized
beds of different depths, profiting from the
communicating vessel principle or the "diving
bell" principle, as you will see in Figure 1.
(The figures are produced simply to give a
theoretical explanation of the process.)
Figure 1 shows a horizontal grate
supporting a fluidized bed (A) of a certain
depth and containing slag at its base. I shall
call this bed the "principal fluidized bed." On
the right, communicating with the first one by
opening (O), is a shallow fluidized bed (B), also
containing slag at its base, which I shall call
the "auxiliary fluidized bed."
As you will notice, the auxiliary fluidized
bed is contained in a closed space (C), consti-
tuting what I shall call a "fluidization cell."
Since gas escaping from the auxiliary
fluidized bed finds no outlet to the right, it
escapes on the left through opening (O) which
communicates with the principal fluidized
bed.
Under these conditions, the auxiliary fluid-
ized bed is in hydrostatic equilibrium with the
principal fluidized bed; i.e., the pressure is the
same at corresponding levels.
1-5-2
Theoretically, the surface of the auxiliary
fluidized bed should thus be horizontal; but
the fluidization gas leaving this bed and flow-
ing laterally through its surface toward one
end causes the surface to assume a concave
shape, dropping away considerably toward the
opposite end, as shown.
Figure 2 shows the same theoretical layout,
but an air intake is provided at the end of the
fluidization cell by opening (d) in order to
increase the flow escaping through (O) toward
the principal fluidized bed (A). The result is
that the surface of the auxiliary fluidized bed
(B) drops away much mor,e quickly than shown
in Figure 1, until it lays bare the slag lying on
the bottom of the auxiliary fluidized bed.
Perfect separation is thus achieved between
slag particles and the fluidized matters.
To practically implement the new process,
we use a cylindrical or cylindro-conical
reactor, fitted at its base with a circular fluid-
zation grate which supports a principal fluid-
ized bed and a very shallow auxiliary fluidized
bed, the latter being contained in a small
fluidization cell immediately underlying the
principal fluidized bed.
The grate and the fluidization cell are in
relative motion to one another; the fluid-
ization cell is wide open at the base to permit
communication between the two fluidized
beds, which are in hydrostatic equilibrium as I
mentioned previously.
The direction of the motion of the grate in
relation to that of the cell is such that the slag
deposited on the grate penetrates into the cell,
where slag particles are decanted, extracted,
and sorted outside the principal fluidized bed
and finally poured into an ash pit in a state of
quasi-purity.
These principles offer the possibility for a
variety of embodiments, of which I shall
mention only two.
The isometric projection in Figure 3
represents a cylindrical reactor (a) with a
horizontal grate (b) revolving around its ver-
tical axis (x-y). The fluidization cell (c) is fixed
-------
and provided with an outlet (d) toward an out-
side ash pit, not shown.
The reacting and fluidizing gas is injected
by pipe (e); gas resulting from the reaction is
evacuated by pipe (f) after having passed
through a cyclone (g), which assures collection
and reinjection of fine dust into the fluidized
bed. Pipe (h) supplies the reactor with gran-
ulated solid material to be treated.
The grate revolves counter-clockwise,
which causes the slag particles deposited on
the grate to penetrate into the cell via opening
(O). The size of this opening depends on the
size of the apparatus, and may reach 30 or 40
cm (1 ft) for a large apparatus in order to avoid
blocking by slag.
For a better understanding of the manner
in which the apparatus shown in Figure 3
operates, I should like to ask you to refer to
Figure 4 which shows an isometric projection
of the fluidization cell at a larger scale. This
cell is provided with vertical side-walls (a-b-c-
d) topped by an upper wall forming a sealed
cover.
Inside the cell, the grate supports the
auxiliary fluidized bed which communicates
with the principal fluidized bed via opening
(O) situated at the base of wall (a), as
mentioned previously.
The shape of the various vertical cell walls
must be adapted to their specific functions, as
follows:
1. Wall (a) is open at the base at (O) to
establish the communication between the
two fluidized beds and to provide a passage
for the slag.
2. Wall (b) is identical to the reactor cylinder
wall; it contains an opening (g) giving
access to an ash-pit placed outside the
reactor (not shown).
3. Wall (c) is preferably laid out in the shape
of a logarithmic spiral, with its convex face
turned toward the outside of the cell to
force slag particles brought in by the
revolving motion of the grate toward
opening (g). This arrangement profits from
the very special fact that this curve has, at
all points, a constant angle in relation to
the incident direction of slag brought in by
the grate (for this purpose, an angle of 30 °
is very favorable). Another essential func-
tion of wall (c) is to prevent the principal
fluidized bed from entering the cell on this
side and flooding it.
4. Wall (d) serves as a simple connecting ele-
ment between walls (a) and (c).
This is only a schematic outline of the cell
for, in fact, the cover of the cell will be prefer-
ably in the shape of a peaked roof, following
the angle of repose or natural slope of products
being processed.
The second example which I should like to
illustrate is shown in Figure 5. Here you see a
cylindro-conical fluidization reactor (a)
containing in the center an ash pit surrounded
by a fixed annular horizontal grate (b).
The fluidization cell (c) revolves around its
vertical axis (x'-y') and has an evacuation
toward the inside ash pit situated at (d), which
is also rotating.
Reacting gas is injected into the reactor at
(e); gas resulting from the reaction is
evacuated by pipe (f) after having passed
through a group of two cyclones (g) mounted
in series; the latter collect and reinject flue
dust into the fluidized bed.
The reactor is supplied with granulated
solid material in this example through pipe (h)
so that solid material circulates by counter-
current flow with relation to gases in interest
of heat recovery.
In this second type of installation the grate
is fixed; the relative motion between grate and
cell results from the cell's own clockwise
motion to assure the penetration of slag into
the cell by opening (O').
The arrangement of the fluidization cell (c)
is based on the same principle as in the
1-5-3
-------
preceding example (Figure 4). However, wall
(c) in Figure 6 is curved in the shape of a
logarithmic spiral with the concave face
turned toward the inside in order to force slag
particles toward the inside ash pit.
You will note that wall (a') is wide open at
the base in (O') to permit inserting the
fluidized bed and the slag particles. Vertical
wall (V) is circular. Vertical wall (c') is laid out
in the shape of a logarithmic spiral with
concave part turned inwards, as already
mentioned. As with the preceding cell, it is
closed by a sealed cover.
From the point of view of the mechanical
fabrication of the various parts, I feel that it is
unnecessary to enter into details concerning
them, except to mention that the cell must be
cooled by circulating water or by vaporizing
water.
Circular fluidizing grates are of the
ordinary type and may consist of refractory
cast-iron links supported by an appropriate
frame.
Startup is realized by a gas or fuel oil
burner (not shown) placed in the upper part of
the reactor, fired on top of the bed and is then
static. Fluidizing gas (in fact, air) must be
injected only when the temperature of the bed
has reached a sufficiently high level.
For example, if the fluidized bed consists of
coal, processing capacity may be estimated at
about 2 to 3 tons/hr of coal gasified/sq meter
of fluidizing grate.
The resulting processing capacity could be
for an 8-meter diameter reactor, for example
150 to 200 tons of coal gasified/hr at
atmospheric pressure. This capacity could be
considerably increased to reach an equivalent
power production of 1000 MW if pressure
gasification is used.
Naturally, reactors designed according to
this new technique may be perfected in various
manners, particularly in view of obtaining
high temperatures which are necessary for
producing cement, by heating the fluid bed
with coal fines or fuel oil and recovering heat
carried off by gases with care. For this
purpose, one might use a series of
superimposed cyclones in which processed
solid material circulates against the current of
reaction gases (Figure 5).
In addition, residual heat of reaction gases
may be recovered by diverting these gases to a
heat exchanger.
Finally, for gasification or for other
chemical reactions at very high temperatures
(500 to 1000 °C), reacting gases may be
injected by nozzles (Figure 5, "j") into the
truncated part of the reactor. This injection, if
sufficiently great, may give rise to a dilute
fluidized bed, which obviously must be
stabilized at the top by cyclones adapted for
this purpose. This latter use of dilute fluidized
bed permits, as it is well known, a
considerable increase in unit production.
From all that has been stated, I feel that
one may conclude that the process offers the
advantage of an exceptionally simple achieve-
ment with a consequent low cost, given its
production capacity.
This process has undergone successful cold
experiments on different types of 1:20 scale
models. The development is so recent that no
industrial nor semi-industrial testing has been
carried out but a pilot unit is now in the course
of realisation for the gasification of 250 to 750
kWhr of coal. So far, no major difficulties
have appeared.
1-5-4
-------
PRINCIPAL FLUIDIZED-BED
(C)
FLUIDIZATION CELL
'" uiiiiBiiiiiiiiiiiiiiiiiii|iiiiiiin iiiiiniiii.il i nun linn in niiiuiniiinniiiiiiiiiiiii i
(0)
OPENING
(B)
AUXILIARY
FLUIDIZED-BED
Figure 1. Fluidized-beds of different depths.
1-5-5
-------
. .
PRINCIPAL FLUIDIZED-BED •.
"
*•/
(.0) \
.OPENING (B)
AUXILIARY
FLUIDIZED-BED
Figure 2. Fluid ized-beds of different depths with air intake.
(D)
AIR INTAKE
1-5-6
-------
(D) VERTICAL
' SIDE-WALL
VERTICAL SIDE-WALL
(0) OPENING
(A) VERTICAL SIDE-WALL
(G) OPENING
(B) VERTICAL SIDE
(F)
DIRECTION
OF FLOW
(E) HORIZONTAL
GRATE
Figure 4. Detail of isometric projection of the fluidization cell.
1-5-7
-------
(F) EVACUATION OF
REACTED GAS
G) CYCLONE
(A) HORIZONTAL
GRATE
(H) SUPPLY OF
GRANULATED
SOLID MATERIAL
(B) VERTICAL AXIS
(C) FLUIDIZATION
CELL
(D) OUTLET
(E) INJECTION OF
REACTING AND
FLUIDIZING GAS
-(0) OPENING
Figures. Cylindrical reactor.
1-5-8
-------
(0) OPENING
(A) SIDE WALL
(D) SIDEWALL
(G) OPENING
(C) VERTICAL
WALL
(F) DIRECTION
OF FLOW (B) CIRCULAR
VERTICAL
WALL
E) HORIZONTAL
GRATE
Figure 6. Detail of fluidization cell.
1-5-9
-------
(G) CYCLONE
(J) INJECTION OF
REACTING GASES
(F) EVACUTION OF REACTED GAS
I) GRANULATED
SOLID
MATERIAL
(G) CYCLONE
(A) FLUIDIZATION REACTOR
(B) HORIZONTAL GRATE
-(D) ASH PIT
"(C) FLUIDIZATION CELL
) REACTOR
(I)
(0) OPENING
Figures. Cylindrical reactor.
1-5-10
-------
6. KINETIC STUDIES RELATED TO THE USE OF
LIMESTONE AND DOLOMITE
AS SULFUR REMOVAL AGENTS IN FUEL PROCESSING
E. P. O'NEILL, D. L. KEAIRNS, AND W. F. KITTLE
Westinghouse Research Laboratories
ABSTRACT
A pressurized thermogravimetric analysis system adapted to handle corrosive gases has been
used to obtain sulfur removal and regeneration data for limestone and dolomite. The kinetic feasi-
bility of the desulfurization processes proposed for fluidized-bed gasification and combustion has
been demonstrated at ten atmospheres. Calcium carbonate can be regenerated for gasification
processes. Regeneration for the combustion system requires further study.
INTRODUCTION
Power generation through fluidized-bed
fossil fuel processing, which uses limestone or
dolomite in the bed to capture sulfur, has the
potential to meet SOa, NOx and particulate
pollution abatement goals at reduced energy
cost. The application of fluidized beds to the
gasification and combustion of oil and coal at
elevated pressures, with combined cycle power
generation, and to oil gasification at atmos-
pheric pressure for retrofit on conventional
plants, is being developed.1 The efficacy of
sulfur removal is a major criterion in eval-
uating these concepts.
Four aspects of the chemistry of sulfur
removal will directly affect the usefulness of
limestones or dolomites as traps for sulfur in
fluidized-bed gasification and combustion of
fossil fuels. The rate of sulfur removal from
the bed gases during gasification or combus-
tion and the capacity of the stones are primary
concerns for ' all the proposed systems;
regeneration of spent stone in an active form
with concomitant sulfur recovery and disposal
of waste stones which contain sulfur are of
complementary importance.
The cycle of reactions in Figure 1 encom-
passes the reactions of concern. Table 1 lists
the reactions and the conditions for the
process options which are being assessed by
Westinghouse.1 Thermodynamic feasibility is
necessary but insufficient for success of the
proposed processes; a knowledge of the ki-
netics of the essential reactions under the pro-
cess conditions can only be obtained by exper-
iment. Despite atmospheric pressure data
obtained by Pell et al.4 on hydrogen sulfide
reactions with dolomite, and various studies
on sulfur dioxide. "sorption" by limestones
and dolomites6'7'8 there is a need for primary
kinetic data with which the behavior of the
fluidized-bed desulfurization of fuels can be
predicted and explained.
The objectives of the current program are
to:
1. Establish which reactions occur,
1-6,1
-------
2. Determine the reaction kinetics, 7. Recommend optimal operating condi-
tions.
3. Determine stone utilization achievable, In this paper we describe how a thermo-
gravimetric analysis (TGA) system, designed
4. Study the effect of regeneration on stone to obtain the desired data is being used to
reactivity, survey the chemistry of the processes, and how
.. „ , . , . it is focusing attention on critical areas where
5. Probe reaction mechanisms, ,, , f.. ,- r xl
successful use of the reactions requires further
6. Assess the influence of side-reactions, study.
1-6-2
-------
Table 1. REACTIONS AND CONDITIONS OF WESTINGHOUSE PROCESS OPTIONS
Reaction
Sulfur removal
1 tiiy., 4- QD i 1/? n *r*^fri _L
'• rorn - l'^U2-»-LabO4 +
CaLU3 ^^2
9 CaO + j_| 5_»Qa2 + ^ Q + —
papr\ ^ ^ r*O
OdLrvy-i L* vj -i
J z
Stone regeneration
•3 CaSOi + . ? ^-CaQ + ^O i ?V,
co 2 CO 2
4. CaS04 + 4H2 -*CaS + 4H20
4 CO 4C02
(followed by reaction 5, 6, 7 or 8)
5. CaS + H20 + CO2-»CaC03 + H2S
6. CaS + H20 + C02-*CaC03 + H2S
7. CaS + 3C02-»CaO + SCO + S02
8. CaS + 3/2 02-*CaO + S02
9. CaS04 + C + H20-*CaC03 + C02 + H2S
10. CaS + 2 O2-*CaS04
Operating conditions
1300°F2000°F
T<1600°F
T<1500°F
P>5atm
T~200°F
atm
T>2000°F
atm
T>1800°F
atm
Applicable
fuel processing
option
Combustion
Gasification
Combustion
(low S02 concentration
at elevated pressure)
Combustion
(at elevated pressure)
Combustion/gasification
(at elevated pressure)
Combustion/gasification
(not recommended due to low
temperature & water purification)
Combustion/gasification
(not recommended due to low SO
concentration, ~2%)
Combustion/gasification
(low S02 concentration
at pressure)
Combustion
(advanced concept)
All gasification
-------
EQUIPMENT
The design conditions for both gasification
and combustion (i.e., pressures of 10 to 30
atm, temperatures up to 1200°C, and the
corrosive gas compositions of hydrogen sulfide
and sulfur dioxide) define an area in which
little or no kinetic investigations of gas-solid
reactions have been reported. For this study, a
duPont thermogravimetric balance was
mounted inside a pressure shell, so that it
could record continuously the weight changes
of a solid suspended in a reacting gas stream
of pre-selected composition at temperatures
up to 1200°C. Figure 2 shows the apparatus;
Figure 3 shows a closeup view of the balance;
and Figure 4 is a diagram of the system. The
corrosion prevention system is based on that
described by Ruth.5 Despite this precaution,
the balance has a limited lifetime owing to
corrosion; alternative designs are currently
under investigation.
The advantages of a TGA are the ability to
isolate chemical reactions for study, the
relative ease with which the desired conditions
can be attained and controlled, and the
accuracy and rapidity with which reactions
can be studied, from the point of view of the
solid.
The chief disadvantage for our purposes
lies in the fact that it is not a fluidized bed;
translation of TGA results into likely fluid-bed
behavior is difficult. A second disadvantage is
the small size of the stone samples used (~ 10
mg)* which makes it occasionally difficult to
use chemical analysis to assess the importance
of minor competing reactions; there is little
product for such physical characterizations as
BET analysis.
Operation of the pressurized TGA was
tested by studying limestone calcination.9
When the operating conditions were suitably
modified, kinetic results in reasonable
agreement with studies at atmospheric
pressure were obtained.
MATERIALS
Solids used in the experiments were lime-
stone 1359, Tymochtee dolomite, dolomite
1337, sieve fractions (420 to 590 Mm) except
where stated. The gases, N2, Oi, CO, CO2,
H2S, and SO 2 were taken from commercial
cylinders. Steam was generated in the mani-
fold.
RESULTS
The strategy adopted was to examine each
reaction in turn before proceeding to study
cyclic behavior.
A preliminary survey of the principal reac-
tions of Table 1 has been carried out; Table 2
summarizes the results.
Sulfur Removal
Combustion:
CaO + S02 + l/2 O2 -* CaSO4 (1)
The reaction between sulfur dioxide and
calcined dolomite 1337 in excess oxygen
followed closely the results obtained by Borg-
wardt7 in a differential bed reactor; e.g., rate
at 10 percent Ca utilization:
TGA results (850°C) 2.40 x 10 "5 [mole SO 3
•gram calcined dolomites-sec-1 ]
Borgwardt's data 2.10 x 10'? [mole SO 3
•gram calcined dolomites-sec"1 ].
The rate of reaction became increasingly
smaller at 40 percent Ca utilization at
pressures of 1 atm, but at 10 atm rapid
reaction continued beyond 70 percent
reaction, as shown in Figure 5. The initial rate
in the pressurized case is probably limited by
the supply of sulfur dioxide to the solid in our
particular reaction system.
*10 mg is kinetically optimal for particle sizes being studied;
balance can accept up to 1 gram samples.
Gasification:
CaO »
(2)
CaCO
H2S
CaS + H20 +(C02)
1-6-4
-------
Table 2. THERMOCRAVIMETRIC ANALYSIS DATA SUMMARY
Reaction
(1)
CaC03 + S02
+ 1/2 02
-> CaSOi, + CO2
CaO + S02
+ 1/2 02
-» CaS04
(2)
CaC03 + H2S
-* CaS + H20 + CO2
(4)
CaSOft + 4 CO
-> CaS + 4 CO2
(5)
CaS + H2O * CO2
-» CaC03 + H2S
(10)
CaS + 2 02
— CaSO,,
Original
substrate
500-^m dia
Tymochtee
Dolomite
Dolomite
1337
Tymochtee
Dolomite
Dolomite
1337
Tymochtee
Limestone
1359
Tymochtee
Dolomite,
Dolomite
1337
Tymochtee
Dolomite,
Dolomite
1337
Limestone
1359,
Tymochtee
Dolomite
(4000 ^m -
1000 ^m)
Pressure,
atm
1
10
1
10
1
10
1
10
1
10
1
1
Cas . a
composition
0.25% SO2
4%02
0.25% SO2
4% 02
15% H2S
1.5% H2S '" N2
0. 5% H?S in No
C02
25% CO
CO/'CO2 = 2/1
CO2, H2O + 25%
Air
Air
Temp.,
°C
800-
850
800-
850
600-
845
760
750,
820
850
590-
700
400-
950
400-
800
Results
40% sulfation at 1 atm.
90% sulfation at 10 atm
Comparable to CoCO3
results
• 90% sulfidation
Some suppression at
high H2S concen-
trations
normal calcination
-90% yield
Depends on source of
CaS
7% yield
Up to 90% yield
Conclusions
Makes once-through
systems competitive
Attractive utilization
and rate achievable
under CO.2 pressures
equilibrium
Gives poor substrate
for RX5
For gasification
promising; for
combuslion.
unproven
Appears impractical
as impervious
sulfate shell forms.
Suitable for disposal
al 800 °C
Future
work
Higher temperature
composition
Investigate utilira-
lion .n low SO>
concentration
High pressure.
fuel gas effect
Investigate SO>
loss and effect
of initial sulfate
porosity
Seek higher
conversion
Chock loss
of SO 2
T*
VI
The balance is nitrogen.
-------
Sulfidation of dolomite at atmospheric
pressure gave results in broad agreement with
Pell's work10 (Figure 6). Rapid rates and high
yields are attained. Some suppression of
reaction at high hydrogen sulfide concentra-
tions and lower temperatures (700°C) was
observed here, confirming Pell's finding.
In limestone sulfidation at pressure (10
atm) complete reaction was achievable when
the stone was calcined; a shrinking shell
model empirically describes the kinetics.
Sectioning of partially reacted stone reveals an
outer layer of calcium sulfide and an
unreacted core of lime. No sulfidation was
apparent before the normal calcination
temperature.
REGENERATION
The two-stage regeneration of calcium
carbonate from calcium sulfate has been
studied. Reduction of calcium sulfate
according to reaction
CaSO4 + 4CO -* CaS + 4CO2 (4)
was essentially complete at temperatures
between 750 and 850°C, and 10 atm total
pressure (Figure 7).
Regeneration of carbonate from the sulfide
produced in reaction (3) above proved to be
difficult. AfteF 25 percent regeneration, the
rate became extremely slow (Figure 8). By
contrast, yields greater than 70 percent were
obtained when the dolomite was directly
sulfided with hydrogen sulfide (Figure 9). The
latter regenerated carbonate was calcined, and
its reactivity with sulfur dioxide was tested. It
proved to react as rapidly as freshly calcined
dolomite.
Unregenerated sulfide is concentrated at
the core of the particle, as evidenced by
sectioning the reacted stone.
STONE DISPOSAL
Spent limestone or dolomite from gasifica-
tion processes contain calcium sulfide. This
compound liberates hydrogen sulfide on
contact with carbonated water, preventing its
disposal in an untreated form. The problem is
important for once-through and regenerative
processes.
Conversion of the calcium sulfide to
calcium sulfate before disposal, has been
proposed.
CaS + 2O2 -*CaSO4 (10)
This reaction has been studied at atmos-
pheric pressure using sulfided limestone,
sulfided dolomites, and air as reactants.
Sulfided limestone 1359 cannot be oxidized
to sulfate, probably due to formation of the
same impervious layer observed when the
stone is directly sulfated.11 Surface reaction is
observed but it rapidly decays. The reactivity
can be renewed by lightly crushing the stone
and repeating the reaction, which proceeds to
an additional degree of oxidation about equal
to the first stage. At high temperatures
(900 °C), about 1 percent reaction occurs
extremely rapidly, followed by complete
cessation of reaction. The exothermic nature
of the reaction may have formed an
impervious dead-burned lime layer on the
solid surface.
Sulfided dolomite may be oxidized to
calcium sulfate. At low temperatures (550 °C)
an initially rapid rate of reaction falls off after
14 percent of the sulfur has been oxidized
within eight minutes. By contrast, 92 percent
of the sulfide is oxidized within three minutes
at a nominal temperature of 800°C. The stone
temperature may be higher (Figure 10).
MODELS
The development of kinetic models which
encompass the data, with a view to making
predictions of the course of reaction in
fluidized beds, is one of the goals of our
investigation. Empirically, limestone sulfida-
tion fits a contracting sphere model, in
agreement with the physical form of sulfided
stones. Reduction of sulfated dolomite by
carbon monoxide is apparently first-order in
sulfate. However, sufficiently detailed
1-6-6
-------
variation of the controlling parameters has not
yet been sufficiently studied to permit predic-
tions of reaction behavior over the wide range
of parameters applicable to the proposed
processes.
CONCLUSIONS
The TGA has provided information on the
rates of reaction and the degree of utilization
which can be achieved by the proposed sulfur
removal, stone regeneration, and solid
disposal processes.
Removal of sulfur dioxide and hydrogen
sulfide: high stone utilization (>70
percent) has been achieved within 30
minutes.
Regeneration of calcium sulfate by the
two-step, low temperature process:
calcium sulfate can be reduced to cal-
cium sulfide ( > 90 percent). Calcium
carbonate has not been successfully
regenerated from the sulfide — <30
percent regeneration at practical tem-
peratures.
Regeneration of calcium sulfide: calcium
carbonate can be regenerated from
calcium sulfide produced from H2S—
70 percent regeneration in 15 minutes.
minutes.
Stone disposal: calcium sulfide has been
oxided (90 percent) to calcium sulfate
using dolomite.
These results are for a limited range of
operating conditions. Further work is required
to assess these reactions over the full range of
operating conditions projected for the
processes.
Future work is planned to:
1. Assess the low-temperature CaSO4
regeneration process.
2. Study the one-step, high-temperature
CaSO4 regeneration process and other
alternative processes.
3. Assess the activity of regenerated stone
as a function of the number of sulfur
removal/regeneration cycles.
4. Further study sulfation of stones for
disposal.
5. Translate the kinetic data to fluidized-
bed systems with the aid of experi-
ments on a 2-in. hot-model fluid bed
and pilot plant experiments conducted
by Westinghouse and others.
6. Understand reaction mechanisms.
ACKNOWLEDGEMENTS
We thank Dr. D. H. Archer for guidance
and support. We also acknowledge the
technical assistance of Drs. F. P. Byrne and
C. R. Wolfe of the Westinghouse Analytical
Chemistry Department.
REFERENCES
1. Archer, D.H., et al. Evaluation of the
Fluidized Bed Combustion Process.
Summary Report. Westinghouse Research
Laboratories, Pittsburgh, Pa. Prepared for
the Environmental Protection Agency,
Research Triangle Park, N.C. under Con-
tract Number CPA 70-9. November 1971.
2. Keairns, D.L. Fluidized Bed Gasification
and Combustion for Power Generation.
In: Proceedings Frontiers of Power Tech-
nology, Oklahoma State University, Still-
water, Oklahoma, October 1972.
3. Archer, D.H. Fuel Processing Tailored to
Environmental Needs. (Presented at
American Chemical Society National
Meeting. September 1972.)
4. Pell, M, R.A. Graff, and A.M. Squires.
(Presented at meeting of American Insti-
tute of Chemical Engineers. Chicago.
December 1970.)
5. Ruth, L.A., A.M. Squires, and R.A. Graff.
(Presented at American Chemical Society
Meeting. Los Angeles. March 1971.)
1-6-7
-------
6. Coutant, R.W., J.S. McNulty, R.E.
Barrett, G.G. Carson, R. Fischer, and
E.H. Lougher. Summary Report. Pre-
pared For National Air Pollution Control
Administration, Cincinnati, Ohio, under
contract Number PH-86-67-115, 1968.
7. Borgwardt, R.H. Environ. Sci. Technol.
4(1): 59, 1970.
8. Borgwardt, R.H. and R.S. Harvey.
Environ. Sci. Technol. 6(4): 350, 1972.
9. O'Neill, E.P., W.F. Kittle, C.R. Wolfe,
and L.M. Handman. Unpublished results.
10. Pell, M. Ph.D. Thesis, City University of
New York. 1970.
11. Davidson, D.C. and J. Highley. The
National Coal Board, London, England,
Final Report to the Environmental Pro-
tection Agency, Research Triangle Park,
N.C. Appendix ;8, September 1971.
H2S
S0202
• SULFUR REMOVAL
STONE REGENERATION
Figure 1. The calcium carbonate/sulfur cycle basic reactions.
1-6-8
-------
OS
GO
Figure 2. Thermogravirnetric analyzer for high temperature and pressure reaction studies
on limestone and char.
-------
Figure 3. The duPont 950 thermogravimetric balance.
1-6-10
-------
*
COOLING
COILS
SAMPLER-
BALANCE
MECHANISM
FURNANCE
V1\1\\\V\VV
=! SAMPLE
iiiinm
COOLING COILS
PRESSURE
" GAUGE
BY-PASS
PREHEATERS
10 ATMOSPHERES
REACTION
GAS
MANIFOLD
I I I I
NITROGEN
' PURGE
RECORDER
CONSOLE
Figure 4. Diagram of the TG system.
1-6-11
-------
850°C, 5000 pptn SO?
4«02
10 rug DOLOMITE
10 12 14 16 18 20
Figure 5. SC>2 reaction with calcined dolomite 1337.
1-6-12
-------
ATMOSPHERIC PRESSURE
-30+40 TYMOCHTEE
DOLOMITE
15% H2.S IN N2
845°C
EXPERIMENTAL
PELL'S FIRST
ORDER MODEL
10 20 30 40 50
TIME, sec
Figure 6. Sulfidation of calcined dolomite.
ea
UJ
UJ
o±
LU
Q_
90
80
70
60
50
40
30
20
10
0
OTG 95 38% SULFATED ~
*TG 96 26% SULFATED
CO/C02 = 2/1
CO=20%
P=10atm
T=820°C
DOLOMITE 1337
CaS04+4CO-*CaS44C02 —
I I I I I I I I I
0 10 20 30 40 50 60 70 80 90
TIME, min
Figure 7. Reduction of sulfated dolomite.
Ld
(£
LU
30
25
20
15
10
5
0
10 20 30 40 50 60 70
TIME, minutes
Figure 8. Regeneration of dolomite.
DOLOMITE 1337
(H20)«10%
=10%, 23%
= 650°C
10atm
TG94
CaS MgO + C02 + H20 — CaC03 MgO + H2S
80
90
100
1-6-13
-------
LU
0.8
o 0.7
QQ
ce
o
0.6
3 0.5
§ 0-4
o
LU
S 0.3
~ 0.2
o
i 0.1
o
^ 0.0;
DOLOMITE 1337
-30+40
86% Ca2+SULFIDED
700° C, 10 ATM' —
[H20]=[C02]=20%
"CaS- MgOf H20+C02 = CaC03- MgO+H2S"
04 8 12 16 20 24 28
TIME, n\\n
Figure 9. Regeneration of carbonate from
sulfided dolomite.
O TG 121 DOLOMITE 1337,55% Ca SULFIDED
A TG21 LIMESTONE 1359, 99% Ca SULFIDED
ATMOSPHERIC PRESSURE
AIR
70Q°C
CaS+202—CaSOa
100
TIME, sec
Figure 10. Oxidation of sulfided limestone and dolomite.
1-6-14
-------
SESSION II:
Non-Coal Fluidized-Bed Combustion Processes
SESSION CHAIRMAN
Mr. A. Skopp, Esso Research and Engineering, Linden, New Jersey
II-O-l
-------
1. COMBUSTION m A
CIRCULATING FLUID BED
H. W. SCHMIDT
Lurgi Chemie und Huettentechnik GMBH
INTRODUCTION
To reduce the high heat consumption as
compared with the rotary kilns normally still
used for calcination of alumina, Vereinigte
Aluminium Verke AG (VAW) has developed a
fluid-bed process with direct combustion in
cooperation with the Lurgi Companies.
The main characteristic of this process,
which is generally suited to the application of
endothermic processes at low particle diam-
eter, is the circulating fluid bed. Compared
with classical fluid beds with constant bed
height and a defined surface, a gas/solid mix-
ture of varying concentration in a circulating
fluid beds fills up the complete reactor room.
The solid discharged at the furnace top to-
gether with the combustion gas is fed back to
the reactor after being collected from the gas
flow in a recycling cyclone. Only by permanent
recycling is the system maintained in a con-
stant condition.
This type of reactor has the advantage of
passing large gas volumes through relatively
small reactor sections.
Since the combustion gas and the fluidiza-
tion gas are identical, the problem is to opti-
mally coordinate the fluid dynamic conditions
for the fluid bed, such as solid recycling and
distribution of concentration, with the pyro-
technical requirements, such as mixing and
combustion.
II-l-l
-------
CIRCULATING FLUID-BED PILOT
PLANT
To investigate these processes, a pilot plant
(24 ton/day AhO3) for fundamental studies
was erected prior to the construction of the
first industrial plant (500 ton/day A12O3).
Figure 1 shows a process flow sheet with the
most important process steps.1
The essential part of the process consists of
the calcination circuit which works on the
principle of a circulating fluid bed. Optimum
thermal efficiency of the whole process is
reached by preheating the solid with the com-
bustion gas flow, and the combustion air with
the solid discharge flow.
The feed hydrate is preheated by the waste
gas flow and partially dehydrated in two
Venturi dryer stages. The alumina, which is
preheated to about 400°C and has a loss on
ignition of 7-8 percent, is subjected to final
calcination in the fluid-bed furnace.
The energy required for the endothermic
process, water evaporation, dehydration heat,
radiation and discharge losses, is supplied by
direct combustion of heavy fuel oil in the fluid-
bed furnace.
The combustion air is divided into primary
air and secondary air. The two air streams are
preheated in a multi-stage fluid-bed cooler by
the discharged alumina. The secondary air
flow, which fluidizes the alumina in the cooler
chambers, is preheated directly. The primary
air flow, which is passed through tube bundles
immersed in the fluid-bed chambers, is pre-
heated indirectly.
Figure 2 shows a sketch of the material
balance of a circulating fluid bed. The
alumina m A discharged at the furnace top
together with the combustion gas is composed
of the circulating flow niR and the throughput
flow mD. The ratio 9!" circulating flow to
throughput flow m R: m D is decisive for the
mean retention time of the product. The mean
retention times of the alumina may be
adjusted between about 10 and 60 minutes
n-i-2
depending on the intensity of circulations of
the circulating solid.
This adjustment of the alumina retention
time needs a defined distribution of solid
concentration in the axial direction of the
fluid-bed furnace, corresponding to the
division of primary and secondary air streams.
Because of the high gas velocities compared
with conventional fluid beds, the problem is to
burn the fuel completely by the time it reaches
the furnace top. This must be achieved by
optimum mixture of fuel and air. The studies
mentioned in this paper therefore concentrate
on the combustion efficiency at varying
conditions for the mixing ratios of fuel and air.
The phase condition of the circulating
fluid bed is illustrated in the fluid-bed phase
diagram (Figure 3) as a function of the Fr^
number and the Re^ number. The perimeter
is determined by parameters K and M. K and
M are nondimensional parameters which
result from a combination of the Fr^ number
and the Rek number which derive from the
fluid-bed phase diagram.
The range shown in the diagram covers the
adjusted test conditions by means of the K and
M lines. This range is located in the zone of
the aggregative fluidization, a transition phase
between classical fluid bed characterized by
the particulate fluidization of the individual
particle, and of pneumatic transport which
starts with the limiting characteristic 3/4-Fri?
7G/(/K •y&)= 1- The cross-hatched zone
between lines K = 3.52 and K = 2.1
characterizes at a mean particle diameter dkm
= 45 fim of the alumina used, the phase
condition of the lower furnace section whose
concentration is determined by the primary air
flow. Lines K = 626 and K = 0.264 result
from the minimum and maximum particle
diameters, dkmin = 8 ion and d|jmax = 120
urn. Lines M = 0.035 and M = 1.33 are
determined by the lowest and highest
fluidization velocities.
It can be inferred from this illustration in
the fluid-bed phase diagram that the fine
particles are preferably discharged from the
-------
lower furnace section and subjected to a
higher amount of circulation than the larger
particles, since the K lines of the smaller
particle diameters remain above the boundary
line for pneumatic transportation. However, it
can be demonstrated by tests that no
separation of the fine particles occurs during
circulation.3 This means that under
appropriate flow conditions all particles come
into the range of pneumatic transport. By
impact and frictional forces of the particles
and by separation of particle clusters,
conditions of fluidization may be changed.
Thus, the transport condition is delayed at
various times and in various places, resulting
in inconstant distribution of solid concentra-
tion in the axial direction.
MEASURING EQUIPMENT AND TEST
PERFORMANCE
The following parameters were varied for
studying the combustion process.
1. Primary air velocity, Wp
This is the velocity referring to the free
cross-section in the lower range of the
fluid-bed furnace below the secondary air
inlet.
2. Impulse of the secondary air flow, Is
This is the force by which the secondary air
jets penetrate horizontally into the fluid
bed produced by the primary air.
3. Mean furnace temperature, Tm
4. Air ratio of the combustion
For the measurements, exhaust openings
were arranged in six measuring sections at
various levels of the fluid-bed furnace. It was
possible to take gaseous and solid samples by
two measuring directions arranged perpen-
dicular to each other, with water-cooled bleed-
off lances. Figure 4 shows the measuring
arrangement for the tests. The alumina
sucked off together with the gas is separated in
a cyclone. After cleaning and drying, the gas is
passed to continually operated gas analyzers
for determining the concentration of the
components CO, CO2, and O2.
The central part of the lance is composed of
a thermocouple for ascertaining the
temperature prevailing at the corresponding
measuring point. By control of the suction
pump the exhausted gas rate can be adjusted
in accordance with the temperature at the
measuring point in the lance section so that
isokinetic suction conditions exist.
With this measuring equipment, the
following measuring variables can be
determined in a radial direction of each
measuring section:
1. Gas concentrations of CO, CO2, and O2.
2. Temperature, Tm.
3. Concentration of solid, CM-
INFLUENCE OF SECONDARY IMPULSE
ON THE EFFICIENCY OF COMBUSTION
From the radial profiles measured, mean
values of the measuring section areas are
formed by integration over the furnace cross-
section. From the integral mean values of the
various section areas, the axial distributions of
gas concentrations, solid concentrations, and
temperatures can be determined over the
height of the fluid-bed furnace.
In Figure 5 the axial temperature distri-
butions are stated for various mean furnace
temperatures (Tm). For the greater part of the
fluid-bed furnace, the lines show a constancy
of temperature which does not occur in
conventional fluid beds and can only be
explained by the intensive solid circulation.
The decrease in temperature in the lower part
of the furnace is caused by the solid feed and
the remaining dehydration heat of the
calcination process still to be applied in the
range. Tests at a higher ratio of circulating
solid to throughput solid mR:mD show that
the temperature drop in the lower furnace
section can hardly be observed any longer.
Distributions of solid concentration in axial
direction at varying velocities referred to in the
free cross-section are represented in Figure 6.
In the lower furnace section, the curves show
increasing solid concentration CM (x) at
decreasing primary air velocities (w«). Starting
II-1-3
-------
with coordinate X/D = 1.4, the solid concen-
trations of all curves show a significant de-
crease with volume flow increase by the
secondary air flow (Vs). Downstream of co-
ordinate X/D = 3.9 the axial distributions of
the solid are almost constant up to the furnace
top.
To analyze the efficiency of combustion, it
is necessary to determine the combustion rate
of the fuel components.4 Except for the local
gas analysis values CO, CO2, and O2, it was
possible to rely upon the unburnt carbon
percentage which was analyzed from the
sucked off alumina. These deposits of unburnt
carbon on the alumina are mainly found in the
lower measuring section areas, decreasing in
accordance with the course of combustion up
to the furnace top.3
The active y-alumina coming from the
preheating section at a temperature of about
400° C is also fed to the zone of the fuel which
is directly injected into the fluid bed. Due to
the catalytic activity of this y-alumina, the
cracking process initiating the combustion of
the heavy fuel oil is greatly enhanced. The
catalytic effect on the cracking process must
be exclusively caused by the activity of the y-
alumina since the remaining alumina, which
has a temperature of 1100°C, is no longer
active as a catalyst.
On the basis of the burnt carbon
percentage, a molar balance of the oxygen
required for combustion, and the measured
gas analysis values for CO, CO2, and O2, it is
possible to establish the necessary equations to
determine the development of combustion.
From the radially measured values, such as
shown for characteristic tests in Figure 7, the
. distribution of combustion rates in axial
direction of the fluid-bed furnace are obtained
by radial and axial integration. Figures 8 and
9 show the combustion curves from which the
main factors of influence on the degree of
combustion can be seen.
Figure 8 shows efficiencies of combustion
with constant impulse of the secondary air jets
and varying primary air velocities (wp), which
are proportional to the velocity referring to the
free cross-section (WG). At decreasing values
of the primary air velocity (wp), a steeper
course of the combustion curves can be
observed. This result is attributed to the
retention time of the combustion gases in the
furnace, which increases at decreasing
velocity.
Figure 9 shows percentage of combustion
at a constant primary air velocity (wp), but
varying impulses (Is) of the secondary air jet
penetrating into the fluid bed. Although the
high primary air velocity (wp) shortens the
retention time of the combustion gases,
substantially steeper combustion degrees can
be reached at increasing secondary air impulse
(Is) than in the case of curves as per Figure 8.
This result is due to the more intensive radial
mixing of the fuel components with the
combustion air. The radial distributions in
Figure 7 in the two different measuring section
areas clearly show this influence. At a lower
secondary air impulse an increase of the CO 2
concentration exists only in the zone near the
furnace wall. At a higher impulse value a
homogeneous distribution of concentration
over all the measuring section areas exists.5
The results show that by suitable
distribution of primary and secondary air
streams, the combustion reaction can be
substantially speeded up. This will permit an
increase of the gas throughput and a higher
specific load of the fluid-bed furnaces, along
with a reduction of the height of the fluid-bed
furnace.3
In conclusion, it should be mentioned that
apart from the combustion process as studied
in a pilot plant, two industrial plants, each
having a capacity of 500 metric ton/day
AhOs, are successfully operating according to
the described process; another three plants,
each with a capacity of 650 metric ton per day
Al2O3, are under construction and will be
started up in the course of the next year.
H-l-4
-------
SUMMARY
In a pilot plant for the calcination of
alumina according to the circulating fluid-bed
method, the combustion process is studied
experimentally. Since high gas velocities occur
in the circulating fluid bed, the retention times
of the combustion gases are short at direct
combustion in the fluid bed. By a suitable
division of the primary and secondary
combustion air flows, it is possible to increase
the radial mixing and to substantially speed
up the combustion of the fuel components. It
is thus possible to increase the specific
throughput capacity by raising the gas rate.
During the tests, deposits of unburnt
carbon on the alumina were observed in the
lower furnace section. Consequently, the
endothermic process can only be performed
according to the principle of circulating fluid
bed since the carbon is not completely burned
until the fuel reaches the furnace top.
NOMENCLATURE
CM = Solid concentration, kg/Nm3
D = Diameter of reactor, m
dfc = Particle diameter, m
= Froude number (-)
= Acceleration of gravity, m/sec2
gy = Burned-out portion of fuel (-)
gu = Unburned portion of fuel (-)
Is = Impulse of secondary air flow, m kg/sec 2
g
K =
M =
• Vc
K-VG)
,3
mp
m R
Tm =
Solid output on furnace top, kg/hr
Solid throughput, kg/hr
Circulating solid, kg/hr
Reynolds number (-)
Mean temperature, ° C
VP = Primary air flow, NmVhr
YS = Secondary air flow, Nm3/hr
V F = Waste gas flow, Nm3 /hr
WG . = Superficial gas velocity, m/sec
wp = Superficial velocity of primary air,
m/sec
LM = Efficiency of combustion, gv/gu + gv)
yG = Specific gravity of gas, kg/m3
y'K = Specific gravity of solid, kg/m3
v = Kinematic viscosity of gas, m2 /sec
BIBLIOGRAPHtY
1. Reh, L. Fluid Bed Processing. Chem. Eng.
Progr. 67(2). 1971.
2. Ernst, J., L. Reh, K.H. Rosenthal, and
H.W. Schmidt. Experience with the Cal-
cination of Aluminum Trihydrate in a
Circulating Fluid Bed. (Presented at
American. Institute of Mechanical
Engineers Meeting. New York. Paper
number A-71-4. February 1971.)
3. Schmidt, H.W. Uber den Verbrennung-
sverlauf in zirkulierender Wirbelschicht.
Dissertation, Karlsruhe 1971.
4. Gunther, R. Ausbrand von Strahl-
flammen. Archiv f.d. Eisenhiittenwesen.
39(7):515-519, 1968.
5. Martsevoi, E.P. Spread of a Gas Jet in a
Cross-Flowing Stream of Different Density.
Gas-Institute, Acad. of Sciences USSR,
translated from Teoreticheskie Osnovy
Khimicheskoi Technologic. 3(4):644-646,
1969.
6. Reh, L. Das Wirbeln von kornigem Gut im
schlanken Diffusor als Grenzzustand
zwischen Wirbelschicht und pneu-
matischer Forderung. Dissertation,
Karlsruhe 1961.
n-i-s
-------
CYCLONES
©FAN
/!
ELECTROSTATIC
PRECIPITATOR
FLUID-BED
FURNACE
Al (OH)3
WET
FEEDING
SCREW
\^
SECONDARY
AIR BLOWER
FLUIDIZED
BED COOLER
PRIMARY
AIR BLOWER
Figure 1. Flowsheet of fluid-bed calcining plant.
n-i-6
-------
Figure 2. Material balance of circulating fluidized bed.
IM-7
-------
IM-8
IIP
Figure 3. Fluid-bed diagram.
-------
TO RECYCLING CYCLONE
ANALYZER ANALYZER ANALYZER
FILTER
FLOW METER I
ALUMINA - FEEDBACK
SECONDARY AIR
WORKING PLANE II
FEEDBACK
ALUMINA
ALUMINA INLET
FROM
PREHEATING STAGE
WORKING PLANE I
SUCTION PUMP
DRYER
ALUMINA
Figure 4. Measuring arrangement.
II-1-9
-------
1100
1000
o
o
0=
fi! 900
800
700
600
-O- T= HOOt
-O- T = 990 °C
-•- T = 900 'C
-•- T = 810 T
O
HEIGHT,-^
Figure 5. Axial temperature Tm distribution versus nondimensional height X/D
of fluid-bed furnace.
IM-10
-------
200
180
160
140
j 12°
5
•S 100
60
40
20
Ow
A w
0.91 m/sec, WQ 1.83 m/sec
1.19 m/sec, WQ 1.98 m/sec
1.32 m/sec, WG. 2.05 m/sec
1.43 m/sec, WQ 2.36 m/sec
1.64 m/sec, WG 2.53 m/sec
2.03 m/sec, WG 3.05 m/sec
01234567
NONDIMENSIONAL HEIGHT (X/D)
Figure 6. Axial solid concentration (C|y|) distribution versus nondimensional
height (X/p) of fluid bed furnace.
Il-lrll
-------
240 cm ABOVE GRATE
wp = 0.91 m/sec
ls=1.298m-kg/seez
140 cm ABOVE GRATE
WD = 0.91 m/sec n
ls =1.298 m-kg/sec2
o C02
• 02
* CO
• 02
OCO
ls = 0.551 m-kg/sec2
• C02
nco
14
12
10
CM
O
o
O
CM
o
14
12
10
8
o
o
490 350 210 70 0 70 210 350 490
8 6
4
2
0
490 350 210 70 0 70 210 350 490
RADIUS, mm
RADIUS, mm
Figure 7. Radial distribution of gas concentration 140 cm
and 240 cm above grate at constant primary air velocity
and variable secondary impulse (ls).
II-1-12
-------
CO
o
ls=0.551 m-kg/sec2
T =1100°C
Wp=L64m/sec
Wp=2.03 m/sec
3 4
HEIGHT (X/D)
Figure 8. Efficiency of combustion (a/m) at constant impulse (l/s) and varying primary air
velocities (w/p).
II-1-13
-------
3 4 5
NONDIMENSIONAL HEIGHT (X/D)
Figure 9. Efficiency of combustion (am) at constant primary velocity (Wp)
and variable secondary air impulse (ls).
II-1-14
-------
2. DISPOSAL OF SOLID WASTES BY
FLUIDIZED-BED COMBUSTION
G. G. COPELAND
Copeland Systems, Inc.
ABSTRACT
Those of us who labor in the field of pollution control have grown to realize that solid waste
disposal is the largest pollution problem which society must face in the coming years. For too long,
we have left the problem to municipal fathers, who for one reason or another have either ignored
the problem or have installed equipment which has long since been made obsolete by modern
technology.
We hear every day about the virtues of land filling garbage and using municipal sewage sludge
as soil conditioner, "a la night soil" techniques used in the Far East, because incineration is too
expensive, too dirty, makes smoke, and needs high chimneys. This thinking is obsolete in the face
of present day fluid-bed technology which is substantially less costly than conventional
incinerators, is not dirty, and cannot be made to smoke if operated properly. To our knowledge,
there is not a fluid-bed system in operation anywhere in the world which is backed up with a smoke
stack.
This paper covers the development of the largest fluid-bed solid waste incinerator in the world
— an installation which we feel is the forerunner of the next generation of solid waste disposal
systems and the best means of solving the solid waste disposal problem.
INTRODUCTION
Having been one of the midwives at the
birth of fluidized-bed technology in the 1940's,
it is a pleasure for me to be here today at this
Third International Conference on Fluidized-
Bed Combustion, and to be able to discuss the
technology with so many people who are
knowledgeable in its applications.
As a matter of fact, I was one of the first
persons assigned the responsibility of selling
fluid-bed technology outside of the oil refining
industry.
Looking back on those years of trying to
convince the technical world that we weren't
crazy, and at the same time having to show
sales progress while doing research all day, all
night and weekends, it is a wonder that those
of us who were engaged in this exciting
development did not throw up our hands in
disgust and go to something more certain.
In those, days when the new applications
were basically metallurgical, it seemed that all
fluid-bed installations would always be in
out-of-the-way places like Red Lake, Ontario;
Arvida, Quebec; or Berlin, New Hampshire.
For sure, however, all startups over a five-year
period were in the dead of winter, and could
only be described as miserable.
H-2-1
-------
Today, our startups are in places like Flor-
ida, Japan, South Africa, the Bahamas — as
always, in the winter time; but the climate is
certainly much better on the average, and we
have the technical understanding and support
of people who have learned something about
the technique.
The technology has come a long way in the
intervening 25 years, and I believe it is now
coming into its own with new and exciting
applications popping up all over the world.
Having been active in the business throughout
these years, I think my greatest satisfaction
comes from meeting young engineers, fresh
out of school, who have been taught at least
the basic elements of the techniques and can
go on to understand the many permutations
and possibilities of fluid-bed processing.
The company, which bears my name, has
specialized in air and water pollution control
since the early 1960's and uses fluid-bed com-
bustion technology as the heart of many
patented processes which the company sells on
a worldwide basis.
We now have fluid-bed waste disposal
combustion installations in virtually every
major industry, many on a first time basis, and
have successfully used the technique on both
liquid and solid waste materials.
We like to think that in developing certain
fluid-bed processes to dispose of pulp and
paper mill wastes, we did accomplish the
impossible by burning organic matter in a
fluid bed in the presence of inorganic salts
having a low fusion point. We did, in fact,
what most of the early publications on fluid
beds said was impossible.
Our first such commercial installation was
in pulp and paper pollution control, where we
were required to destroy the organic matter
removed from wood pulp in an inorganic solu-
tion of sodium-sulfur salts. We knew that a
fluid-bed temperature in excess of 1300°F was
necessary to get complete burn-out of organic
(polluting) matter; at the same time we knew
that the inorganic salts would fuse at 1350 to
1370°F depending on the inorganic mix.
By taking advantage of the partial fusion of
inorganics, we were able to force pelletization
of the inorganics to form the fluid-bed me-
dium; our systems actually operate easily
between 1300 and 1350°F. We have several
such plants which have been in successful
operation for over 10 years — providing a
solution to pollution problems which were
considered impossible in the 1950's.
For the most part, our early development of
fluid-bed technology for pollution control
dealt with waste solutions. In 1971, however,
we built what we believe is the largest solid
waste fluid-bed incinerator in the world at
Great Lakes Paper Company in Thunder Bay,
Ontario. We believe this installation broke
some new ground in solid waste disposal by
fluid-bed technology. This paper covers that
installation in some detail.
11-2-2
-------
FLUID-BED
GENERAL
INCINERATORS IN
The principle of fluidizing solid materials
at elevated temperatures in the presence of
and by means of a gas, was first commercially
developed by the oil refining industry in the
form of fluid-bed catalytic crackers. While
fluid-bed incinerators only vaguely resemble
"cat crackers" they do function because solid
particles are set in fluid motion (fluid ized) in
an enclosed space (fluid-bed zone) by passing
combustion air through the fluid-bed zone in
such a way as to set all particles in that zone in
a homogeneous boiling motion.
In this state, the particles are separated
from each other by an envelope of the
fluidizing gas (air for combustion) and present
an extended surface for a gas to solid reaction,
as for example, air to carbon-hydrogen. This
extended combustion surface makes possible
the high thermal efficiency found-in most
fluid-bed reactors.
Capacity is a function of reactor bed total
area, but is usually expressed as fluid-bed
surface area.
At combustion equilibrium, the fluid bed
resembles a boiling liquid and, in fact, obeys
most of the hydraulic laws. The dispersion of
fluidizing gas throughout the fluid-bed zone
by the specially designed orifice plate, assures
complete mixing; temperature variations from
any one spot in the fluid bed to any other will
not normally exceed 10 to 15°F.
The mass of fluid-bed medium is kept at
combustion temperature by oxidation of the
organic material in the feed by the oxygen
contained in the fluidizing air. There is little
or no flame, but rather a glowing condition.
Combustion is virtually instantaneous and the
fluid-bed proper will contain no unburnt
organic material. Complete oxidation is the
key to the control of air pollution.
It is fundamental to any incineration pro-
cess that there be no air pollution problem
resulting from incomplete combustion of
waste solids. In this respect, fluid-bed units
are superior to any other type of combustion.
This is again due to the extended surface
presented by the fluid-bed medium. Copeland
designed units normally will burn in excess of
300,000 Btu/ft2-hr. This could be compared,
for example, with coal burning boilers which
do well to consume 40,000 Btu/ft2 of grate
area/hr.
A commercial unit burning sewage sludge
at 1400° F had the stack gas analyses shown in
Table 1.
Table 1. MASS SPECTR06RAPHIC
ANALYSES^
Component
Volume, %b
C02
A
02
N2
Volume, ppmc
S02
COS
H2S
NOX
Hydrocarbon
Sample number
1
5.17
0.73
14.4
79.7
ND
2
8.33
0.07
9.2
81.8
ND
3
8.86
0.60
7.4
83.0
ND
Dry basis.
Samples were taken at 11:30 a.m.,
October 30, 1968.
cNot detectable by mass spectro-
graphic analysis.
Every installation built in the last ten years
is meeting the air pollution regulations of the
state in which it is located and complies with
the most stringent air pollution regulations of
the country. This applies to both gaseous and
particulate matter in exhaust gases.
Perhaps the most important and least
understood feature of Copeland incinerators is
the ability to pelletize most inorganic ash
residues to form the fluid-bed medium itself.
Many fluid-bed incinerators use sized sand as
the fluid-bed medium, but the presence of low
n-2-3
-------
melting point inorganics makes pelletization a
useful technique which in some cases permits
the recovery of a saleable product uncon-
taminated by sand or other diluents. Pelletized
ash is dust free and easy to handle. For this
reason, where possible, we endeavor to force
pellet growth in our incinerator systems.
In every case of forced pelletization, we find
reduced dust collection problems and a
generally easier operation. Pellet growth is a
function of temperature and surface area in
the bed itself; its rate is controlled by
controlling both temperature and unit area in
the bed. By occasional screen analysis of the
bed product, we can predict rate of growth
and adjust it to the needs of the system.
Fluid-bed systems capacities are a function
of superficial space velocity or the rate at
which the fluidizing gas is forced through the
fluid bed. A unit designed for a 2 ft/sec space
velocity will give 100 percent additional
capacity if enough air is forced through it to
raise the velocity to 4 ft/sec, provided the
velocity is not sufficient to elutriate the bed
material out of the reactor. Space velocities
are chosen to meet a given set of feed
conditions and will generally be in the range of
1.0 to 5.0 ft/sec. Installations have been built
for other purposes where the velocity has
exceeded 10 ft/sec.
Of great importance in basic design is the
opportunity to build into fluid-bed systems a
future capacity at minimal cost. By installing a
false brick lining in the fluid-bed zone at time
of original construction, combustion area is
provided for future use. For example, a project
requiring a 10-ft diameter reactor can get 30
percent future capacity built in by adding a 1-
ft thick false lining, for less than a 5 percent
increase in present cost.
DESIGN CONSIDERATIONS FOR FLUID-
BED INCINERATION PROCESSES
Fundamental to the design of any fluid-bed
unit is a clear understanding of the waste
material to be processed. Fluid beds will not
work on material too coarse in size to be
fluidized or having an ash content with a lower
melting point than the temperature necessary
for complete oxidation of the organic matter.
In this latter case, fluid beds do not differ
from the older type incinerators.
Generally it can be said that fluid beds will
burn anything that can be fed into them and
fluidized. Solid wastes with a minimum of free
surface water are generally blown into the
reactor, whereas drier materials can be fed by
means of a sealing type screw conveyor. Semi-
plastic sludges such as sewage sludge are fed
by screw conveyor or more simply by a pro-
gressing cavity pump. Thermoplastic
materials like grease are most readily fed by
first being melted and then pumped by
centrifugal pump.
Wherever possible in our design, we try to
build into the system enough freeboard
residence time to permit some heat exchange
between the incoming wet feed and the out-
going combustion gases. Since these gases are
generally wasted, evaporation of water from
the incoming feed by direct heat exchange has
the effect of improving thermal efficiency.
Before designing any fluid-bed system, we
pay particular attention to the chemical
composition of the material and look for trace
elements which might have a fluxing or fusion
point lowering effect on the ash content. We
have found, for example, that a fluid bed
burning organic material, the ash content of
which is 100 percent sodium chloride, can be
operated in excess of 1300°F without fusion
problems. Yet another product of the same
type having 1.5 percent sodium chloride could
not be burned in excess of 1150°F without
complete fusion. Obviously, this is a typical
eutectic problem, but critical in any incin-
erat^r design.
VIRTUES OF FLUID BEDS USED IN
SOLED WASTE INCINERATION
The higher combustion efficiency of
fluidized beds is attributable to a number of
II-2-4
-------
characteristics found only in part in other
combustion techniques. These may be
described as follows:
Extended Surface
Total surface area, in or on which the com-
bustion process takes place, is a very
important design consideration.
We have a commercial installation burning
60 x 106 Btu/hr (waste sulfite liquor) which
has surface area in the bed medium (90 tons)
equivalent to the surface of a super highway 70
miles long. Bed medium in this case is pel-
letized inorganic salt recovered as a by-
product of combustion. Pellet size is basically
14 to 65 mesh Tyler screen scale.
Obviously, the more surface area, the
better the opportunity for reaction between
oxygen in the fluidizing air and combustible
material.
Residence Time
Combustion is a time-temperature reaction
which is most efficiently carried out under
conditions which give instantaneous reaction.
Lack of time or temperature will make for
incomplete reaction and produce partial
products of combustion.
i
In systems where temperature must be
controlled at lower limits because of other
thermal considerations, residence time there-
fore becomes an important factor. We have
noted in systems burning waste sulfite liquor
at 35 percent solids that combustion at 1300°F
is instantaneous with no residual carbon left in
the bed. However, at 1250°F combustion is
slow and, if allowed to proceed for any length
of time, carbon build-up in the bed is
noticeable.
TYPICAL APPLICATIONS OF FLUID-BED
INCINERATORS
Fluid-bed incinerators are finding appli-
cation in the combustion of waste solids and
liquids because in many cases these waste
materials could not economically or
practically be incinerated by older more con-
ventional equipment. The very fact that many
waste solid materials have been used for land
fill, rather than completely destroyed by com-
bustion, is usually an indication of some
difficulty with conventional processes.
Since disposal usually implies an outright
cost to the producer, the most efficient system
must be found; fluid beds are being chosen
because of higher thermal efficiencies, better
control of odors and particulate matter
emissions, and a simpler process having
greater design latitude.
Fluid-bed incinerators have been
demonstrated by commercial practice to be
readily applicable to the combustion of the
following types of solid waste materials:
1. Domestic sewage sludge.
2. Municipal garbage.
3. Oil refinery wastes such as:
API separator sludge,
tank bottoms,
waste caustic streams, and
general refuse.
4. Petrochemical wastes such as:
hydrocarbon compound sludges, and
complexed waste inorganics.
5. Water treatment plant carbonate
sludges.
6. Packing house wastes.
7. Distillery slops.
8. Pharmaceutical plant wastes.
9. Clarifier effluents from most industries.
10. The destruction of lethally poisonous
materials.
11. Pulp and paper mill sludges and various
solid wastes.
The foregoing list is by no means complete;
it will be seen that some waste materials could
be destroyed by older techniques. However,
the high thermal efficiency of fluid beds
makes it possible to incinerate these materials
at much higher water contents without the use
of extraneous fuel, thus giving fluid bed
incinerators the nickname "water burners."
II-2-5
-------
FLUIDIZED BEDS VERSUS POLLUTION
CONTROL
Water Pollution
Combustion of waste material by fluid-bed
techniques is the ultimate of disposal tech-
niques in that no liquid effluent need result.
Ash produced by fluid-bed combustion is
completely burned out, impresses no BOD if
used for land fill, and many inorganic ashes
can be reused chemically. Scrubbing of
exhaust gases for air pollution control is a
necessity; but in most cases we scrub with the
waste liquid being combusted and hence
produce no new effluent.
Air Pollution
Most fluid beds used in industrial waste
disposal use forced air fluidization which per-
mits an easier route to high pressure drop
treatment of exhaust gas. We usually install
dry cyclones after the reactor and take up to
15 inches of water drop across them. These are
generally followed by wet scrubbing where up
to 50 inches of water pressure drop are taken.
We find that these systems, although costly
in terms of power consumption, are exceeding
by as much as 50 percent the most stringent
air pollution regulations now in effect. Our
experience indicates that, if sub-micron
particulate matter is to be taken out of
exhaust gas, high pressure drop across
scrubbing equipment is necessary. If sub-
micron particles escape the scrubber, a tail gas
plume will persist in the atmosphere. Any
persistent plume visible to the public is an
open invitation to investigation by pollution
control authorities.
Regardless of present air pollution
regulations, our experience tells us that the
ultimate regulation will demand no visible
plume whatsoever; this may include water
vapor plumes as well.
We believe that fluid-bed combustion
systems, properly designed and incorporating
the newest scrubbing techniques, offer the
best answer to ultimate air pollution
regulations.
A recent trend in municipal sludge incin-
eration, promoted by new EPA regulations,
will require that exit gas from sludge incin-
eration be exposed to a temperature of 1600°F
for 2 seconds to destroy malodorous gases and
chlorinated hydrocarbon compounds which
are produced by incineration at lower temper-
atures. This will make it impossible to dispose
of municipal sludge by older incineration
methods without fuel burning after-burners.
WORLD'S LARGEST SYSTEM FOR
BURNING BARK, DEBRIS, AND SLUDGE
In December of 1971, we brought on
stream a solid waste, fluid-bed incinerator,
which disposes of about 600 ton/day of pulp
and paper mill wet wood waste at a water
content of 65 to 70 percent without the use of
extraneous fuel. The composition of the feed is
given in Table 2.
The Great Lakes Paper Company Limited
at Thunder Bay, Ontario, had two solid waste
disposal problems.
The first problem was a huge pile of bark
(300,000 yd3) which had been accumulating
for years, and was beginning to make its
presence felt by spilling over into the local
river. The second problem arose from the
installation of clarifiers on the effluents from
the groundwood mill, the sulfite mill, and the
Kraft mill, the sludge from all of which would
have to be disposed of.
A conventional boiler was already in use at
the mill for burning bark fresh from the
barking drums. But (apart altogether from
questions of existing capacity) disposal of the
clarifier sludge and pile bark by such
conventional means would require that they be
further dewatered. The old piled bark, how-
ever, contained all manner of junk, including
a generous proportion of stones, which repre-
sented a potential source of damage to existing
bark presses. And the clarifier sludge, on
account of its slimy, fibrous nature, was very
II-2-6
-------
Table 2. PROPERTIES AND DAILY QUANTITIES OF FEED TO COPELAND SYSTEM
Sludge feed
Sludge from groundwood clarifier
Sludge from kraft clarifier
Rejects from groundwood mill
Rejects from kraft mill
Average properties and total amount
Wood waste feed
Bark and wood debris
Surplus bark
Average properties and total amount
Dry solids
%
25
25
25
25
25
35
35
35
Ash, %
2.2
2.3
1.9
1.9
2.2
2.0
2.1
2.0
Btu/lb
7100
7155
7155
7155
7155
8500
8500
8500
Ton/day
40
5
5
5
55
50
20
70
hard to dewater over 25-28 percent solids. At
this low consistency, it would cause
combustion troubles in the conventional bark
burning furnace.
Consequently, Great Lakes' management
turned to the Copeland fluidized-bed process,
which offered as one of its characteristics the
ability to burn woody materials with as little as
30 percent solids without supplementary fuel.
This would make it possible to burn, self-
sufficiently, the impressed bark and/or a
mixture of the unpressed bark and clarifier
sludge. A Copeland unit of 180 BD ton/day
capacity was therefore decided on.
The next question was whether this unit
should be attached to a waste heat boiler for
raising steam. It was recalled, however, that
24,000 Ib/hr of steam was currently being used
to heat 4000 gal/min of hot water for wood-
room showers. Since this heat demand was of
the same order as that expected to be available
by recovery from the Copeland unit, it was
decided to produce hot water directly in the
unit's heat recovery-scrubbing system. In this
way, the mill steam supply available for other
uses was increased, without going to the
expense of installing another waste heat
boiler.
Bark Feed
The bark from the old pile and other wood
debris is picked up in the mill yard by truck,
and combined with slasher sawdust, ground-
wood snipes, and wood scraps in a surge
hopper. From here the waste is mechanically
conveyed to a wood hog, which breaks it into
fragments which can be handled efficiently by
the subsequent pneumatic conveying system
(smaller than 6x6x6 inches). The shredded
waste resulting is then transferred by bucket
elevator to a storage silo. This silo holds 1
day's feed to the system, so that a man for
collecting bark and wood debris is needed on
one shift only. The silo has a live-bottom
vibrating hopper, which discharges the waste
through a vibrating feeder into a pneumatic
conveyor which injects it into the reactor
immediately above the fluid bed.
Stones up to 6-in. diameter in the bark and
wood debris, are conveyed along with wood
waste to the fluid-bed reactor by a pneumatic
feeder. Tramp metal is removed by an electro-
magnet.
Sludge Feed
The wet, fibrous clarifier sludge is
pneumatically conveyed to the disposal system
from the sludge collecting tanks, after
dewatering by filtration. Though it could be
injected directly into the reactor above the
fluid bed, the large amount of air required to
convey it would, under these conditions, enter
the reactor and critically lower its operating
temperature. Normally, therefore, the sludge
11-2-7
-------
is separated from its conveying air in a cyclone
from which it is mixed with the normal wood
wastes.
Combustion System
The fluid-bed reactor, is a carbon steel
vessel lined with insulating and refractory
brick. A five-stage, centrifugal, low-pressure
blower supplies the fluidizing and combustion
air. The fluidizing air is distributed into the
bed by an orifice plate separating the windbox
from the bed section. The fluid-bed zone
contains the fluidizing medium, which is made
up of sand removed from the waste wood.
Above the bed zone, the reactor widens out
to form the disengagement or freeboard zone.
The increase in reactor diameter here is
sufficient — even in view of the increase in
volume of gas phase due to the generation of
water vapor — to reduce the upward velocity
of the gases so that particulate solids will
disengage and fall back into the fluidized bed.
Overbed burners are provided for startups
as necessary. Auxiliary fuel is not required
when the system is fed waste materials at 75 to
120 percent of design capacity.
Sand Handling System
It is necessary, from time to time, to with-
draw excess sand and grit from the bed. This
is done'by discharging it into a sealing screw
conveyor, and then to a storage silo.
Gas Scrubber & Hot Water Generator
The hot combustion gases leaving the
reactor at about 1600°F pass into a two-stage
scrubbing system.
The first stage is an adjustable wetted -
throat venturi scrubber, in which the
scrubbing water is introduced over a weir and
atomized by the energy of the venturi. The
resulting fine droplets contact the ash
contained in the combustion gases and are
separated from the gas in the separator
section. The secondary scrubber consists of
two beds of fluidized packing, on which
scrubbing water is sprayed, thus trapping any
residual particulate matter while at the same
time picking up heat. Finally, the gases pass
through a demister to remove entrained fine
droplets before being vented to atmosphere.
Since no particulate matter, objectionable
gases, or odor have been detected in these exit
gases, a high stack to disperse them has not
been found necessary.
The scrubbing water from both first- and
second-stage scrubbers is collected in an 8500
gallon reservoir below the separator section,
from which it is recycled through a pump to
the two stages of scrubber trays. The recycle
line is provided with a purge line for main-
taining the desired temperature and ash
concentration in the scrubber liquid. From
this reservoir is taken 4000 gal/min of hot
water at 155°F used in the woodroom.
Operation & Control
One man operates the plant from a central
control panel. During startup opera'ting
temperature is reached with auxiliary fuel;
when the feed is ignited the auxiliary fuel is
shut off. Shut down is accomplished by
shutting off the waste feed and the blower
furnishing the fluidizing air. The sand bed
normally loses less than 100°F/day, so that the
unit can be started up without auxiliary fuel
after being down for as long as 6 days.
Cost of the fluid-bed installation was
approximately $1 million. Labor required is
1.3 man-days/day, and maintenance costs are
expected to be low. The operating credit of
24,000 Ib/hr of steam is reckoned at
S100,000/yr. A further credit is the valuable
land that will become available when the old
bark pile has been disposed of. Present feed to
the unit is 125 ton/day, but short runs have
shown that 180 ton/day is feasible. The Air
Management group of the Ontario
Department of Energy and Resources has
monitored the system thoroughly, and has
found the exit gases to contain no particulate
matter or objectionable gases.
II-2-8
-------
SUMMARY
In bringing this large fluid-bed solid waste
incinerator into operation, we have made some
interesting discoveries which lead us to believe
that the same technique can be used on such
other solid waste disposal problems, such as
garbage, etc.
To our great surprise, we have found the
reactor able to accept large numbers of stones
up to 6-in. in diameter, non-magnetic scrap
metal, cans, and a variety of things from
shredded truck tires to welding rods.
We are fluid izing bed rock particles up to
1-in. cubes and still maintaining good bed
stability.
The unit has never produced the first wisp
of smoke; air pollution control is so good that
pressure drop across the double scrubbing
system has been reduced from 50 in. H2O to
15 to 20 in. F^O, while still meeting the very
stringent air pollution regulations of the
Province of Ontario.
The ability of the unit to dispose of large
pieces of solid waste has encouraged Great
Lakes Paper to institute a change of feed flow
which will eliminate the hog completely. When
done, all waste, including waste pulp wood
sticks up to 4 ft. long will be fed directly into
the bed without any size breakdown.
The installation argues well for
uncomminuted garbage incineration by fluid-
beds — a development which is sorely needed
today.
II-2-9
-------
WATER
VENTURI
SCRUBBER
.HOT WATER
TO MILL
ASH
DISCHARGE
PREHEATER REACTOR
DIS. SCREW
BLOWER
FEEDER
Figure 1. Copeland solid waste Incinerator system.
-------
3. FLUIDIZED-BED COMBUSTION OF MUNICIPAL
SOLID WASTE IN THE CPU-400 PILOT PLANT
G. L. WADE AND D. A. FURLONG
Combustion Power Company
INTRODUCTION
A series of experiments is currently being
carried out to develop a large, high pressure
fluidized-bed combustor for incineration of
municiple refuse. This publication is based on
work performed under Contracts PH 86-68-
198 and 68-03-0054 with the National Envi-
ronmental Research Center, Cincinnati, Ohio,
of the Environmental Protection Agency.
The solid waste management concept
known as the CPU-400 is sized to accept 400
ton/day of solid waste from municipal packer
trucks. Solid waste will be conveyed from
receiving pits directly into shredders which
discharge into the air classifier units. Com-
ponents such as metal and glass with high
weight to aerodynamic drag ratios will be sep-
arated out and conveyed to ancillary disposal
or recycling processes. Lighter materials, pre-
dominantly papers and plastics, will be trans-
ported to a storage conveyor of sufficient
capacity to provide a continuous supply of
combustor fuel at a uniform rate.
The shredded and classified solid waste is
fed into the fluidized-bed combustor through
high pressure air-lock feeder valves and a
pneumatic transport line. In the fluidized bed,
inert sand-sized particles are buoyed and
mixed by an upward flow of air coming from
the compressor. Heat released by solid waste
combustion will maintain the fluidized bed
and exiting gas products between 1300 and
1700°F.
Several particulate removal stages will
operate on the hot combustion gases prior to
their admission into the gas turbine. Inert
granular residue will be removed from the
fluidized bed and particle collectors as
required.
The economic basis for the CPU-400 lies
primarily in the recovery of the energy con-
tained in solid waste by virtue of its high con-
tent of paper, plastic, and wood products. The
recovery and sale of electric power from the
disposal of solid waste markedly reduces the
cost of the operation. Depending on the value
of electrical power and other local conditions,
estimated net operating costs range from 2 to
$5/ton compared with current incinerator
costs of 8 to $14/ton. Where a stable market
exists for other reusable materials such as
metals or glass, the CPU-400 will also permit
recovery of these resources; the process acts to
concentrate these recyclable materials by
removing the large volume of combustible
materials. Revenues derived from recycling
will serve to further reduce net operating costs.
The CPU-400 is now in the early stage of its
development; system studies and subscale
experiments have been completed, and
development of the pilot plant is well under
way. The fluidized-bed combustor and partic-
ulate removal stages currently in development
testing are components of the pilot plant.
II-3-1
-------
ADVANTAGES OF FLUIDIZED-BED
COMBUSTION
The fluidized-bed reactor is a relatively new
approach to the design of high heat release
combustors. The primary functions of the air-
fluid ized inert bed material are to promote
dispersion of incoming solid fuel particles,
heat them rapidly to ignition temperature, and
to promote sufficient residence time for their
complete combustion within the reactor.
Secondary functions include the uniform
heating of excess air and the generation of
favorable conditions for residue removal.
The fluidized-bed reactor greatly increases
the burning rate of the refuse for three basic
reasons:
1. The rate of pyrolysis of the solid waste
material is increased by direct contact
with the hot inert bed material.
2. The charred surface of the burning solid
material is continuously abraded by the
bed material, enhancing the rate of new
char formation and the rate of char
oxidation.
3. Gases in the bed are continuously mixed
by the bed material, thus enhancing the
flow of gases to and from the burning solid
surface and enhancing the completeness
and rate of gas phase combustion
reaction.
A significant advantage of the fluidized-
bed reactor over conventional incinerators is
its ability to reduce noxious gas emission. Five
types of noxious gas are of potential concern.
The anticipated ability of the fluidized-bed
reactor to reduce each of these will be
separately discussed.
1. Oxides of Nitrogen. The relatively low,
uniform temperature of the fluidized-bed
reactor (1300 to 1700°F) limits the
formation of oxides of nitrogen. Most
combustion chamber concepts require a
hot, primary combustion zone to assure
good combustion efficiency; it is in these
hot zones that most oxides of nitrogen are
formed. Because of extensive mixing of
the fluidized bed, excellent combustion
efficiency is realized without a hot primary
zone.
2. Oxides of Sulfur. The effectiveness of
limestone for control of SO 2 emission
from coal combustion chambers is being
demonstrated. After injection into the
combustion chamber, the limestone is first
calcinated to lime. The SO2 is oxidized to
SO 3 on the lime surface and then reacted
to calcium sulfate (CaSO^, which remains
with the ash. A disadvantage of this
process for coal combustion is the
relatively large quantity of limestone
required. Test data show that limestone
may be only partially reacted because of
short residence time in the furnace result-
ing in calcium sulfate accumulation only
on the surface of the limestone. This
problem is reduced with the fluidized bed
since the increased residence time in the
fluidized bed strongly favors the capture
of sulfur by limestone. In addition^ solid
wastes average less than 0.5 percent sulfur
and the solid waste inerts already contain
significant quantities of calcium and mag-
nesium oxides. Thus, little if any lime-
stone additive is required when burning
solid waste in a fluidized bed.
3. Hydrogen Halides. The emission of
hydrogen halides, primarily HC1, can be
expected to be a significant problem for
future incinerators; probably more signif-
icant than SO 2 emission. Although
limited experimental work exists on HC1
suppression, chemical considerations
indicate that reactants similar to those
previously described for SO 2 suppression
may b.e effective for HC1 suppression.
4. Carbon Monoxide, and
5. Hydrocarbons. The highly mixed oxygen-
rich environment of the fluidized-bed
reactor provides very favorable conditions
for complete combustion, thus minimizing
carbon monoxide and hydrocarbon
emission.
II-3-2
-------
Finally, the fluidized bed is unique in its
ability to efficiently consume low quality fuels.
The relatively high inerts and moisture
content of solid waste pose no serious problem
and require no associated additional devices
for their removal.
THE CPU-400 PILOT PLANT
The CPU-400 pilot plant development is
currently nearing completion in Menlo Park,
California. A systematic evolution is planned
to ultimately include all major components on
a pilot plant level. The planned high pressure
configuration (with integrated gas turbine)
consists of four primary subsystems. Three of
these, broken down into their major
constituent parts, are illustrated in Figure 1.
These are the solid waste handling subsystem;
the solid waste combustor and gas preparation
subsystem; and the turbo-electric subsystem.
A fourth area, the control system, is both a
part of each of the other three subsystems and
also a separate subsystem in itself, causing the
other subsystems to interact properly with one
another and respond correctly to external
commands.
The purpose of the solid waste handling
subsystem is to prepare the solid waste for
combustion. This includes separation of the
shredded material into two constituents
(materials dominated by combustible and
inert components, respectively), storing the
combustibles until ready for use in the
combustor, and metering the solid waste to the
combustor. Unprocessed municipal waste is
initially loaded onto the shredder conveyor by
a skip loader. The conveyor modulates the
feed to the shredder based upon electrical
loading of the shredder motor. After shred-
ding, the material is fed to the air classifier
where the light, combustible materials are
pneumatically lifted and transported to the
storage bin, while the heavy, inert materials
drop out for subsequent separation and
recovery. Metering of the prepared solid waste
fuel is accomplished through a variable-speed,
servo-controlled outfeed conveyor in the
storage bin along with variable speed transfer
and weighing conveyors to the combustor feed
points.
The second subsystem of the high pressure
pilot plant consists of the solid waste
combustor, three particulate removal stages,
ash removal equipment, and interconnecting
piping and valving. The solid waste is
introduced into the combustor through air-
lock feeder valves. The material is burned in
the solid waste combustor and the resulting
hot gases are then cleaned of suspended solid
material in three stages of separation. Fly ash
material is removed from the particle
separator collection hoppers by pneumatic
transport to a baghouse filter.
The third subsystem is the turbo-electric
unit consisting of a gas turbine, generator,
switch gear, and load bank. The compressor
section of the turbine supplies the cold air for
the solid waste combustor fluidization. The
resulting hot gases, after being cleaned in the
separators, are used to power the compressor
turbine and the power turbine. The generator
is driven by the power turbine and generates
power which is subsequently controlled by the
switch gear. The electrical energy output of
this pilot plant system will be dissipated in a
combination load and light bank. In the full
scale system, the electrical power will be
delivered to a customer for subsequent use in
the municipality.
The control subsystem interacts with these
systems to control their respective outputs in
response to commanded set points. This
system, which features analog controllers
under the supervisory control of a digital
process computer, will also monitor numerous
signals to provide data acquisition, logging,
out-of-tolerance alarming, and status display
functions.
In the low pressure configuration operated
to date, the gas turbine compressor is replaced
by a facility blower and exhaust gases are
cooled by water spray. Consequently, there is
no further discussion of the turbo-electric sub-
system in this paper. Since incorporation of
the process control computer into the pilot
II-3-3
-------
plant is incomplete at this writing, few
references will be made to the control
subsystem.
Solid Waste Handling Subsystem
The solid waste handling subsystem
processes municipal solid waste unloaded by
packer trucks at the pilot plant facility.
Photographic views of the subsystem are
shown in Figure 2. The flow of solid waste is
illustrated in Figure 3. A 16 ft3 skip loader is
used for transfer of the raw solid waste to the
shredder's conveyor hopper. Through elec-
trical current level controls, the conveyor
supplies material on demand to the shredder.
Shredded material is ejected into the air class-
ifier unit. In this unit heavy metallics, rocks,
glass, etc., are dropped out; the remainder is
pneumatically transported up through the
chamber and into the cyclone above the
storage tank where it is disengaged and
deposited in the tank. The exhaust stack of the
cyclone connects to a second blower on a dust
filter cyclone. This unit pulls off vapors and
dust from the center of the storage tank
cyclone, depositing the dust in a container at
floor level and venting excess air to the
atmosphere.
Very little hand sorting of the delivered
solid waste is conducted prior to shredding.
On rare occasions an item too large for the
shredder inlet (e.g., a large truck tire) is
encountered; these are manually removed.
Power limitations in the 75-hp shredder have
dictated that massive metal items and fabric
bundles (both infrequently found in municipal
solid waste) also be removed. Other highly
visible items such as automobile tires and
mattresses, though successfully shredded on
occasion, are normally removed in the
interests of maintaining high through-put.
Under these conditions a nominal rate of 1.5-
2.0 ton/hr has been established, and over 350
tons of solid waste have been shredded and air
classified through August 1972. Concessions
to subscale pilot plant operation essentially
disappear when 1000 hp shredders, such as
planned for the CPU-400, are employed.
The shredded and air classified solid waste
fuel form is a mixture whose visual
appearance is homogeneous and dominated
by identifiable paper products. A typical
pound of the material consists of 0.30 Ib of
water, 0.52 Ib of ash-free combustibles, and
0.18 Ib of inerts (including ash). The latter two
fractions are typically sub-divided as follows:
Description % by weight
Paper
Thin metals,
fine glass,
and dirt
Wood
Dust
Plastics
Textiles
55
15
13
13
2
2
As would be expected, all of the preceding
values are subject to considerable variation.
As an example, moisture content in the final
fuel form ranges from 10 to 40 percent by
weight depending upon the origin of raw
material (e.g., residential or commercial
sources), time of year, weather conditions, and
municipal collection policies.
Air classified solid waste fuel is shown in
the storage tank in Figure 4. The accompany-
ing view of the empty tank interior shows three
of the four rotating bucket chains which sweep
the floor to move material into the outfeed
conveyor located in the slot. Free trailing ends
of the bucket chains allow edge contact with
the stored material pile regardless of pile size.
Variable speed hydraulic drives in the sweep
system and outfeed conveyors are controlled to
maintain material levels in a small hopper at
the outfeed conveyor discharge point.
The hopper volume lies above a transfer
conveyor and upstream of material leveling
devices that produce a fixed material height
on the transfer conveyor. Fuel is delivered by
the transfer conveyor to a weighing conveyor
(equipped with load cells to provide
continuous measurement of fuel flowrate) and
then dropped through a static splitter
assembly -ifrto the two airlock feeder valves.
II-3-4
-------
Variable speed electric drive systems are used
on the transfer and weighing conveyors. Speed
command signals are both slaved to a single
fuel demand signal; this implements a
responsive, continuous volumetric flow
element for a combustor temperature control
system.
Two airlock feeder valves receive the solid
waste fuel from the weighing conveyor and
deliver it to the air transport lines, which in
turn pneumatically deliver the material to the
fluidized bed. Each valve (Figure 5) is powered
by a 25-hp electric motor drive and associated
gear box. As the valves turn, empty pockets in
the valves receive material from the top, rotate
through 180 degrees past a sealing wall, and
deliver the material into the transport lines at
the bottom. Material is continuously delivered
to the combustor through continuous rotation
of the valve.
Solid Waste Combustor and Gas Preparation
Subsystem
The fluidized-bed combustor is contained
within a vertically oriented cylindrical
pressure shell with dished heads. The outside
diameter is 9.5 feet and overall height is 23.5
feet. Three layers of insulation protect the 3/8-
in. carbon steel pressure shell cylindrical
sections from high combustion zone tempera-
tures. A wear-resistant firebrick inner liner is
backed up by a liner of insulating brick. These
refractory layers are separated from the shell
by a thin layer of packed ceramic fiber
insulation designed to isolate the shell from
stresses induced by differential thermal
expansion. Insulation in the top dome is
provided by a castable refractory held in place
by standard hangers.
The fluidized bed is supported by a flat
carbon steel plate welded to the pressure shell
and covered by two layers of castable refrac-
tories that provide insulation and wear resis-
tance. Penetrating this assembly are 161 2-in.
pipes capped with wire mesh air diffusers.
Other penetrations from the air plenum
chamber beneath the plate permit bed
temperature and pressure measurements.
The circular cross section fluidized bed has
an area of 40 ft2' and is designed to operate
with a superficial velocity in the 5 to 7 ft/sec
range. A nominal 2-ft bed (unfluidized state) is
used together with a 12-ft freeboard
(unfluidized bed surface to exhaust duct
centerline).
Penetrations through the pressure shell and
refractory insulation into the fluidized bed
provide for two solid waste feedpoints. Two
feedpipes bolted to outer shell bosses extend
into the bed. Solid waste is fed into the bed
along the length of these pipes via a slanted
cut on the bottom side. The design and
positioning of these pipes is based on earlier
tests where oxygen concentration
measurements established the dimensional
characteristics of combustion zones. The
result is a configuration which, in low pressure
testing, has demonstrated very satisfactory
operation with respect to geysering due to
feedpipe air flow, minimization of local fuel-
rich zones, and reduction of heat release above
the bed.
Other bed penetrations provide for possible
removal of excess bed material and for six oil
guns to permit fluidization combustion of
auxiliary diesel oil. This normally unused
auxiliary fuel, available primarily as a
developmental tool for backup service in
maintaining or establishing desired test condi-
tions, is mixed with air and carried through
the inner of two concentric pipes. The outer
pipe of each gun carries cooling air.
Initial bed heating is accomplished by hot
products of combustion from an oil burner
located in the top dome of the combustor. This
downward firing burner forces hot gases
through the bed in a "back heating" mode
that is capable of heating the bed from
ambient conditions to 1100°F in 90 minutes.
This bed temperature, being above the auto-
ignition temperature of either solid waste or
diesel oil, is an appropriate initial condition
for successful fluidized combustion.
II-3-5
-------
In the low pressure configuration,
fluidizing air is supplied to the combustor's air
plenum by a 125-hp positive displacement
blower which can deliver up to 7000 scfm at 3
psig. The blower also supplies fuel transport
air in a parallel path to the fluidized bed. By
appropriate valving, the same blower is used
to drive the back heating mode.
The top cylindrical section is removable
and contains the exhaust port, instrumenta-
tion, and observation ports. Exhaust into the
first particulate removal stage is carried by a
double-walled pipe with 26-in. carbon steel
outer wall and 20-in. type 310 stainless steel
inner liner. The annular space is packed with
ceramic fiber insulation.
Two particulate removal stages have been
tested in low pressure operations by the time
of this writing. The first, known as the
alumina/sand separator, treats a problem
peculiar to fluidized-bed combustion of solid
waste fuel; the removal of inert particles of
elutriated bed material, and the handling of
particles generated by the presence of
aluminum in the fuel. While the majority of
aluminum is removed by air classification,
about 0.25 percent by weight of the processed
fuel is aluminum in the form of foils, beverage
can pull-rings, and other thin or trapped
particles. The fluidized bed melts, fragments,
and partially oxidizes this material into
particles having molten internal aluminum
and a frozen oxide (alumina) surface shell that
inhibits further rapid oxidation. Previous
testing has shown that elutriated particles of
this type tend to generate sizable deposits with
high aluminum content on impingement
surfaces in the exhaust gas stream. The
possible participation of soft (at combustion
temperatures) bottle glass particles or other
binding agents in this deposition mechanism
should not be discounted, even though present
evidence does not seem to indicate that they
play a dominant role.
A promising solution to the problem of
handling partially molten aluminum and
other sticky particulates is to provide a curved
surface in the first separator stage where the
turning exhaust gas can produce a controlled
deposit. Scrubbing action associated with the
continual impingement and turning of particle
laden exhaust gases acts to promote further
oxidation at the deposit surface. It also
promotes the necessary erosion of resultant
alumina-rich particles to stabilize deposit
shape and size. Most of the inert alumina-
based particles thus generated as well as the
silica-based elutriation particles are then
collected in the settling chamber formed by
the bottom of the vessel. In testing to date,
these separated particles have been allowed to
remain in the settling chamber. For future
tests, residue particles collected in the
alumina/sand separator stage will be removed
on a continuous basis.
Fly ash and fine bed material particles are
removed in two stages of inertial separation.
The first inertial separator stage has been
successfully tested by the time of this writing
and the final stage (similar design features)
will be added in the immediate future. An
inertial separator assembly consists of an
inertial separator tube holder with a residue
collecting hopper, an insulated cylindrical
housing with inlet and outlet flanges, a dished,
insulated flanged head, and stainless steel
liners. As shown in Figure 6, the hot gases
enter the inlet cavity to the cyclone tubes
through an internal circumferential passage.
The entering gas then turns and flows through
the spirally fintied annular section of each
tube. These fins impart rotational motion into
the gas which centrifuges the particles to the
outside wall. Particles spinning along the out-
side wall of the tube will decelerate and fall
through the opening at the bottom of the tube
into the collecting hopper. To avoid plugging,
the gravitational flow of particles from each
tube is assisted by a secondary bleed flow.
Particulates are removed from the hopper
through the opening on the bottom of the
hopper cone. Gas which enters the cyclone
tubes turns 180 degrees and exits through the
center tube into the owtfet manifold of the
vessel. In the first inertia-1 separator, space is
II-3-6
-------
provided for 48 6-in. tubes; half of the spaces
are plugged for low pressure operation,
however. When incorporated, the second
inertial separator will have space for 100 3-
1/2-in. tubes.
Material from the inertial separator hopper
is pneumatically transported to the baghouse
filter through a finned line. The high tempera-
ture baghouse assembly includes a puffback
bag cleaning system, an exhaust blower, a
holding bin, and unloading valve.
A photographic view of current subsystem
components (Figure 7) also shows the gas
turbine in the foreground. Present plans are to
install and integrate this turbine into the pilot
plant by the end of 1972.
SUMMARY OF COMBUSTION TEST
RESULTS
The seventh combustion test conducted on
the low pressure configuration described in
preceding paragraphs was completed in
August 1972. The test featured 35 hours of
fluidized combustion and produced an
extensive quantity of data while consuming
over 69,000 Ib of shredded, air classified solid
waste. During one particular 24-hr period of
special performance interest, a number of
solid waste and exhaust gas stack samples
were drawn for laboratory analysis. Most of
the discussion to follow will be based on data
obtained in this latter period.
A laboratory facility has been established
at Combustion Power Company to perform
many of the required experiments on material
samples drawn from the pilot plant process
operation. Included in the installation are
equipment and procedures for the determina-
tion of:
1. Solid waste moisture fraction,
2. Solid waste inerts fraction,
3. Solid waste heating value,
4. Granular material size distributions,
5. Particulate loading and size distributions
from gas samples, and
6. HC1 gas concentrations in exhaust gas
samples using titration techniques.
Fuel Properties of Prepared Solid Waste
A combination of 12 laboratory samples
and adjustments based on long term pilot
plant mass balance measurements yielded the
following average weight distributions for the
shredded and air-classified solid waste used
during the 24-hr period.
Ash-free combustibles
Moisture
Inerts
0.516
0.301
0.183
A series of 12 bomb calorimeter
experiments in the laboratory on dried parallel
samples produced an average higher heating
value of 6437 Btu/lb. Using an approximate
ultimate analysis of C3QH48OJ9 for the
combustibles fraction and converting to a
lower heating value based on the combustibles
only, a corresponding average value of 8087
Btu/lb of combustibles is found. This result is
in very good agreement with values
determined by applying heat balance relations
to observed pilot plant temperature and flow
measurements. It also correlates well with
expected values for a cellulose-like material
such as approximated by the C3QH48Oi9
formulation.
Compared to more conventional fuels, the
heating value of solid waste is in a sense
degraded by the presence of inerts and
moisture in greater-than-normal concentra-
tions. On the other hand, the greater mass
flows of combustibles and water vapor
required to vaporize and heat the water are
exploited by the gas turbine cycle as a natural
form of water injection. In addition, fluidized
beds have a demonstrated ability to consume
"low quality" fuels and hence to eliminate any
need for a drier or other additional fuel pre-
processing.
With the preceding fuel properties, the
steady state combustor operating point
described in Table 1 is typical. Note the 178
percent excess air level associated with this
operating condition, which is one of the
reasons that very high combustion efficiencies
are realized.
II-3-7
-------
Table 1. TYPICAL OPERATING POINT OF THE
LOW PRESSURE CPU-400 PILOT PLANT CON-
FIGURATION
Solid waste fuel flowrate, Ib/min 36.6
Combustor air flowrate (includes feed line), Ib/min 330
Excess combustor air, % 178
Solid waste fuel inlet temperature, "F 70
Air inlet temperature, °F 130
Exhaust gas exit temperature, "F 1,500
Total combustor heat release, Btu/min 152,700
Heat loss to ambient (two vessels), Btu/min 5,000
Mole fraction of major exhaust gas contituents:
Oxygen
Carbon dioxide
Water
Inert mix of nitrogen and argon
0.122
0.063
0.099
0.716
The excess oxygen ratio is strongly
influenced by fuel moisture fraction. If
moisture fraction is increased to 0.50, for
example, and inerts fraction is reduced to
0.131 (as would happen if the added moisture
was obtained by direct addition of water to the
previous fuel mixture), then the required solid
waste flowrate jumps to 64.9 Ib/min, and the
excess oxygen percentage drops to 119
percent. Further increases in moisture fraction
produce even sharper drops in excess oxygen
ratio so that it is not possible to use moisture
fractions much over 0.60. An interesting inter-
pretation of the preceding numbers is that the
combined injection of 18.5 Ib/min of water
and 30 percent moist solid waste raises the
consumption of the latter (and hence the total
heat release) by 27 percent to 46.4 Ib/min.
FIuidized-Bed Combustor Performance
With a large, relatively shallow fluidized
bed served by one or two feed pipes, a small
but significant portion of the total combustion
process occurs in the lean phase or freeboard
above the dense bed. For a 40 ft2 bed area 2
feet deep, for example, bed to exhaust temper-
ature rises of 260° and 160°F were observed
with one and two feed pipes in operation,
respectively. Since these levels of gas phase
combustion (afterburning) are quite stable for
well controlled solid waste injection, it has
been concluded that satisfactory operation can
be obtained in either case at low pressure
conditions.
Very high overall combustion efficiencies
have been demonstrated by several
approaches. First, careful heat balance calcu-
lations based on process measurements have
consistently produced apparent heating values
that, when compared to laboratory
calorimeter results, produce efficiency values
slightly in excess of 100 percent. Flow
measurement errors and occasional high
heating value components in the solid waste
not found in laboratory samples are the
probable explanations. A second method
based on a carbon balance as indicated by
continuous exhaust gas measurement of CO2,
CO, and hydrocarbons plus post-test analysis
of free carbon in the separator residue stream,
generated an average efficiency value of 99.76
percent. Finally, there is no visible exhaust
smoke, odor, or other common evidence of an
inefficient process.
System Inert Material Balance
After installation of the combustor was
completed it was loaded with a 2-ft "starter"
bed of commercial 16-mesh beach sand having
a bulk density of about 100 lb/ft3. After 18
hours of fluidized operation on solid waste fuel
in the first six tests, the cooled bed was
observed to still be 24-in. deep and free from
excessively large particles or clinkers. No
material had been removed except through the
elutriation process. The size distribution was
somewhat larger (i.e., more fines and more
coarse particles) and the bulk density had
dropped to 91 lb/ft3.
The bed generated in this fashion was
successfully used in the following 35 hr test.
Again, no material was removed except by
elutriation. Post-test analysis and inspection
then showed that the bed was 25-in. deep and
had an average bulk density of 87.5 lb/ft3.
Thus the 7,300-lb bed processed and
elutriated more than 12,300 Ib of inerts in 35
hours while only growing about 10 Ib. This
natural replenishment of bed material by the
inert content of the fuel is a phenomenon
peculiar to solid waste combustion. Earlier
long duration tests (240 hours) on a 2.2 ft 2 bed
H-3-8
-------
coupled with computer simulation of transient
size distribution histories had shown that the
steady state bed could be expected to have
quite acceptable properties. Consequently, the
bed maintenance and elutriation control
problems appear to be grossly simplified.
In the 35-hour test, 72.0 percent of the
elutriated material (i.e., 8890 Ib) was collected
in the settling chamber of the alumina/sand
separator vessel. This inert granular residue
has considerable promise as a construction
material. Another 25.4 percent (3140 Ib) was
collected by the first stage inertial separator in
the form of fine fly ash. The remaining 2.6
percent, 320 Ib of very fine fly ash, left with
exhaust gases. Most of this latter material will
be collected by the second stage inertial
separator in subsequent tests.
Gas samples were withdrawn from the first
stage inertial separator outlet at 3-hour
intervals for laboratory analysis of particulate
loading. Early samples showed a total loading
of 0.057 gr/scf and a loading of 0.025 gr/scf
for particles greater than 5 f*m in size; both
values are indicative of satisfactory first stage
performance. Later markedly higher values
confirmed post-test findings that the hopper
beneath the 24 active tubes has failed to drain
properly and therefore progressively plugged
the ash discharge sections of some tubes,
rendering them ineffective. Twin vibrators
have been installed to solve this problem in
future tests.
Exhaust Gas Composition
A set of instruments has been installed for
on-line concentration measurements of six
specific constituents of the exhaust gas. A
seventh gas chromatograph instrument for the
on-line measurement of HC1 is under develop-
ment and will be added to replace current
laboratory titration procedures. The gas
sampling, conditioning, and distribution
systems are integrated with the instruments
and analog recorders in a mobile, rack
mounted, complex. The six current instru-
ments, measuring principles, and associated
gases are:
1. Beckman 715, Polarography, Oxygen;
2. Beckman 315B IR, Infrared, Carbon
dioxide;
3. Beckman 400, Flame lonization Detec-
tion, Hydrocarbons;
4. Beckman 315B IR, Infrared, Carbon
monoxide;
5. Theta LS-800AS, Electrochemical
reaction, Sulfur dioxide; and
6. Theta LS-800AN, Electrochemical
reaction, Nitrogen oxides.
The sampling probe leads into a stainless
steel sampling train that includes particle
removal elements*," gas cooling and drying,
controlled reheating to 100°F, a common
manifold, and flow control elements for each
instrument. Various calibration and zero
adjustment methods have also been
incorporated.
Measurements from the recent pilot plant
test are presented in Table 2 together with
pertinent emission standards. All instrument
records were relatively steady and free from
apparent anomalies considering the potential
heterogenous composition of the fuel form.
The 12 discrete laboratory samples (2-hr
intervals) and subsequent HC1 analyses
showed more variance from a low value of 40.5
ppm to a high of 122.4 ppm.
The first two entries of Table 2 correlate
well with analytical predictions such as
contained in Table 1. The very low values for
the second two, coupled with the average 45.8
ppm of free carbon (weight basis), are indica-
tive of the high combustion efficiency. Sulfur
dioxide presently appears not to present a
pollution control problem, probably owing to
the relatively low, sulfur content of municipal
solid waste and the apparent capture of
existing SO 2 by the bed material.
Measured NOx levels are much closer to
proposed standards and somewhat higher
than originally expected. As confirmed by
other investigators, much of the nitrogen
II-3-9
-------
Table 2. EXHAUST GAS CONSTITUENTS AND
REFERENCE DATA
Constituent
Oxygen
Carbon dioxide
Hydrocarbons
Carbon monoxide
Sulfur dioxide
Nitrogen oxides
Hydrogen chloride
Proposed or expected
EPA emission standards
(maximum)
800 ppm
2000 ppm
300 ppm
NAa
NA
Mole fractions
measured during
24-hour CPU-400
pilot plant test
11.7%
6.8 %
14.4 ppm
4 1.7 ppm
20 ppm
162 ppm
90.3 ppm
a Not available.
emitted in these compounds derives from the
fuel rather than the combustion air. Other
experience, however, indicates that the
concentrations may be expected to drop as
pressure is increased.
Depending on emission standards yet to be
established, the measured levels of HC1 pose a
potential control requirement. There appears
to be a number of effective remedial measures
that rely upon fluidized-bed characteristics,
however, if forthcoming pressurized combus-
tor tests should firmly establish a requirement.
CONCLUSIONS
Shredding and air classification operations
on municipal solid waste produce a fuel form
having very satisfactory physical and chemical
properties for energy recovery through
combustion in a fluidized-bed reactor. A solid
waste handling subsystem with reliable
components has been developed and
extensively operated.
The fluidized-bed combustor has fulfilled
its promise as a highly efficient, easily fed,
readily controlled reactor of simple design and
capable of utilizing low quality fuels.
Grade 16 silica beach sand is an acceptable
starter bed material. The inert content of
typical solid waste provides a natural bed
make-up material that leads to a satisfactory
steady state bed composition. As a result, the
need for elaborate bed maintenance or anti-
elutriation devices is minimized if not
eliminated.
Test results to date show no problem with
fluidized-bed residue buildup. Relatively large
metal and inert particles entering the active
bed experience gradual oxidation or attrition
to typical bed size particles and eventually are
elutriated to be collected by particle
separators. No bed material agglomeration is
experienced in the 1270 to 1450 °F bed
temperature range.
Gas phase combustion above the fluidized
bed has been reduced to acceptable values
without resorting to undesirable remedies
such as extensive internal bed structures,
numerous fuel feed points, deeper bed, multi-
stage combustors, etc.
Earlier combustor freeboard and exhaust
system deposit problems due to the aluminum
content of solid waste appear to have been
solved.
Performance of the first stage inertial sepa-
rator has been very promising and is expected
to further improve at high pressure conditions.
Reliable handling and transportation of
removed hot, fine fly ash has posed develop-
ment problems.
Exhaust gas sampling instruments indicate
that the pollution control effort required for
the CPU-400 process can be expected to be
minimal.
No evidence of serious hot gas subsystem
material corrosion or erosion problems has
been found.
II-3-10
-------
a
w
SOLID WASTE PROCESSING
SHREDDED
WASTE
STORAGE
ALUMINA/SAND
SEPARATOR
FIRST
INERTIAL
SEPARATOR
GENERATOR
EXHAUST DUCTS
CONTROL ROOM
BAG HOUSE
FILTER
SECOND
INERTIAL
SEPARATOR
AIR INLET
Figure 1. CPU-400 pilot plant pictorial.
-------
SHREDDER CONTROLS
wassail
CONVEYOR
CONVEYOR
H CONTROLS
Figure 2. Photographic views of the solid waste handling subsystem.
II-3-12
-------
H
oo
PACKER TRUCK
DUST KOP
LOADER
EIDAL SHREDDER WITH
ANTI-BALLISTIC HOOD \
BLOWER
AIR CLASSIFIER
SOLID WASTE SYSTEM CONTROL PANEL
CYCLONE
WEIGHING CONVEYOR.
SHREDDED SOLID WASTE
STORAGE TANK
AIRLOCK FEEDER
ATLAS CONVEYOR
INSTRUMENTATION PANEL
Figure 3. Pilot plant solid waste handling subsystem schematic.
-------
Sweep Bucket Chain
Shredded, Air Classified Solid Waste Fuel
II-3-14
Figure 4. Photographic views of solid waste storage tank interior.
-------
Figure 5. Airlock feeder valve.
II-3-15
-------
Figure 6. Inertial separator schematic.
II-3-16
-------
Inertia!
Separator
Alumina/Sand
Separator
Gas Turbine
(uninstailed)
Figure 7. View of solid waste combustor and gas preparation subsystem.
-------
4. FLUIDIZED-BED COMBUSTORS
USED IN HTGR FUEL REPROCESSING
B. J. BAXTER, L. H. BROOKS, A. E. BUTTON,
M. E. SPAETH, AND R. D. ZIMMERMAN
Gulf General Atomic Company
ABSTRACT
High-temperature gas-cooled reactors (HTGR) utilize graphite-base fuels. Fluidized-bed
combustors are being employed successfully in the experimental reprocessing of these fuels. This
paper presents a general discussion of the reprocessing method and describes the two types of
fluidized beds being used.
INTRODUCTION
The high-temperature gas-cooled reactor
(HTGR), as developed at Gulf General Atom-
ic, is a helium-cooled, graphite-moderated
reactor. The fuel in an HTGR consists of fis-
sile microsphere particles containing U-235,
recycle microsphere particles containing U-
233, and thorium fertile particles contained in
a hexagonal fuel element, shown in Figure 1.
The HTGR fuel recycle operation consists of
shipping spent fuel to recycle facility, repro-
cessing the fuel to recover the U-233 and U-
235, refabricating the U-233 and U-235 into
recycle fuel, shipping the refabricated fuel
from the recycle facility to the reactor, and
ultimately storing the radioactive fission pro-
duct wastes.
The fuel reprocessing sequence starts with
the head-end operation shown in Figure 2, in
which the fuel in the HTGR fuel element is
separated from the graphite body by crushing
and fluidized-bed burning. Subsequent
head-end operations separate particles
containing U-235 from ash containing U-233,
thorium, and fission products. The metal
oxide ash is dissolved to create a solution of
uranium, thorium, and fission products; the
silicon-carbide-coated U-235 is the residue.
The U-235 is separated mechanically, and the
uranium and thorium are recovered
individually from the fission products. The
recovered U-233 and thorium are stored for
reuse as fuel. .The radioactive wastes are
disposed of in appropriate storage facilities.
II-4-1
-------
HEAD-END REPROCESSING
Head-end reprocessing for HTGR fuel con-
sists of a crush-burn-leach process.1 Fuel ele-
ment size reduction is the first step in head-
end reprocessing. Two major criteria govern
this step: (1) the fuel must be crushed to a
suitable size for maintaining fluidization
quality in the fluidized-bed burners, and (2)
the crushing system must minimize fuel parti-
cle breakage to prevent undesirable crossover
of fissile and fertile product uranium.2
A three-stage crushing system has been
adopted for the reprocessing plant, based on
the experimental testing of commercially
available equipment using full-sized fuel ele-
ments. This crushing system is presently being
tested.
Primary reduction is done in a large, over-
head eccentric jaw crusher; secondary reduc-
tion in a small, overhead eccentric jaw
crusher; and tertiary crushing in a double-roll
crusher. The tertiary crusher product,
nominally minus 3/16 in., is pneumatically
conveyed to the fluidized-bed burner feed
hoppers.
Crushed fuel is fed to the top or base of a
continuous, exothermic fluidized-bed burner,
shown in Figure 3, by an auger feeder. (The
term exothermic is used to describe the burner
that generates sufficient heat to maintain
operating temperature.) The feed rate is auto-
matically controlled by the off-gas carbon-
monoxide concentration, which has been
shown to be proportional to the graphite sur-
face area exposed in the bed.
Both the crushed graphite and the silicon-
carbide-coated fissile particles serve as the
fluidizing media. The heat generated by
burning is removed by forced-air cooling in a
clamshell jacket surrounding the burner and
an off-gas heat exchanger. The fluidizing gas
fed to the burner is oxygen with a small
amount of inert gas (i.e., CO2, N2), and the
flow is automatically controlled to maintain
the bed temperature. The burner product
removal rate is automatically controlled by the
bed pressure drop, which is proportional to
the bed weight.
The burner off-gas with entrained fines is
passed through a cyclone separator and a
sintered metal filter for fines removal, before
being cooled and proceeding to off-gas treat-
ment. Off-gas treatment removes the fission
products including noble gases before release
to the environment. Fines are presently being
recycled to the burner in the experimental
program.
If the exothermic burner is operated with
top feed and no fines recycle, the elutriated
fines from the burner (both TRISO/TRISO
and TRISO/BISO flowsheets) are also added
to the feed stream for the last burning step.
This mixture now constitutes the feed material
for the endothermic (requiring heat input
from a furnace to maintain operating temper-
ature) fluidized-bed burner.
The exothermic burner product is fed to a
batch-operated endothermic fluidized-bed
burner, shown in Figure 4, where the
remaining graphite is burned and the thorium
and uranium-oxide kernels are exposed. The
silicon-carbide-coated fissile particles serve as
the inert fluidizing media. The feed stream to
the endothermic burner will not sustain
exothermic burning to the low carbon level
required in the subsequent processing steps.
The burning in the endothermic burner,
therefore, proceeds from exothermic condi-
tions, with the heat removed from a clamshell
surrounding the burner, to endothermic con-
ditions, with heat supplied by resistance
heaters located in the clamshell. The off-gas
from the burner is treated in the same manner
as that from the exothermic burner. The pro-
duct from the endothermic burner is pneuma-
tically conveyed to the leaching system.
The thorium and uranium oxides are dis-
solved in acid thorex [13M HNO3 - 0.05M HF
- 0.01M Al (NO3)3] in a steam-jacketed
cylindrical vessel with gas sparge mixing. This
leaching vessel is run as a refluxing, batch
leacher.
11-4-2
-------
The insoluble silicon-carbide-coated fissile
particles and the unburned carbon must be
separated from the mother liquor before the
solution can be fed to the solvent extraction
system for uranium purification and thorium
recovery. A centrifugal separator receives the
entire slurry from the leacher. Solids retained
on the centrifuge screen are washed with fresh
leach solution, which becomes the leach solu-
tion for the next batch of solids from the
endothermic burner. The washed solids from
the leacher are then air-dried and transferred
to a screen classifier, where the fissile particles
are separated from the silicon-carbide hulls.
The waste solids are processed as wastes, and
the fissile particles are stored for later
processing, by a method similar to that for the
fertile particles, to recover and purify the
uranium.
The clarified leach solution is evaporated
and steam-stripped to an acid-deficient condi-
tion for use as feed to an acid thorex3'4
extraction process. The acid thorex solvent
extraction process is used for the
decontamination and purification of the U-
233 and thorium and for the separation of the
U-233 and thorium from each other.
FLUIDIZED-BED BURNERS
Exothermic Fluidized-Bed Burners
Figure 3 depicts the exothermic fluidized-
bed burner presently being used in the
experimental program. Exothermic burners
with both 4-in. and 8-in. diameters are being
used; construction of a larger burner is
planned for early next year. The large burner
will become the full-sized commercial plant
test unit. Preliminary nuclear criticality
calculations have shown that this burner can
be about 16 inches in diameter.
Operability
The exothermic burners have been
operated on a routine basis for the last 18
months. Startup is initiated by heating a
charge of coke (1200 g for the 8-in. and 400 g
for the 4-in.) to ignition temperature (700°C)
with a carbon monoxide-oxygen gas mixture.
This gas mixture is introduced into the burner
with a standard cutting torch that is ignited by
two spark plugs. After the coke is ignited,
fluidizing oxygen and the graphite-base feed
are introduced.
The steady-state bed properties of the
exothermic burners are listed in Table 1; the
bed contains 2 to 5 percent burnable carbon.
Table 1. AVERAGE BED PROPERTIES
FOR EXOTHERMIC BURNERS9
Size
-3/16 in.
-1/8 in.
-869 urn
-550nm
-420 jim
-375 jxm
-250 Mm
-125 /^m
-74|^m
-44jxm
-, 1/8 in.
+ 869 ^m
+ 550nm
+ 420 ^m
+ 375pm
+ 250 (^m
+ 125^01
+ 74 /urn
+ 44 ^m
Wt%of
total sample
0
0.2
67.9
4.9
0.2
1.0
5.5
4.3
2.8
13.2
a Burnable Carbon: 3.0%
Average Particle Size: 590 Mm
Steady-state operation is achieved for experi-
mental purposes in about 4 hours by adding
the estimated steady-state bed composition
directly to the burner immediately - after
startup. Product removed after 4 hours is also
near steady-state and exhibits about the same
properties as the bed. The present feeding
method is a variable-speed auger controlled
automatically by the carbon-monoxide
concentration in the off-gas. The nominal off-
gas concentrations are 1 to 3 percent carbon
monoxide, 0 to 6 percent oxygen, and 60 to 95
percent carbon dioxide. This composition is a
function of the inert fluidizing gas diluent (i.e.,
air versus CO2). Product removal rate is
automatically controlled using the pressure
drop across the bed and regulating a variable-
speed drive motor on the product removal
auger. Bed temperature is automatically
controlledrby regulating the fluidizing oxygen
II-4-3
-------
supply to the bed. Temperature profiles, off-
gas compositions, bed and filter pressure
drops, and mass flows of important streams
are monitored continuously.
Fluidization quality has been difficult to
define. Normal operation is with a well-mixed
bed that occasionally slugs. Distributor plates
are not presently being used but will be fully
tested in the near future. At present, operation
is with a cone base and a ball check valve.
Preliminary tests of perforated plates, bubble
caps, and sintered metal screens were all
successful to some extent, and the beds
appeared to maintain good fluidization
quality.
Feed
The feed to the burners is presently defined
as minus 3/16-in. graphite-based material.
This feed size was established by gradually
increasing the size from minus 1/16 inches to
minus 1/4 inches. Poor fluidization occurred
with minus 1/4-in. feed, as witnessed by local
"hot spots" in the bed. Returning to minus
3/16-in. feed eliminated this problem. Table 2
shows the average properties of the exothermic
burner feed.
Table2. AVERAGE PROPERTIES OF
EXOTHERMIC BURNER FEED3
Size
-3/16 in. + 869f*m
-869 J^m + 550 /urn
-550 ^m + 420 (xm
-420 |i*m + 375 urn
-375 /xm + 250 fxm
-250 j^m
Wt%of
total sample
64.6
25.0
1.5
1.5
1.5
5.9
% burnable
carbon in
fraction
100
21
78
97
98
98
aTop Density: 1.25g/cm3
Bulk Density: 1.08g/cm3
Angle of Repose: 35°
Average Burnable Carbon: 80%
Average Particle Size: 854 jjtm
Heat transfer
The heat transfer problem encountered is
somewhat different from that occurring in
most fluidized-bed work; because of nuclear
criticality considerations, the cooling medium
is limited to air. Figure 5 shows the heat
balance for a typical 8-in. diameter
exothermic burner run. The over-all heat
transfer coefficients for the off-gas heat
exchanger and clamshell cooler are also
shown.
Fines recycle
One of the major problem areas in
exothermic burner operation is the burning of
fines that have elutriated from the burner.
Since burning efficiency and radioactive hot-
cell constraints play a major role in the
process, work has focused on burning the fines
by recycle to the bed rather than in a separate
fines burnup cell.
Fines carryover has been about 22 percent
of the burn rate when the furnace has
operated at normal conditions; i.e., at a burn
rate of approximately 200 g/min and a
superficial fluidization velocity of 3 to 4 ft/sec.
Of these fines, about 98 percent is collected by
a cyclone and 2 percent by a filter chamber.
The elutriated fines are described in Figure 6.
The present operating mode is to recycle
the fines by blending them with the graphite-
base feed stream. This composite is fed to the
bottom of the fluized bed, and the fines
appear to burn successfully when steady-state
is achieved.
The nominal burn rate for the 4-in. and 8-
in. exothermic burners is about 33 g
carbon/hr-ft2, which corresponds to 50 and
200 g carbon/min, respectively. A burn rate of
50 g carbon/hr-ft 2 (corresponding to 75 and
300 g carbon/min, respectively) is planned for
the 4-in. and 8-in. exothermic burners. Burn
rates of 125 and 475 g carbon/min for the 4-in.
and 8-in. burners (about 84 g carbon/hr-ft2)
have been achieved for short periods.
Operation at these high burn rates- is not
II-4-4
-------
possible for long periods because of
limitations with existing equipment. Future
studies will be aimed at defining these values
over long-term run conditions.
Scale-up
The design considerations for the larger
exothermic fluidized-bed burner include both
a theoretical approach and scale-up factors
from the 4-in. and 8-in. burners. To date, the
problems encountered for scale-up have been
in defining both a suitable transport disen-
gaging height and exact heat transfer values.
Although the theoretical and experimental
predictions of heat transfer values in the area
of the bed-wall-clamshell are in agreement, it
is difficult to define the proper heat transfer
coefficient in the transport disengaging height
in which the overall heat transfer coefficient
rapidly decreases with reactor height. These
values will be determined experimentally in
future experiments in which "sectioned"
clamshell coolers will be utilized. *
Endothermic Fluidized-Bed Burners
Figure 4 depicts the 4-in. diameter
endothermic fluidized-bed burner presently
being used in the expermental program. It is
planned to convert the 8-in. diameter
exothermic burner to an endothermic burner
by adding resistance heaters to the clamshell
interior and moving the filter chamber to
directly above the burner for future scale-up
testing.
Operability
The endothermic fluidized-bed burner has
been operated as a batch burner on a routine
basis for the last 18 months and is presently
being automated. The automation consists of
the furnace temperature control loop and a
burner control system. The burner control
system is a repeating unit of automatic tem-
perature control (by regulating the fluidizing
oxygen flow) and a series of programmed
events. The programmed events control fluid-
izing gas flow, batch product dump valve,
batch pneumatic feeder, and refluidizing
oxygen flow until the automatic temperature
control loop takes over and the sequence starts
again. This control cycle is repeated automati-
cally and provides a batch-continuous oper-
ation. Although the endothermic burner is
similar to the exothermic burner, the fines are
burned by containing them within the burner.
The low fluidizing velocities used during the
endothermic burn stage allow burning the bed
to less than 1 percent carbon.
Feed
The feed to the endothermic burner is the
product from the exothermic burner. The size
distribution of this feed is highly variable,
depending on the flowsheet being processed,
but at all times it is easily fluidized. The
average particle size of the feed varies from
200 to 400 jum for the various flowsheet
varieties.
Heat transfer
The burn rates achieved in the 47in.
endothermic burner are about one-half that of
the 4-in. exothermic burner, because the bulk
of the heat transfer occurs in the transport
disengaging height. Also, the operating period
in the endothermic stage of burning is a slow
burning process. Burn rates of 25 to
35 g carbon/hr-ft2 are achieved during-the
exothermic burn period and 5 to
10 g carbon/hr-ft2 in the endothermic burn
period. An average burn rate of about 20 to
25 g carbon/hr-ft 2 or 20 to 25 g carbon/min is
achieved during batch-continuous operation.
REFERENCES
1. Nicholson, E. L., et al. Burn-Leach
Processes for Graphite-Base Reactor Fuels
Containing Carbon-Coated or Oxide Parti-
cles. Oak Ridge National Laboratory. U. S.
Atomic Energy Commission. Oak Ridge,
Tenn. Report Number ORNL-TM-1096.
1965.
2. Steward, H. B., et al. Utilization of the
Thorium Cycle in the HTGR.
In: Proceedings, 4th Geneva Conference
on Peaceful Uses of Atomic Energy. 1971.
II-4-5
-------
3. Blanco, R. E., L. M. Ferris, and D. E. 4. Blanco, R. E., L. M. Ferris, C. D. Watson,
Ferguson. Aqueous Processing of Thorium and R. H. Rainey. Aqueous Processing of
Fuels. Oak Ridge National Laboratory. U. Thorium Fuels - Part II. Oak Ridge
S. Atomic Energy Commission. Oak Ridge, National Laboratory. U. S. Atomic Energy
Tenn. Report Number ORNL-3219. 1962. Commission. Oak Ridge, Tenn. Report
Number ORNL-3418. 1963.
II-4-6
-------
_E=\
31.22 in.
196 lb=89 kg GRAPHITE
235 Ib =116 kg TOTAL
CROSS SECTION
BURNABLE
POISON ROD
COOLANT
CHANNEL
FUEL ROD
STACK
HELIUM
FLOW
Figure 1. HTGR fuel element.
II-4-7
-------
oo
OFF-GAS
OFF-GAS
SECONDARY
CRUSHER
(JAW)
TERTIARY
CRUSHER
(DOUBLE ROLL)
SECONDARY
BURNER
(ENDO)
f I COOLANT OUT
SCREEN
DISSOLVER SEPARATOR
ACID
TO LIQUID-
•LIQUID
EXTRACTION
FISSILE PARTICLES
INSOLUBLE WASTE
Figure 2. Head-end reprocessing simplified flow diagram.
-------
CYCLONE
HEAT
EXCHANGER
PLATES
COOLING
AIR, IN
COOLING
AIR, OUT"
THERMOWELL
BALL"
02 AND N2
±H—*• OFF-GAS
16 FILTERS
SOLIDS LEVEL INDICATORS
AUGER
PRODUCT REMOVAL
Figure 3. Exothermic fluidized-bed burner.
II-4-9
-------
FILTERS
COOLING AIR
HEAT EXCHANGER
FLUIDIZED BED
02 AND
=: v OFF-GAS
FEED HOPPER
N2-PULSE
PNEUMATIC FEEDER
MOVABLE ELECTRIC
FURNACE
INTERMEDIATE SINGLE
BATCH HOPPER
PNEUMATIC
TRANSPORT
Figure 4. Endothermic fluid-bed burner.
II-4-10
-------
SENSIBLE HEAT
DIFFERENCE IN
PRODUCTS AND
REACTANTS AND INERTS
16 kcal/min (1.0
HEAT OUT FROM
HEAT EXCHANGER SKIN
15 kcal/min
(0.9%)
HEAT OUT FROM
SPOOL PIECE SKIN
125 kcal/min
(7.7%)
HEAT OUT FROM
CLAMSHELL SKIN
81 kcal/min
5.0%
RADIATION
FROM CONE
181 kcal/min
(11.1%)
Ahr
V
A H OF HEAT EXCHANGER
COOLING AIR
150 kcal/min
' (9.2%)
U=lBtu/hr-ft2-°F
, Ahr=1630 kcal/min
(83% ACCOUNTED FOR)
AH OF CLAMSHELL
COOLING AIR
780 kcal/min
U=20Btu/hr-ft2-°F
Figure 5. Heat balance and heat transfer coefficients for a typical 8-in.
exothermic burner run.
II-4-11
-------
44g/min
EXO BURNER
200 g/min
BURN RATE
-250 ji+125 )>
•125 ji -f74|i
-74jj+44}i
-44 ;i
1
CYCLONE
FILTER
CHAMBER
43 g/min
95% CARBON
wt%
0.25
0.25
4.4
95.1
1 g/min
96% CARBON
0.0
0.0
4.0
96.0
Figure 6. Elutriated fines description.
II-4-12
-------
5. STUDIES ON THE COMBUSTION
OF NATURAL GAS IN A FLUID BED
W. E. COLE AND R. H. ESSENHIGH
Pennsylvania State University
ABSTRACT
Natural gas has been burned with air in a fluid bed of 1.4 ft2 cross-sectional area, using
expanded alumina of 14 to 16 ASTM mesh. The gas/air mixture is supplied premixed. Initial
problems of operation, now solved, concerned rapid ignition, and uniform distribution. Light-off
initially required two hours or more before combustion was uniform throughout the fluid bed. This
period is now consistently 5 to 10 minutes. With very uniform gas distribution, obtained with a
distributor of novel design, fluid bed depths were reduced from 6 to 1 inch for complete
combustion, even up to superficial (hot) flow velocities of 12 ft/sec. Combustion intensities at
2000°F and 100 percent excess air were in the region of 106 Btu/ft3-hr based upon bed volume.
Experiments were carried out with excess air ranging from 60 to about 150 percent, with extinction
at the higher value. Gas rates ranged from 6 to 11 ft3/min. Bed temperatures ranged from 1700 to
2400 °F, rising with fuel/air ratio. Air rich extinction boundaries were mapped over a range of
fuel/air ratios. Bed temperatures, gas analysis, and pressure drop on a vertical axis through the
bed have also been measured. Superadiabatic temperatures and a lowering of the lean flam-
mability limit have been observed. These two observations are explained qualitatively by the pre-
heat effect of the fluid-bed particles. Reaction rates are significantly faster than for free-burning
gas in a premixed flame at comparable temperatures and gas concentrations. The effect is
attributed to the bed particles. Data on the physical behavior of the bed, with good theoretical
agreements, are also given.
INTRODUCTION
This paper describes experiments on the
combustion of natural gas in a fluid bed to
investigate problems of ignition, even distri-
bution, and combustion speed. Combustion
applications of fluid-bed technology can
include incineration, steam raising, possibly
heating of crucibles, billets, etc. However, use
of gas in such applications can present prob-
lems.
Fluid-bed combustion of natural gas is,
according to conventional belief, beset with
difficulties by comparison with combustion of
oil or coal. Reasons generally cited are: (1)
excessive initial heat up time of the bed, up to
24 hours in -the case of very large units; (2)
inefficient (incomplete) combustion in the bed,
leading to (3) excessive freeboard (overbed)
temperatures from final burnup; (4) high bed
temperatures, said to be necessary compared
with liquid or solid fuels to allow an adequate
margin of safety above the commonly
accepted ignition temperatures for gas; resul-
ting in (5) excessive thermal stress on heat
exchangers included to improve the overall
thermal efficiency.
II-5-1
-------
The most significant problem is combus-
tion efficiency, once light-up has been com-
pleted. In large units it is customary to avoid
premixing the gas and air, because of the
explosion hazard; so the gas is supplied
directly to the bed through one or more supply
ports. A considerable fraction of the gas may
then bubble to the surface (2) and, in burning
overbed, generate the excessive freeboard
temperatures (3) that can overstress heat
exchangers (5). Such bubbling behavior paral-
lels remarkably the behavior of coal volatiles,
if a high volatile coal is fed too quickly into a
fluid bed at too restricted a point.
As explanation of the incomplete combus-
tion inside the bed, the bubbling effect sug-
gests immediately that it would be due to
inadequate mixing before the gas breaks the
surface. This view is substantiated by esti-
mates of transit time through the bed which
considerably exceed the expected reaction
times. Nowhere, however, were we able to find
any direct substantiation of this expectation in
any published literature on gas combustion in
fluid beds. Indeed, information on this topic is
conspicuous by its absence. There were a few
references available indicating that gas could
be burned in fluid beds, but not very satisfac-
torily; and nowhere was there any source of
quantitative data, so far as we could establish,
that could serve as any basis for engineering
design involving commercial use of gas in fluid
beds. Furthermore, although our findings as
reported here have now substantiated the
expectation of very fast reaction in the bed
once the gas and air have mixed, at the outset
of the project there were some few indications
that slow mixing might not be totally responsi-
ble for the incomplete combustion in the bed.
The argument was based on the usual behav-
ior in the bed immediately after light-off. The
bed acted as a flameholder with all reaction
above it and no reaction whatever inside it
until the whole bed had heated up to a definite
temperature. Clearly, the cold particles were
providing thermal quenching and/or chain
termination. Equally clearly, once hot, the
particles could be expected to reverse their
action and become sources for thermal and/or
chain initiation of reaction. However, this still
did not rule out the possibility that the pres-
ence of very large surface areas, even of hot
solids, in the middle of the reaction zone
might so significantly alter the reaction mech-
anism that the reaction rate could still be
appreciably hindered or accelerated. Our
findings do, in fact, suggest that the reaction
rate may be increased; there are also indica-
tions of some other interesting aspects of
behavior, such as wider combustion limits and
super adiabatic temperatures.
Specifically, however, our starting point
was the problem of even distribution and com-
bustion. In the process of investigating this, in
conjunction with developing an alternative to
the conventional distributor plate, we found a
means of reducing the light-off time. The unit
thus developed, having very even gas distribu-
tion and good fluidization, was then suitable
for more detailed measurements of gas tem-
peratures and analyses in the bed, leading to
the results indicated above.
What we have to report therefore are po-
tentially valuable data for engineering design
and data of more fundamental significance
but, withall, not without relevance to design.
This is, we believe, the first public report on
quantitative behavior of gas burning in a fluid
bed.
n-s-2
-------
PRELIMINARY EXPERIMENTS
The first experiments carried out, sum-
marized here, led to a new design concept
described below. The unit used for these initial
experiments utilized a square, 3/8-in. steel
distribution plate, of 13-in. side, carrying 81
perforated studs (a 9x9 array on 1.5-in.
centers). The walls were uncooled, constructed
of 2.5-in. series superduty (Rockspar) fire
brick. Air and gas were supplied separately to
mix in the bed. Air was supplied via an air box
under the distribution plate, flowing into the
bed through 64 of the distributor studs. Gas
was supplied through the remaining 17 studs
which were connected by a progressively bifur-
cating line from a 15 psi supply point. The
perforations in the distributor studs were 1/8-
in. diameter, drilled horizontally from the
outside to meet hollow centers. The bed
material mostly used was Type 8-F blown
alumina of 6 to 12-in. depth in the initial
experiments. Instrumentation included: gas
and air meters, wall thermocouples, wall
pressure taps, suction pyrometer for gas
temperatures, and sheathed thermocouples
for bed temperatures.
With the unit as described, the bed
fluidized satisfactorily in cold flow; but light-
off, with combustion predominately in the
bed, was never initially achieved. The gas all
burned above the bed with the bed acting as a
flameholder. It was clear that mixing was the
problem because the flames stabilizing on the
top of the bed appeared in a set of rings of
flame whose pattern was determined by the
layout of the gas supply studs. Various means
of overcoming this were tried. Unsuccessful or
partially successful means included: other bed
materials including a fine sand that cascaded
through the stud perforations into the wind
box; and rotary stirring of the bed by
mechanical and pneumatic means. (For the
latter, a set of four supplementary air pipes
were lowered into the bed to produce rotation
of the bed by horizontal jets aimed at an
imaginary circle. This was partly successful.)
Success was finally achieved by resting the
fluid bed on an underlying bed of limestone
rocks (Type 2B) 4.5 inches deep. This was
found to be an excellent mixer and distributor,
and with this arrangement light-off became
possible with reaction ultimately drawn back
into the bed. (It was at this point that most of
the instrumentation was added.)
Initially, however, light-off was still the
excessively long process generally claimed,
taking anything up to 2 hours (for a 6 to 12-
in. bed). Furthermore, as the flame was drawn
into the bed it exploded in a random sequence
of strikebacks followed by blow-off (i.e., with
the flame oscillating between the top and bot-
tom of the bed). In some instances, the explo-
sions were violent enough to displace some of
the wall bricks even though they were mor-
tared into position and partly held there by
steel angle frames. Once hot, the bed burned
evenly without explosion.
Light-off, however, was ultimately reduced
to 10 minutes or less (3 to 4 minutes is about
the shortest time so far). It was clear that the
flame would not strike back into the bed until
the whole of the bed had reached a
temperature that would permit it. (The exact
temperature is not known.) The heatup took a
long time because the particles were only
heated at the top of the bed (where the bed top
acted as a flameholder) and were cooled again
by the cold fluidizing air and gas as they
mixed back into the bed. The solution was to
fluidize the bed in progressive stages and, with
correct gas and air settings, the process is
virtually automatic. The gas and air in
stoichiometric proportion are initially set at
about 10 percent of the rate required for
incipient fludization in the cold; the gas is lit
to burn over the bed with the bed acting as a
flameholder. The gas and air rates are then
promptly increased to about half of the cold
fludization requirements. The over-bed flame
heats the top (unfluidized) layers of the bed
and the rising gas and air. When the top bed
particles are hot enough, reaction can then
start just inside the top of the bed. The
consequent local rise in temperature of the
II-5-3
-------
gases increases their velocity to above fluidiza-
tion velocity, fluidizing the top layer of
particles. The same process then operates to
fluidize the next static layer; this is then
repeated, with smooth, non-explosive progres-
sion of the fluidized interface down to the
bottom of the bed. The gas and air supplies
are then adjusted to the required operating
levels, and start-up is complete. There is no
reason why the same procedure could not be
used to cut light-off time on a commercial-
scale bed by an order of magnitude, or more.
Solution of the even distribution and rapid
light-off problems were the main tasks under-
taken in these preliminary experiments. Work
was then started on combustion behavior in
the bed itself, with initial indications that
reaction was substantially slower than it would
be in free-burning gas at the same
temperature. However, problems due to
leakage, mainly through patched cracks
produced during the explosive ignition tests,
made measurements erratic. An entirely new
bed was therefore constructed for further
measurements, as described in the equipment
section.
Preliminary to the reconstruction, however,
some cold model tests of air distribution and
fluidization were carried out to aid design.
(1) The first model used water as the
fluidizing medium, fed into the bottom of a
12-in. high, 3.5-in. diameter, Plexiglas
cylinder, through a 1-in. pipe containing five
4-in. sections of 3/8-in. copper tubing to serve
as flow straighteners. An air line to a hyper-
dermic syringe in the center of the flow
straighteners served as a fine bubbler for
visual flow tracing in the mixing studies. The
fluid bed was simulated by glass beads on a
wire screen, about 1 inch above the 1-in. inlet
pipe. Beads of 3, 4, and 5mm diameter were
used. With water supply up to 6 gal./min., the
behavior of fixed, fluidized, and spouting beds
could be observed.
(2) The second unit was an air model of a
3-in. diameter feed pipe feeding into a square
clay pipe, of 13-in. side, or into 6-in. diameter
Plexiglas pipe, again using glass beads to
represent the distributor with finer beads or
actual bed material above.
These studies indicated the importance of
uniformity of the underlying distributor rock,
but provided the information to indicate that a
single supply point of relatively narrow
diameter compared with the fluid bed diam-
eter could be used as long as the distributor
rocks (or coarse particles) were deep enough.
EQUIPMENT
The present unit is illustrated in Figure 1.
It consists essentially of a refractory lined
cylinder of castable refractory, cast inside two
oil drums, standing on a welded steel platform
30-in. high. This construction provides an
outside diameter of 22 in. and a height of 47
in. The shell is wrapped with 540 feet of 3/8-
in. copper tubing for cooling and monitoring
the wall loss. Walls, floor, and roof are 3-in.
thick, of Hydrecon Tabcast, a 3600°F erosion-
resistant castable refractory. The air is
supplied to the unit from two staged 24-oz.
blowers (360 scfm capacity) through a 3-in.
diameter pipe cast in the floor, and the
exhaust gases leave through a metal cased flue
liner, 12-in. diameter and 2 ft long, leading
into a 16-in. diameter stack.
The bed is blown alumina, of mesh size -14
+ 16 ASTM, resting on a 6-in. deep bed of
crushed refractory, of mesh size -3 + 5 ASTM.
At the top of the 3-in.diameter air supply pipe,
the crushed refractory is supported on a 1/16-
in. perforated steel plate. The crushed
refractory provides all necessary air and gas
distribution across the full width of the unit.
An ignition burner, at a height of 22 inches,
is provided for startup and safety. To observe
the bed, two inspection ports closed with 2-in.
Vycor discs are provided in the top; the top
also contains two holes for insertion of probes.
For access to the bed, the top can be removed
using a 1/2-ton differential chain hoist.
II-5-4
-------
The air and gas are fed to the bed through
a mixer, after being metered by rotameters
reading to 280 and 15 scfm full scale,
respectively. The mixer is a 2-1/2-in. diameter
(Pyronics) venturi unit, followed by a Tee to
carry the mixture into the bottom of the bed;
the other leg of the Tee provides a sump for
any material falling through the inlet hole. To
permit increased flow rate through the bed, if
required, a bypass valve is provided around
the air rotameter, with an "Annubar" flow
meter to measure these higher flow rates.
Cast into the furnace at heights of 9 and 12
inches above the base are two rows of 11 holes
capped with 1/4-in. pipe nipples for
instrument access or solid feed. Twenty-three
pressure tap holes lined with 1/4-in. porcelain
are cast into the side at 1/2-in. intervals from
the bottom to a height of 8 inches, and every 1
inch thereafter to a height of 16 inches. Before
casting, 1.5-in. pieces of stainless steel were
soldered to the oil drum shells at all pressure
tap and thermocouple stations to protect the
porcelain from breakage. Shielded thermo-
couples flush with the inside wall are provided
at elevations of 1, 3, and every inch thereafter
up to 19 inches. Two sets of thermocouples are
also mounted at intervals of 1/2-in. depths
into the wall to monitor the wall temperature
profile. All thermocouples are Chromel-
Alumel,'with read out by a Leeds & Northrup
24-point recorder. Pressure is monitored by a
36-in. tube, well type manometer. A Chromel-
Alumel thermocouple is used for bed
temperature measurements with readout by a
single channel Honeywell recorder; a check on
bed temperature can also be made by optical
pyrometer.
A draft gauge with range from + 0.005 to -
0.015-in. we is connected to the horizontal
flue-run for pressure monitoring. Two water
cooled probes 48-in. long are used for gas
sampling, one in the stack and the other in the
bed. Gas is continuously monitored for CO2
and CO with infrared instruments, and for C»2
with Thermox analyser using a fuel cell
as the sensor element. Incomplete combustion
of the gas is also determinable directly by an
MSA total combustibles analyser.
Initial startup, when the unit was first
completed, was accomplished without
difficulty using the technique described
earlier. However, some difficulty was initially
experienced in maintaining fluidization for
any length of time. The bed would start to
blind at some point. Fluidization would be
lost; the gas/air flow would be diverted from
those regions which would cool, thus tending
to prevent refluidization. Careful analysis of
the problem suggested that it was probably
due to fines in the bed material. At all events,
removal of fines by screening on a 16-ASTM
sieve eliminated the problem.
Progressive improvements to the design
and method of operating the unit resulted in a
steady improvement in fluidization; as fluidi-
zation improved, the bed depth required for
complete combustion dropped steadily. In the
preliminary experiments on the first unit, bed
depths up to 12-in. were used. In the current
unit, beds were initially 4 to 6-in., but were
finally reduced to 2 inches and then to 1 inch
with the even fluidization achieved. For
heating purposes (in the bed) or solid wastes
incineration greater depths are required. For
our immediate purposes here, however, since
no interesting combustion behavior occurred
above the bottom 1 or 2 inches of the bed, the
top layers were "omitted" to allow easier
access by probes to the regions of interest. It
should be emphasized, of course, that until the
present unit was completed, and the very
uniform fluidization obtained, the evidence
indicated that combustion required 3 to 6
inches.
PHYSICAL PROPERTIES OF THE FLUID
BED
In the course of the combustion
investigations, the porosity and other
properties of the hot bed were measured, some
of which were obtained as a matter of
necessity. Because of their potential
engineering value for design they are
11-5-5
-------
summarized here. It is emphasized that these
are measurements made at high temperatures,
up to 2400°F, with combustion following;
since the bed depths were small, the data
should be valid for all bubble free conditions.
The data of principal interest were the
expanded bed heights, porosities, gas
velocities, and residence times, all of which are
needed for interpretation of the combustion
data.
Values of pressure drop (AP) between the
top of the bed and a point below the top of the
bed were first measured by traversing the bed
with an open-ended, water-cooled probe. (The
kinetic head contribution is too small to cause
any determinable error.) Measurements were
taken for six different fluidizing velocities,
with bed temperatures ranging from 1800 to
2350°F; the normalized results of AP against h
are shown in Figure 2 where h is measured
from the bottom of the bed. The normalizing
parameters used were the total pressure
drop APtand the (expanded) bed depth L. The
top of the bed was identified by a break in the
slope of the AP against h line. At first glance it
is evident that the plot of Figure 2 is respec-
tably linear, in agreement with theoretical
predictions quoted in standard tests; e.g., 1, 8.
With closer inspection, the slight curvature of
the line is self-evident; the curve is probably
due to the temperature variation and hence
the velocity variation through the bed. How-
ever, the departure from linearity is small
enough that it can be neglected for our pur-
poses.
The normalizing parameters, APt and L0,
were also found to obey simple theory.
Equations given by Davidson and Harrisoni
have been used. Renormalizing their equation
(1.11) against the initial bed depth (LQ) at
incipient fluid ization velocity (Uo) when the
initial porosity is £o, we obtain
where APot is the pressure drop across the bed
at incipient fluidization. However, (APt/AP0t)
is unity in the fluidization region, as Figure 3
illustrates. Note also that the value of AP0t is
below that required to balance the bed weight
This is in accordance with Trevedi and Rice's
experiments.5 A further simplication of
equation (1) is possible by using equation
(1.22) of reference 1, again renormalized to
yield
= (£/£o)3(l-£o)/(l-£)
(2)
Figure 4 substantiates this equation,
illustrating the linearity obtained by plotting
£3/(l-£) against U, where U is determined
under the hot condition. Figure 4 also shows
the actual variation of £ with U. Values of E
were calculated from the particle denisty (a)
and the bulk bed density (p) (calculated from
bed weight, depth, and areas) using
=l-p/«a=l-0.464
(3)
The incipient fluidization porosity (£o) is
0.536, which is somewhat above the value of
0.476 for cubic packing of uniform spheres.
Since the alumina particles are quite good
spheres it is not unrealistic to assume this less
dense packing is due to particles bridging void
areas.
Substituting equation (2) in equation (1) yields
L/Lo = (l-£o)/(l-£).
(4)
This relation is substantiated by Figure 5. The
slope, equal to (1 <- £o), has a value 0.46
yielding £Q= 0.54, which is good agreement.
Figure 5 also shows the variation of (L/LO)
with U, with the fitted curve calculated from
equation (2) adopting the experimental values
of U0 and £Q. The gentle curve could be well
approximated by a straight line, which is a
consequence of e3 /(I-e) varying almost linearly
with !/(!-£); it is approximately proportional
to U.
II-5-6
-------
Absolute calculation of the incipient fluid -
ization velocity is not quite so satisfactory
although the experimental value can be
bracketed. Again, using an expression given
by Davidson and Harrison,1
U0 = 0.00081 (cgd2/n) (cm/sec)
(5)
where d is the particle diameter (=0.0510 in.)
and (•< is the dynamic viscosity. Uo was
calculated for ambient temperature and
1800°F, knowing the average weight of one
particle. The predicted values were 3.28 ft/sec
and 1.21 ft/sec (at the higher temperature).
This bracketed the experimental value of 2.0
ft/sec. Since the numerical factor of 0.00081
was determined for experiments on fluidizing
with water,1 the agreement is acceptable.
From the data given, calculation of the
actual (as opposed to the superficial) velocity is
straight forward using a porosity correction;
the same occurs for the residence time in the
bed Os). The results of the two calculations are
given in Figures 6 and 7, respectively, with
bounding values for£0 and £ = 1 included for
comparison. The residence time data, in
particular, are needed for discussion of the
combustion behavior.
The general agreement with theory
established here would support the use of the
equations tested for engineering design and
scale-up, if used with care.
COMBUSTION BEHAVIOR
General
The problems of light-off have been fully
covered above. Once lit, combustion can be
maintained indefinitely as long as no blinding
of the bed by fine particles occurs. In the
earlier experiments, when fluidization and
mixing of the fuel and air were relatively poor,
the over-bed gases were periodically flecked
with yellow as bubbles of gas broke the surface
and burned in the freeboard space. This
behavior was progressively eliminated by
improved fluidization and mixing. With the
present arrangement, utilizing premixing in a
venturi mixer, there can be no bubbles of fuel
rich gas. However, it was clear that the speed
of reaction was still strongly influenced by the
fluidization quality. When this was poor, 4 to
6 inches were apparently required for com-
bustion. As fluidization was improved, the
space required was progressively reduced
until, as mentioned above, it could be accom-
plished well within a 1-in. bed. Under these
conditions, all that can be seen through the
sight glass in the top is the red hot bed-top in
continuous motion without any spouts, and
with particles welling up and disappearing
again.
In the experiments next described there
were three objectives. With the expectation
that gas-fired fluid beds will be increasingly
used in commercial practice, attention was
first directed at two aspects of safety: (1) if fuel
to a bed is cut off, at what minimum tempera-
ture will it relight? and (2) if a surge of air or
neutral gases leans out the fuel-air mixture, at
what gas percentage and bed temperatures
will there be extinction? With information on
these first two, attention was then given to
behavior in the bed in an attempt to determine
whether combustion in a particle-filled volume
is affected in any way by the presence of the
particles.
Relight
To investigate relight behavior, the
procedure was to set up the bed in normal
operating condition and then to switch off the
fuel and the air, or to decrease the air. The bed
would cool; and periodically at recorded
temperatures the fuel supply would be
restarted. The observation then made was
whether or not the bed temperatures would
start to rise again. This was taken as a
condition of relight.
The experiments on relight were carried
out at a fairly early stage in the investigation,
with beds 4 to 6-in. deep, and mostly at fuel-
air ratios closer to stoichiometric than to the
lean limiir Under these conditions, relight was
ii-s-7
-------
always successful down to bed temperatures of
750 °F (400°C). Lower temperatures than this
were not investigated for safety reasons. This
is substantially below the values generally
listed in standard data tables; e.g., 4, for auto
(spontaneous) ignition temperatures.
Reference 4, for example, quotes: 1290°F
(700°C) for the stoichiometric mixture; greater
thai* 1200°F (650°C) for the most easily
ignited mixture; and adds the comment that
under pressure, the temperature is never less
than 880°F (470°C). It seems fairly clear that
the method of determination is too different
from a fluid bed for the results to be relevant.
Extinction
The procedure for determining extinction,
at the low limit, was to start with the bed in
normal operating condition at some suitable
fuel and air rate, and then to lean out the
mixture by stepwise reduction of the fuel rate.
When the flame extinguishes the temperature
falls rapidly. The extinction point can only be
judged between two steps, always approached
from the flame side of the boundary. After
extinction the bed was relit, the air rate reset,
and the sequence repeated.
A typical set of results is shown in Figure 8
which illustrates plots of temperature
(maximum observed by thermocouple) against
the superficial velocity (calculated utilizing the
maximum temperature) through the bed for
several run sequences. The extinction-
temperature boundary is clearly marked as a
heavy dashed line. The lightly dashed lines
represent constant fuel-air ratio in per-
centages by volume. The continuous line
marked 5.3 percent is the conventional low
limit. It can be seen that a considerable
number of combustion points lie below the low
limit. This is more clearly seen on Figure 9,
showing continued combustion down to 4 per-
cent methane, more than 1 percent below the
normal low limit. Figure 9 also includes the
theoretical adiabatic flame temperature line
with some temperatures exceeding the
adiabatic value. The source of these
unexpected peculiarities lies in the heat
exchanger effect of the fluid-bed particles, as
discussed below.
Bed Profiles
The same super-adiabatic behavior also
occurs in the bed itself. Figure 10 is a typical
temperature traverse down through the bed.
Traverses were made both with sheathed and
bare thermocouples, and a displacement of
the profiles was observed. By using a special
sheathed couple set at right angles to the
holder, it was established that conduction
down the sheath could result in spuriously
high temperatures at a given point. The data
are, in effect, translated by about 1/4-in.
However, this may not entirely account for the
differences. Temperatures exceeding the
theoretical adiabatic by 50 to 150°F have been
recorded in many of the temperature traverses
made.
Clearly, the rate of heat removal must
exceed the rate of reaction at locations above
the temperature peak (in the regions marked
C and D) in order to allow the temperature to
decline. Furthermore, the reaction is probably
totally completed at the location of the
temperature peak. To check this, gas analyses
were taken throughout the bed; Figure 11
illustrates one method of representing the fuel
consumption calculated from the gas analyses.
The graph represents the unburned fuel, on a
log-linear plot, declining with distance up
through the bed. The fuel unburned was back
calculated from the CC*2 analysis; the CO was
never found to exceed 0.75 percent and was
disregarded in the calculation. The two curves
represent two different bed temperatures,
185_0°F (1010°C) and 2350°F (1290°C), with
fuel burn-up easily followed through the bed,
although the times represented are only of the
order of milliseconds. The fuel concentration
in each case decays more or less exponentially
through the bed. Reaction is fast with 90
percent reaction in 16.0 msec and 2.2 msec for
the lower and higher temperature beds,
respectively. These figures are in agreement
with the temperature profiles and with the
prediction that combustion is mostly complete
II-5-8
-------
before the temperature peak is reached, and
all (detectable) combustion is completed
within the bed.
DISCUSSION
The physical behavior of the bed is in good
agreement with expectation from established
theory and requires no further comment. The
combustion behavior, on the other hand,
shows a number of unexpected features.
(1) The lean limit extension and super-
adiabatic temperatures indicated in Figures 8,
9, and 10 were particularly unexpected. They
are, however, simple to explain. They depend
on a heat recovery or heat exchanger effect
due to the bed particles moving up and down-
stream. To understand the behavior, consider
first the effect of a heat exchanger in the
exhaust of an otherwise adiabatic flame
system. With the heat exchanger not
connected, the gas exit temperature of the
flame system is the adiabatic flame tempera-
ture. If the heat exchanger is allowed to heat
the incoming combustion air for the flame
system, the gas exit temperature must be
boosted to adiabatic plus the preheat. Overall,
of course, there is no gain because the extra
heat in the exit gases is removed by the heat
exchanger (exactly, in a no-loss system); the
gases now leave the heat exchanger at the
adiabatic flame temperature. This is, of
course, no more than the usual application of
the heat exchanger although the potential of
heat exchangers for producing super-
adiabatic temperatures is not always
recognized for what it is.
In the case of the fluid bed, the net heat
exchanger effect is clearly evident, but it is not
particularly efficient in this role since the
temperature excess is only 50 to 150°F. The
preheat influence is believed to be most
marked in the early stages of the temperature
rise. In these regions the supply of bed
material must be predominantly from above
(i.e., from hotter zones), whereas further Up in
the bed there can be as much material
supplied from below as above, thus contri-
buting to cooling of the upper zones.
Nevertheless, the overall consequence is to
accelerate the rate of heating, and therefore
the rate of combustion in the bed.
(2) The same preheat effect is responsible
for the extension of combustion below the
usual low limit. Again consider the case of air
preheated to a very high temperature (say, by a
heat exchanger). If the temperature is high
enough, any quantity of fuel, however small,
injected into that air stream cannot be
prevented from reacting completely. Between
this extreme condition and the condition of
the normal lean limit there is a range of rising
temperatures permitting a progressive
lowering of the lean limit to zero. Weinberg,7
for instance, recently quoted a system in which
stable combustion is maintained at a gas
concentration of methane in air of 1 percent;
combustion contributed 250°C and preheat
contributed 1000°C.
Clearly, the heat exchanger effect can be at
least partly responsible for the widening of the
combustion limit found in the fluid bed.
However, the magnitude of the effect — a
drop of over 1 percent in the low limit — does
seem to be rather large for the relatively small
temperature increase over adiabatic, of 50 to
150°F. Some other effect, as discussed below,
may also be involved. This view is supported
by a few measurements using beds of double
the depth. The maximum temperature
increased by about 50 °F, but the limit mixture
at extinction was unaffected. It may also be of
significance that the super-adiabatic tempera-
tures were only obtained at the higher flow
rates, presumably because the heat exchanger
effect was stronger.
(3) Indications that factors, other than
those already mentioned, could be influencing
the reaction were obtained from estimates of
reaction time. Table 1 lists some estimates of
time to complete 90 percent of the reaction
(based on the Figure 11 plots). Included for
comparison are some data from other
sources.2-3'6 The most directly comparable
II-5-9
-------
Table 1. ESTIMATES OF REACTION TIME IN A FLUID BED AND IN CONVENTIONAL FLAME SYSTEMS
tyi
H*
e
Investigator
Present work
Dixon-Lewis
Levy
Van Tiggelen
Van Tiggelen
CH4. %
4.25
5.92
5.03
5.4
Stoichio-
metric
Stoichio-
metric
02, %
20.1
19.88
19.94
20.0
Stoichio-
metric
Stoichio-
metric
Inert, %
75.65 (N2)
74.2 (N2)
75.03 (N2)
74.6 (AR)
73 (AR)
65 (AR)
t
°K
1290
(adiabatic)
1650
(observed)
1530
1528
1950
2110
2370
°F
1865
(adiabatic)
2450
(observed)
2300
2300
3050
3340
3800
V
'ft/ sec
2.7
/
2.2
0.17
1
1.1
2.2
cm/sec
83
74
5.2
31
34
66
T
msec
22a
16a
11
6
2.2
17 + 3
4.9 + 1
0.060
0.015
Total bed
residence time
Temperature peak
90% reaction
Total bed residence
time using Tmgx
Temperature peak
90% reaction
using Tmax
Temperature peak
90% reaction
using T
max
90% reaction
90% reaction
Mean molecular
residence time
Temperature does not show a significant rise until after 8 msec have elapsed.
-------
data are those given by Dixon-Lewis.2 For 90
percent reaction, 17 ± 3 msec are required in a
flat-flame system, while only 2.2 ±0.5 msec are
required in the fluid bed at very close to the
same temperature (2300°F), and only 4 ±1
msec for 99 percent reaction. This factor of 6
or 7 difference is clearly significant.
The most probable explanation that comes
to mind is, of course, enhanced reaction due to
the particles. This could be a result of either
initiating more gas-phase reaction or catalytic
surface reactions.
(4) Some choice between catalytic surface
reactions or enhanced gas-phase reactions
may be possible from further data developed
from the fuel consumption curves of Figure
11. Assuming that the heat from consumption
goes exclusively into the mixture and the
products at the local level, a temperature
profile through the bed was calculated. Figure
12 shows how this compares with the
measured profile. The deviations at the top
end have already been explained as a result of
the preheat effect. The discrepancy at the
lower temperatures, with the predicted
temperatures substantially in excess of
measurement, was totally unexpected.
*
In accounting for the observed
discrepancies a number of explanations were
considered; all but two were discarded. The
simplest explanation is that the gas analyses
may show spuriously high CO2 values because
of continued reaction in the sampling probe.
Against that, however, is the matter of the low
temperatures involved so that any substantial
cooling of the gases would freeze the compo-
sition. The other explanation is more involved
but is also considered more likely. It is based
on the assumption of significant temperature
difference between the particles and gas,
which is quite possible considering the rapid
translation of hot particles into the cooler bed
zones and the short times involved for re-
equilibration. If, therefore, the hot particles
stimulate surface catalytic reaction so that
most of the heat released goes directly into the
particles, the lead in the particle temperature
above the gas temperature will be maintained
until reaction decays. Either bare or sheathed
thermocouples will then take up a tempera-
ture intermediate between the gas and the
particles, but the sheathed couple can be
expected to be more responsive to the particle
temperature because of the enhanced heat
transfer coefficient between particles and a
surface. This would then help to explain the
discrepancy noted above between the bare and
sheathed couples as a factor additional to
conduction as noted.
(5) Finally, a brief examination of the rele-
vance of this information to engineering
applications is in order. The outstanding point
is that reaction in unpremixed systems will
clearly be dominated by the mixing behavior;
the time for reaction can be virtually ignored
unless temperatures are very low indeed.
Clearly, future experiments should include
measurements down to about 1000°F or lower,
which may well be achieved on occasion if a
very wet slurry or sludge is incinerated.
However, on the unit used, controlled
variation of the bed temperature has been very
difficult. There could be some advantage in
reducing the bed size to provide better bed
temperature control. The other aspects of
possible engineering significance are the wider
combustion limits and enhanced reaction rate
(i.e., speed of ignition) due to the hot particles.
These could also increase the risk of serious
explosion of any large bubbles of premixed
fuel and air, if such bubbles are ever permitted
to form. The extinction and relight conditions
are also important; consequently,
development of analytical models for
experimental test is now required to provide a
more reliable basis for extrapolation. (Two
models have been developed, but they are still
too limited in their assumptional basis to be of
much value yet.) Beyond that, what is mainly
needed now for engineering purposes is an
understanding of the behavior of jets and mix-
ing in fluid beds.
In conclusion, therefore, the results
developed iii this paper substantiate the
II-5-11
-------
reasonable expectation that reaction of gas in
fluid beds is fast, and that problems of
incomplete bed reaction must be due to poor
mixing in the bed. In addition, and
unexpectedly, it was also found that reaction
seems to be accelerated by the presence of
particles which also can widen the combustion
limits and generate super-adiabatic tempera-
tures. Other results include the development
of a means of rapid light-off and a
demonstration that the physical behavior of
the bed is in general accordance with expecta-
tion from available theory, in spite of the
simultaneous presence of combustion.
ACKNOWLEDGMENTS
We have pleasure in acknowledging
support for this work from Consolidated
Natural Gas Service Corporation under Grant
No. C-70-29-2. We also gratefully acknow-
ledge assistance in the construction of the
equipment from Messrs. R. Frank, C. Martin,
and D. Simpson. The first author also wishes
to thank Mr. M. Kuwata for assistance he so
ably provided during this work.
LIST OF SYMBOLS
c jf = Initial fuel concentration, by volume
h = Location in bed measured from
bottom, in.
L = Bed thickness, in.
AP = Pressure drop, in. EbO
Tp = Flame temperature
U = Superficial velocity (calculated at bed
maximum temperature), ft/sec
V = Flame velocity
£ = Bed porosity
p = Bed bulk density, lb/ft3
* — Completeness of combustion
a = Particle mass density, lb/ft3
II-5-12
Ts = Gas residence time in the bed, sec
Subscript
o = Incipient fluidization values—also
used to denote fixed bed values
where applicable (AP, e, L).
REFERENCES
1. Davidson, J.F. and D. Harrison. Fluidised
Particles. Cambridge, Cambridge
University Press, 1963.
2. Dixon-Lewis, G. and A. Williams. Some
Observations on the Combustion of
Methane in Premixed Flames. (Presented
at the llth Symposium (International) on
Combustion. 1967. pp 951-958.)
3. Levy, A., J.W- Dredege, JJ. Tighe, and J.F.
Foster. The Inhibition of Lean Methane
Flames. (Presented at 8th Symposium
(International) on Combustion. 1961. pp
524-533.)
4. Spiers, H.M. (ed.). Technical Data on Fuel.
Edinburgh, British National Committee
World Power Conference, 1962. p. 260.
5. Trivedi, R.C. and WJ. Rice.'Effect of Bed
Depth, Air Velocity, and Distributor on
Pressure Drop in an Air Fluidized Bed,
Fluidized Bed Technology, Chemical
Engineering Progress Symposium Series,
American Institute of Chemical Engineers.
62, 1966.
6. Van Tiggelin, A. and J. Deckers. Chain
Branching and Flame Propagation.
(Presented at 6th Symposium
(International) on Combustion. 1957. pp
61-66.)
7. Wemberg. Combustion Temperature: The
Future. Nature. 233:239, 1971.
8. Zabrodsky, S.S. Hydrodynamics and Heat
Transfer in Fluidized Beds. Cambridge.
MIT Press, 1966.
-------
ADDENDUM ON RELIGHT BEHAVIOR
Since this paper was written further data
have been obtained concerning the relight
behavior of the bed. This further information
amplifies the results quoted in the section
titled "Combustion Behavior: Relight."
The method of experiment was as follows.
In the tests described in the section on
"Relight" the bed was first fired up normally
till thermal equilibrium was obtained. Then
the gas flow was decreased to a value below the
lean flammability limit so that the
temperature declined slowly. Combustion was
still occurring in the bed, but the heat release
rate was insufficient to maintain equilibrium.
We also conclude that combustion in the bed
was not quite complete, because the reaction
was evidently continuing on surfaces in the
freeboard area, such as the stainless-steel
sheathed temperature probe. It was observed
that the probe was glowing red when the bed
was black and, therefore, presumably cooler.
When the gas flow was increased (to the level
of stoichiometric gas-air mixture), relight was
always obtained down to 750°F (400°C), as
already reported. The glowing thermocouple
sheath did not originally appear to be
important since the temperatures reported
were presumably those of the thermocouple
(and sheath); the reported temperatures, being
greater than those in the bed, provided a
conservative margin.
These tests were recently repeated, but with
the gas flow turned completely off. The
temperature declined much more rapidly since
there was no combustion occurring in the bed
to retard the temperature decline rate. This
rapid drop made temperature estimation
difficult which was the reason for trying to
control the temperature rate of fall in the
original experiments. There was also no
combustion in the over-bed region to provide
an over-bed ignition source as in the first case.
Relight under these new conditions was not
obtained even at 1400°F (750°C).
These results indicate that in the first case
the hot areas in the over-bed region were
providing the ignition. Thus, for engineering
applications, an over-bed ignition source
should be present for safety reasons.
The relight temperatures obtained under
the second set of conditions are evidently at or
above the values generally quoted for auto-
ignition. The experiments thus underline the
very significant distinction between auto
ignition and reactor ignition. They show very
clearly the safety of the system and ease of
relight down to very low temperatures, even if
the gas concentration is very substantially
below the low limit. Risk occurs only in the
event of total failure of the gas supply, which is
easily guarded against by standard safety
precautions.
II-5-13
-------
IGNITION
BURNER
HP GAS
WALL
THERMO-
COUPLES;
18 TOTAL
TYPEK
WATER COOLED
GAS SAMPLING PROBE
TEMPERATURE PROBE,
TYPE K THERMOCOUPLE
DRAFT GAUGE
FLUE GAS EXIT
12 in.
WATER COOLED WALL
ALUMINA BED (1 in.)
ROCK DISTRIBUTOR
NET (6 in.)
CASTABLE
REFRACTORY
AIR-GAS PREMIXER
AIR INLET
GAS
INLET
Figure 1. Details of experimental apparatus used for these tests.
II-5-14
-------
0.0
!
Q.
LU
V)
£
£
o
«o
0.2
1.0
DIMENSIONLESS BED DEPTH, h/L
Figure 2. Variation of dimensionless pressure drop Ap/Apt with bed depth h/L.
Data taken with water cooled probe in a thin fluidized bed of 1-in. fixed-bed
thickness. Combustion in the bed produced a temperature variation of 1800-2350°F,
and velocities of 3 to 5 times incipient fluidization velocity (2 ft/sec).
II-5-15
-------
o
CM
1.0
0.8
CO
V)
O
ce
u
-------
o
ce.
o
Q-
0.3
HOT SUPERFICIAL VELOCITY, ft/sec
3
Figure 4. Variation of porosity € (see equation 3) and£ / (1 -Q (see equation 2)
with hot superficial velocity. The second group shows correlation between theory
and experiment.
II-5-17
-------
HOT SUPERFICIAL VELOCITY, ft/sec
0.0 2.0 4.0 6.0 8.0 10.0 12.0
1.0
VALUES OF
(1-6)
Figure 5. Variation of the dimension I ess bed thickness L/Lo with both
the superficial velocity U (curve calculated from equation 2 using experimental
values of U and L ) and with 1/(1-G) as substantiation of equation 4.
II-5-18
-------
o
o
2.0
10.0
HOT SUPERFICIAL VELOCITY, ft/sec
Figure 6, Variation of actual velocity with hot superficial velocity u.
The dotted lines indicate what the velocity would be if the bed had a
constant porosity equivalent to fixed bed (€=0.536) or if no bed were
present (€=i).
II-5-19
-------
o>
o
LLJ
a
LlJ
CO
INCIPIENT
FLUIDIZATION
10.0
5.0
10.0
HOT SUPERFICIAL VELOCITY, ft/sec
Figure 7. Variation of bed residence time Ts with superficial velocity.
II-5-20
-------
UJ
ct
ffi
0-
2300
2200
2100
2000
1900
1800
EXTINCTION LINE
6.0
8.0
10.0
HOT SUPERFICIAL VELOCITY, ft/sec
Figure 8. Variation in temperature with hot superficial velocity.
The initial fuel concentration is also indicated. Extinction occurs
at the low temperature and low initial fuel concentration.
II-5-21
-------
COLD SUPERFICIAL VELOCITY
2400) O 1.27 ft/sec
A 1.49 ft/sec
V 1.75 ft/sec
2300 D 2.02 ft/sec
2200
2100
2000
1900
1800
LOW
LIMIT
I
INITIAL FUEL CONCENTRATION, Cif: %
Figure 9. Temperature as function of initial fuel concentration for differing cold
superficial velocities. Also indicated are the adiabatic line and the generally
accepted lower flammability limit.
H-5-22
-------
TIMESCALE,T: msec
16
2000
ADIABATIC
TEMPERATURE
0.4 0.6 0.8 1.0 1.2
HEIGHT ABOVE BOTTOM OF BED, h: in.
Figure 10. Temperature as a function of height in the bed with inlet and adiabatic
temperatures indicated. An auxiliary axis shows the^elapsed time in milliseconds.
II-5-23
-------
o
o
o
ce
ca
0.4 0.6
HEIGHT ABOVE BOTTOM OF BED, h : in.
0.8
1.0
Figure 11. Unhurried fuel concentration (calculated from CC>2 analysis)
as a function of height in the bed for two initial fuel concentrations.
II-5-24
-------
2000
ADIABATIC
TEMPERATURE
TEMPERATURE
FROM C02
ANALYSIS
1.4
HEIGHT ABOVE BOTTOM OF BED, h : in.
Figure 12. Actual temperature measurements and temperature calculated
from the completeness of combustion as a function of height in the bed.
II-5-25
-------
SESSION III:
Gasification/Desulfurization
SESSION CHAIRMAN:
Dr. E. Gorin, Consolidation Coal
III-O-l
-------
1. FLUIDIZED-BED GASIFICATION-PROCESS
AND EQUIPMENT DEVELOPMENT
J. T. STEWART AND E. K. DIEHL
Bituminous Coal Research, Inc.
INTRODUCTION
As part of its broad gas generator research
and development program sponsored by the
Office of Coal Research," U.S. Department of
the Interior, Bituminous Coal Research, Inc.,
(BCR), is developing a multiple fluidized-bed
coal gasification process for the production of
low-Btu fuel gas. The goal of the multiple
fluidized-bed system is the gasification of both
caking and non-caking coal, with fuel gas
being the only product.
Several government and industry co-spon-
sored coal gasification programs are at the
pilot plant stage. These include the IGT Hygas
process, the BCR BI-GAS process, and the
Consolidation Coal Company's COa acceptor
process. Such processes are designed to gener-
ate high-Btu gas, i.e., gas having a heating
value in excess of 900 Btu/scf. Gas of this
quality can be used as a direct substitute for
natural gas. Steam boiler and gas turbine
applications for electrical power generation do
not, however, need this high-Btu gas. More-
over, the optimum fuel gas heating value
required for combined cycle applications can
be of the order of 150 Btu/scf. The primary
purpose of the multiple fluidized-bed coal
gasification system is, then, the production of
low-Btu fuel gas for the generation of electri-
cal energy by means of the combined cycle.
This paper describes the BCR fluidized-
bed gasifier concept and summarizes the work
done by BCR in its development program. The
program began with laboratory scale kinetic
experiments and has progressed through the
semi-continuous operation of a small fluid-
ized-bed batch reactor.
With the aid of an engineering subcontrac-
tor, a process and equipment development
unit (PEDU) has now been designed. The
PEDU, to be located at the BCR Research
Center at Monroeville, Pa., consists of three
fluidized-bed reactors, a gas quenching and
scrubbing system, facilities to preheat reactor
inlet gases, and solids handling equipment.
Designed to gasify 100 Ib coal/hr, 'the PEDU
will operate at 250 psia and at temperatures to
2100°F.
ra-i-i
-------
THREE STAGE FLUIDIZED BED
CONCEPT
The goal of the multiple fluidized-bed
system is the gasification of both caking and
non-caking coal, with fuel gas being the only
product. End use of the gasification product
dictates the operating conditions as well as the
gasifying medium. Thus, gasifying with air
and steam will yield a low-Btu fuel gas;
gasification with oxygen and steam can yield a
higher Btu gas containing a greater proportion
of combustible components; and gasification
with carbon dioxide can yield a carbon
monoxide-rich gas that could serve as a fuel
for power generation by MHD.
A 3-stage system was chosen as one with
the probable minimum number of stages
necessary to meet the requirement of starting
with any rank of coal and producing no tars or
oils as a waste or by-product. Figure 1 is the
design material balance for the proposed
fluidized bed PEDU. Stage 1 receives raw coal
and functions as the pretreatment step. The
devolatilized coal flows by gravity to Stage 2
arid then to Stage 3, which operates as the
final carbon burn-up reactor. Several
pretreatment mediums have been investigated
by others and have been shown to be effective.
These include air alone, and steam or carbon
dioxide diluted with nitrogen and containing
small amounts of air. In this scheme, Stage 3
flue gas is used as the fluidizing medium for
Stage 1.
Stage 2 is the major gasification stage. The
devolatilized coal is gasified with air and
either steam or carbon dioxide to generate the
desired product gas. In addition, Stage 1 flue
gas is fed to Stage 2 where the entrained tars
and oils are gasified. Stage 3 operates at the
highest temperature and serves to maximize
carbon utilization. The ash discharged from
Stage 3 will contain a minimum amount of
carbon. Hot flue gas from Stage 3 flows to
Stage 1 and completes the cycle.
LABORATORY INVESTIGATIONS
The literature abounds with information
regarding the kinetics of carbon/steam and
carbon/carbon dioxide reactions, but there is
often little agreement among the results of
different investigators. For example, reported
activation energies for the carbon/carbon
dioxide reaction range from 25 to 90
kcal/mole. This wide variance may be
attributed to the different types and ranks of
carbon used, the different temperature and
pressure ranges investigated, and the various
simplifications and interpretations of the
observed data. The different results are
indicative of the different reaction
mechanisms and rate controlling steps that
occur under - different experimental
conditions. No correlations were found in the
literature that could adequately describe the
kinetics of the gasification reactions to the
devolatilized coal or "char" that would be
produced in the first step of the multiple
fluidized-bed gasification scheme. Therefore,
laboratory scale kinetic studies were begun.
Six widely different chars were chosen as
the basis for the laboratory studies. The chars
ranged in volatile content from 3 to 12 percent
and were produced from all ranks of coal,
from lignite to a highly caking Pittsburgh
seam coal. Table 1 shows the chemical
analyses of the selected chars.
EXPERIMENTAL EQUIPMENT AND
PROCEDURE
A thermogravimetric balance was used to
obtain the kinetic data. A schematic view of
the TGA is shown in Figure 2.
Char is placed in the sample holder, a
crucible or flat pan, which is connected to and
suspended beneath the transducer coil and a
precision spring. This entire assembly is
mounted inside a quartz and Pyrex
housing. During operation, the sample is
located inside the well of the furnace where
temperatures are continuously monitored by a
chromel-alumel thermocouple.
m-i-2
-------
Table 1. ANALYSIS OF CHARS USED IN REACTIVITY TESTS
Char sample
number
2455
2469
2927
1963
2280
2655
Dry basis, %
Proximate
Volatile
matter
2.85
11.6
2.1
11.3
5.9
4.2
Fixed
carbon
88.9
58.4
74.6
71.4
82.5
83.5
Ash
7.65
30.0
22.5
17.3
11.6
12.3
Ultimate
Carbon
88.7
56.1
73.5
77.1
82.0
83.4
Hydrogen
0.8
1.92
1.35
1.03
1.8
1 .42
Nitrogen
1.3
1.16
0.62
0.49
-
0.9
Sulfur
0.6
9.39
2.89
0.90
0.37
0.76
Ash
7.7
30.0
22.7
17.3
11.6
12.3
Oxygen
0.9
1.43
0.0
4.18
-
1.22
-------
The char is brought to the chosen reaction
temperature in an inert atmosphere. The inert
gas stream is turned off; simultaneously the
reaction gas stream is turned on. Sample
weight loss is then recorded as a function of
time. Data precision is checked and
maintained within ±0.1 percent weight
loss/unit time by duplicating all experimental
runs.
As the test progresses, changes in sample
weight cause an extension or contraction of
the spring which changes the positional rela-
tionship of the armature and transducer coil.
A resulting electrical signal proportional to
the change in sample weight is developed,
amplified, and fed to the vertical (Y) axis of
the recorder. The input to the horizontal (X)
axis of the recorder is proportional to time or
temperature.
MATHEMATICAL MODEL AND
REGRESSION ANALYSIS
The purpose of this kinetic study is to
establish the rate-controlling step and thus
rate equations for the char/steam and
char/carbon dioxide reactions. The study will
also determine the effects of temperature,
reacting gas concentration, and particle size.
The conversion of solids in a heterogeneous
gas-solids reaction can follow one of two
extremes. At one extreme the diffusion of
gaseous reactant into the particle is rapid
enough, compared to the chemical reaction
rate, that the reaction takes place at the same
time and at the same rate everywhere. This is
called the continuous reaction model. If
1
diffusion into the particle is slow, the reaction
is restricted to a thin shell which moves from
the outside of the particle inward. This is the
unreacted core model with diffusion
controlling.
The appropriate model may be chosen by
determining the time needed for complete
conversion of solids of different sizes, as
summarized below:
The parameter tc, defined as the time
needed for complete conversion is:
1. Independent of particle diameter for
the continuous reaction model.
2. Directly proportional to particle dia-
meter for the unreacted core model,
with chemical reaction at the reaction
front as the rate controlling step.
3. Directly proportional to the square of
the particle diameter for the unreacted
core model, with diffusion through the
ash layer as the rate controlling step.
Experiments were carried out with particles
ranging in size from 70 mesh (2\Q\nm) to
minus 325 mesh, (44^m). Within the
temperature ranges investigated, the reaction
rate was independent of particle diameter.
Therefore, the continuous reaction model was
chosen to develop a rate equation, as follows:
the rate of carbon is proportional to the
concentration of reacting gas around particles
times the amount of carbon left unreacted.
In terms of the fraction of carbon reacted, X,
this becomes
dt
(1)
where: Cn is the concentration of reacting gas
to some power, n.
Rearranging equation (1) gives:
Integrating equation (2) yields:
i = -kCnt
or
(3)
The proportionality or rate constant, k, may
be assumed to vary with temperature in an
Arrhenian fashion:
- E
k = ae
RT
(4)
Combining equations (3) and (4) gives:
"
ffl-1-4
-------
Letting X now represent the amount of char
reacted, the complete rate equation becomes:
.E (5)
RT
(1 - X) = Ash + (1 - Ash) e ' ae °nt
where:
X = fraction of char reacted
T = temperature, °K
G = concentration of reacting gas
t = time, minutes
, . . Ib C reacted
k = apparent reactivity, ^ c ^^ min
Ash = weight percent of ash in unreacted char
cal
= activation energy,
mole
cal
R = the gas constant = 1.987 mole OK
n = the order of reaction with respect to
reacting gas
a = the Arrhenius constant (frequen-
cy factor), mhr1
The experimental data are in the form of a
plot of (1 - X) as a function oft. The constants
a, E, and n may then be found from equation
(5) when (1 - X), Ash, T, C, and t are known
variables. The regression analysis proceeds in
two steps. For each value of Ash, C, and T, (1 -
X) versus t is solved for the constant M as
follows.
-W/WI
,n
-E/RT
Let M = ae CJ
Mt
then (1 - X) = Ash + (1 - Ash) e
and (1-X-Ash) ,.,
ln (1-Ash) =Mt
The square of the error in this equality is:
(6)
(7)
where: c- is the error and the subscript i
denotes individual data points. £ is minimized
by differentiating and setting it equal to zero.
= S tjln 1-Ash
M =
I tjlnd-X-AshMnd-Ash) Ztj
(9)
Once M is found from the above equation
for each C and T, a multiple linear regression
approach may be used to find the constants
-E/R, a, and n by solving
M = ae
-E/RT
(Cn)
(10)
or In M = ln(a - E/RT - n)ln C
Again, the partial derivatives of the error
function with respect to M, T, and C are set
equal to zero; the three resulting equations are
solved simultaneously for a, -E/R, and n.
EXPERIMENTAL RESULTS
A series of photomicrographs gave the first
physical clues leading to the selection of a
reaction model. Figures 3 through 6 show
unreacted char, 50 percent reacted char, 80
percent reacted char, and ash. It is
immediately apparent that the average
particle diameter does not change with
increasing carbon burn-off. The particles
become increasingly porous, but even the ash
residue retains a skeletal structure similar in
overall dimensions to the unreacted char.
Along with these physical observations, the
experimental data showed that the time
needed for complete reaction was independent
of particle diameter. Rate equations were thus
developed for the six chars. For example, the
rate equation for the reaction of char No. 2455
with steam is:
(l-Ash)e-k(CH
The significance of this rate equation is
demonstrated in Figures 7 and 8. The
activation energies are of the order of 40
kcal/mole. Both the char-steam and the char-
carbon dioxide were found to be approxi-
mately half order with respect to reacting gas
concentration.
ra-i-5
-------
The char-oxygen reaction was also studied.
Within the temperature range of importance
to this investigation, the reaction was
controlled by mass transfer from the gas phase
to the surface of the particle. Again the
particles do not shrink as the reaction
proceeds. Diffusion of reactants and products
into and out of these small, porous chars is so
fast compared to the chemical reaction step
that the reaction may be thought of as taking
place continuously throughout the particle.
The next step in the laboratory studies was
the construction and operation of a small
fluidized-bed batch reactor. Figure 9 is a
schematic diagram of this system, and Figure
10 shows the actual equipment. The reactor
was made from 1-1/2 in. schedule 40, type 310
stainless steel pipe. External heating was
provided for operation at a maximum reactor
temperature of 2200°F. The system is
designed for operation at pressures to 10 atm.
Table 2. MATERIAL BALANCE FOR FLUIDIZED-
BED BATCH REACTOR AIR-BLOWN GASIFICA-
TION TEST NUMBER 4a
Component
Oxygen
Nitrogen
Carbon dioxide
Carbon monoxide
.Hydrogen
TOTAL
Feed
Mole.%
12.4
46.4
41.2
0
0
100.0
g moles/ min
0.00846
0.03160
0.02824
0
0
0.06830
Product
Mole.%
0.12
30.72
21.45
47.71
0
g moles/min
0.0001
0.0316
0.0221
0.0491
0
0.1029
Feed gas rate: 1 530 sml/min
Product gas rate: 2300 sml/min
Total g moles Carbon
Total g moles Oxygen
Total g moles Nitrogen
In
0.02824
0.03770
0.03160
Out
0.0712
0.0467
0.03160
a Carbon gasification rate, 0.0429 g moles/min
Carbon Dioxide utilization, 21.7 percent
Reactor Pressure, 68 psia
Reactor Temperature, 940°C
Initial Charge to Reactor, 20 g char, BCR Lot 2455
The batch reactor tests had a threefold
purpose:
1. To verify the proposed rate equations.
2. To determine the degree of steam or
carbon dioxide decomposition that could be
achieved in a reactor of reasonable size.
in-i-6
3. To provide physical data such as
minimum fluidizing velocity, attrition rate,
elutriation losses, etc.
Table 2 is a material balance from a typical
batch reactor test. The gasifying medium was
air and carbon dioxide; At the reaction
temperature of 940°C, the gasification rate of
0.0429 g moles/min was consistent with the
predicted value of 0.0519 g moles/min. The
product gas had a gross heating value of
approximately 153 Btu/scf.
The success of the laboratory studies led to
the design of the 100 Ib/hr process and
equipment development unit.
PROCESS AND EQUIPMENT DEVELOP-
MENT UNIT (PEHU)
The purpose of the PEDU is to provide the
necessary parameters for the design and
operation of a pilot scale or larger unit to
demonstrate the process and the economic
feasibility of fluidized-bed gasification for the
commercial production of low-Btu fuel gas.
The PEDU was designed to conform with the
desire of the Office of Coal Research to have a
flexible system designed with a nominal
capacity of 100 Ib coal/hr. Feed to the unit can
be either coal or char with air or oxygen, or a
mixture, as the oxidant and steam and/or
carbon dioxide as the moderator.
Figure 11 is a schematic diagram showing
the major process equipment. Coal is metered
from the pressurized lock hopper through a
rotary air-lock feeder and flows by gravity into
the first reactor. Stage 1, the smallest reactor,
has a reaction zone inside diameter of 10
inches and a disengaging zone of 16 inches.
Stage 2 is the largest reactor with a reaction
zone inside diameter of 16 inches and a 24-in.
disengaging zone. The reaction zone diameter
of Stage 3 is 12 inches with a 16-inch
disengaging zone. All three reactors are
approximately 11 feet high.
Refractory-lined cyclones are provided for
stages 2 and 3 to recycle entrained solids to the
bed. Solids are scrubbed from the product gas
stream in a venturi scrubber, and the gas flows
through iron oxide boxes for hydrogen sulfide
-------
removal and thence to a thermal oxidizer for what extent and in what manner this project
disposal. will continue.
The next step in the development program ACKNOWLEDGMENT
is the construction and operation of the
PEDU. Detail engineering, procurement, and This paper is based on work carried out at
erection will take approximately 12 months. A Bituminous Coal Research, Inc., with support
definitive cost estimate is currently being from the Office of Coal Research, U.S.
prepared for this phase of the program. The Department of the Interior, under Contract
Office of Coal Research will then decide to No. 14-32-0001-1207.
-------
Stream number
Coal (ash free)
Ash
H2
CO
C02
HjO
N2
HzS
Stage 1 oil gas
Total .
Average mole wt
Temperature,0?
Pressure, psia
sclm
aclm
1
Coal leed
Ib
93.8
6.2
100.0
wttf
93.8
S.2
100.0
77
2
Char from
stage 1
Ib
72.E
6.2
78.8
wt*
92.1
7.9
100.0
1200
3
Stage 1
flue gas
moles
0.95
1.53
0.37
0.55
2.48
1.06
6.94
VOl*
13.7
22.0
5.3
7.9
35.7
15.4
100.0
23.3
1200
250
43.8
8.2
4
Stage 3
Hue gas
moles
0.95
1.53
0.37
0.55
2.48
5.88
vol 3
16.2
26.0
6.3
9.4
42.1
100.0
23.9
2100
250
37.1
10.7
5
Product gas
moles
4.7
5.16
1.04
2.00
9.14
0.05
22.1
voIX
21.3
23.4
4.7
9.0
41.4
0.2
100.0
22.3
2000
250
140.0
38.9
6
Air to
stage 2
moles
8.43
29
1000
300
53.0
7.3
7
Char Irom
stage 2
Ib
26.4
6.2
32.6
wt!?
81
19
100
2000
250
8
Char
staj
Ib
3.6
6.2
9.S
from
e3
wt*
37
63
100
2100
9
Air to
stage:
moles
3.14
29
1000
300
19.8
2.7
10
Steam to
stage 3
moles
1.50
18
1000
300
9.5
1.31
11
Steam to
stage 2
molts
2.5
18
000
300
15.8
2.17
PRODUCT GAS
^ 5—1
BASIS COAL FEED
/K/h
TAM I
STAGES
2100 °F
4-10-
STEAM
9— AIR
AIR
Figure 1. Material balance for gasification with air and steam.
IIM-8
-------
SPRING
TRANSDUCER
ARMATURE
TRANSDUCER
COIL
SWEEP GAS INLET
REACTION GAS INLET
INERT GAS INLET
CRUCIBLE
FURNACE
t
DEMODULATOR J
RECORDER
Y • AXIS
X - AXIS
(U.
t
5
GAS
OUTLET
TIME
TEMPERATURE
SWITCH
I
TEMPERATURE
CONTROL
„ SAMPLE
THERMOCOUPLE
Figure 2. Schematic view of thermbgravimetric analysis equipment.
ra-i-9
-------
Figure 3. Photomicrograph of unreacted char.
Ill-1-10
-------
SHREDDED, AIR CLASSIFIED SOLID WASTE FUEL
.
Figure 4. Photographic views of the solid waste storage tank interior.
III-l-ll
-------
Figure 5. Airlock feeder valve.
III-1-12
-------
I I I I I I I I I
1.0 cm
Figure 6. Photomicrograph of char ash (100% burn-off).
III-l-13
-------
i
LU
0
10
20
30
40
50
60
70
90
100
REACTING GAS: 100% STEAM
TEMPERATURE: 1000 "C
SOLID LINE: EXPERIMENTAL DATA
BROKEN LINE : PLOT OF PROPOSED RATE EQUATION
12
15
18
21
TIME, min
Figure 7. Typical correlation of reactivity data.
-------
30
40
50
60
REACTING GAS: 100% STEAM
TEMPERATURE: 900 *C
SOLID LINE: EXPERIMENTAL DATA
BROKEN LINE: PLOT OF PROPOSED RATE EQUATION
TIME, min
Figure 8. Typical correlation of char reactivity data.
-------
FLUIDIZED-BED
BATCH REACTOR
PRESSURE REGULATOR
o
BACK PRESSURE
REGULATOR
TO IN-LINE
CHROMATOGRAPH
Figure 9. Flow scheme for fluidlzed-bed batch reactor
-------
Figure 10. Fluidized-bed batch reactor system.
HI-1-17
-------
TO SULFUR REMOVAL
AND THERMAL OXIDIZER
oo
PRODUCT GAS
i
COOLER'
SCRUBBER
A
COAL
FEED
HOPPER
STAGE 1 OFF GAS
PRODUCT
GAS
COOLER
.1
1
CYCLONE
SEPARATOR
STAGE 3 FLUE GAS
1
»i-J
",r
STAGE 3
r
.^^^
2100°F
\
ASH
LOCK
HOPPER
ASH
F
CYCLONE
SEPARATOR
AIR AND STEAM
BAG
FILTER
CHAR
FINES
HOLD
BIN
\y
Figure 11. Fluidized-bed gasification PEDU process flow diagram.
-------
2. HOT SULFUR REMOVAL FROM
PRODUCER GAS
F. G. SHULTZ AND P. S. LEWIS
Morgantown Energy Research Center
U.S. Bureau of Mines
U. S. Department of the Interior
ABSTRACT
Sulfur-free gas for power generation or catalytic conversion to pipeline gas is needed to meet
near term energy and antipollution requirements. Gasification of coal with air or oxygen and
steam at elevated pressures supplies the gas, but cleaning is required to remove sulfur and
participate matter. A stirred-bed pressurized producer is described, and results are discussed for
caking coals. Progress is reported in developing a process using a regenerable solid sorbent for
removing hydrogen sulfide from hot producer gas with recovery of elemental sulfur formed during
regeneration.
INTRODUCTION
Gasification and gas cleanup must be con-
sidered jointly, in view of today's clean
environment regulations, because their com-
bined action is required if clean gas is to be
obtained from coal. Much of the coal sulfur
appears in the gas; in addition, solid and tar
particulates are present in concentrations that
vary with the gasification process and coal
composition. Gasification concepts under-
going development include gas cleanup in the
overall processing scheme. Innovations are
introduced mainly in the gasification step, and
in some cases desulfurization is incorporated
at this point. In other cases, gas purification
could take place after gasification, and
existing commercial systems may be satisfac-
tory. However, new technology may be needed
to meet more stringent demands.
Probably the least complicated system for
converting coal into either low-Btu fuel gas or
high-Btu pipeline gas is the one described
herein. It bears the suggested name MORGAS
(Morgantown gas). It incorporates pressure
gasification in a stirred bed of mine-run coal,
which may have any free-swelling index from
low to high. Hydrogen sulfide is removed by
contacting the hot gas with a bed of solid sor-
bent containing iron oxide; elemental sulfur is
recovered during regeneration of the sorbent.
These two basic elements can be combined
with other unit operations as required by the
end use of the gas.
in-2-i
-------
EQUIPMENT
The Bureau's stirred-bed producer
resembles the conventional fixed-bed
producer except it uses mechanical deep-bed
agitation. Principal dimensions and layout are
shown in Figure 1. The fuel is supported on a
revolving grate with an area of 9.6 ft2. The
bed depth varies between 6 and 7 feet; the
depth is maintained by frequently adding coal
in batches weighing 200 to 250 lb.; cinder is
removed as needed. The water-cooled stirrer is
balanced by a counter weight and supported
in the pressure vessel by a thrust bushing
sealed by a packing gland. Compound motion
is imparted by combined horizontal rotation
and vertical reciprocation, which can be con-
trolled with respect to speed of rotation and
vertical movement. Figure 2 shows the stirrer
in greater detail. The two lower arms are water
cooled, but the top arm is not cooled as it
normally remains in a reduced temperature
zone. Steady gasification conditions usually
have been obtained by rotating the stirrer at
one-half revolution per minute and limiting
the vertical travel through a vertical distance
of 2 feet. In practice, the stirrer passes through
the bed in 15 minutes, but this rate can be
slower or faster, as optimum rate varies with
coal properties. The lowest point reached by
the stirrer is usually set 2 feet above the top of
the grate, but the limit of travel is within 1 foot
of the grate.
Nuclear density gauges are used as shown
in Figure 3 to indicate conditions within the
pressure vessel. Ash zone, bed level, and voids
in the bed are detected. Control of operating
conditions are simplified by the use of these
instruments; centralized, fully automated
controls seem to be feasible for multiple units.
Continuous stirring of the bed maintained
a dense fuel bed, giving good quality gas
having constant composition. Vertical
movement was as important as rotation for
operating the experimental producer, but
vertical movement may not be necessary for
full-size units. Stirring was needed to break
large clinkers that formed in the combustion
zone, as well as coke in the gasification zone.
As shown in Figure 4, the torque applied to
rotate the stirrer varies directly with the bed
depth covering the stirrer. At maximum depth
the normal torque was 1,300 foot-pounds,
reaching momentary peaks of 1,700 foot-
pounds. Measurements were obtained on
Upper Freeport coal, which gives a very hard
coke. No significant difference in the torque
load was found for double-screened 1/4- x 1-
1/2-in. Upper Freeport and run-of-mine 0- x
1-1/2 in. Gasifying mine-run coal in a stirred
bed was a significant advancement because
the size limitation, heretofore believed
necessary, can be eliminated. More of the
market supply will be available for gasifica-
tion, and preparation will be less costly. A
screen analysis of run-of-mine Upper Freeport
coal is given in Table 1. Twenty-five percent of
the sample passed through a 1/16-in. sieve
and 5 percent through 100 mesh. Some fine
coal particles were entrained in the gas, but
most were removed by a cyclone separator.
Gas vented to the atmosphere and burned had
a dust loading of about 0.5 to 0.7 lb/1000 ft3.
Table 1. SCREEN ANALYSIS, UPPER FREEPORT
COALa
Screen size
2-1/2x2 in.
2x1 in.
1 x 1/2 in.
1/2x1/4 in.
1/4x 1/16 in.
1/16 in. x 50 mes
50 x 100 mesh
100 x 200 mesh
200 mesh x 0 in.
Analysis, %
Direct
2.5
12.1
12.2
17.6'
30.2
h 16.9
3.5
2.1
2.9
Cumulative .
2.5
14.6
26.8
44.4
74.6
91.5
95.0
97.1
100.0
Free swelling index No. 8-1/2
RESULTS
Experimentally determined gas yields for
moderately caking Illinois No. 6 coal are
shown in Figure 5. Gas production was limited
when the gas flow reached a velocity at which
loss of fuel by entrainment becomes excessive.
m-2-2
-------
This plot shows that the quantity of air
limiting gas yield increases with increased
pressure.
Table 2. SORPTION OF H2S FROM DRY
PRODUCER GAS BY SINTERED IRON OXIDE-
FLY ASH8
A mixture of iron oxide (hematite
and fly ash was the best sorbent found among
more than twenty materials tested. Primary
requirements were that the sorbent be readily
available and relatively inexpensive, have
reasonable sorption capacity and useful life,
be easily regenerated for repeated use, and be
resistent to fusion or disintegration over the
useful temperature range. Fly ash as received
could be formed into a durable and
regenerably sorbent, but its sorption capacity
was improved by adding iron oxide, increasing
the concentration to 36 from 15 percent
originally present. Other oxides present and
inactive included silica 35 percent, alumina 18
percent, and small percentages of oxides of
calcium, magnesium, sodium, potassium, and
titanium. Iron oxide concentrations greater
than 40 percent were unsatisfactory because
the bed fusion temperature was lowered and
fusion took place during normal operations.
Pilot quantities of the fly ash-iron oxide
sorbent were made by two catalyst manu-
facturers by mulling and extruding the
mixture to form 1/4-in. diameter cylinders
with 1/4 to 1/2-in. lengths, which were then
sintered to develop hardness. Mercury
porosimeter measurements showed pore
volume of new sorbent was 0.36 cm3/g, but
this decreased to 0.13 cm3/g and remained
constant after 30 regenerations, as shown in
Figure 6. Surface area measured by nitrogen
absorption ranged from 4.2 to 6.5 m2/g. Sorp-
tion of hydrogen sulfide from dry simulated
producer gas is given in Table 2 for materials
of essentially the same composition but made
by three laboratories.
The sorbent made by MERC was tested
through 174 regeneration cycles using
simulated producer gas and bed temperatures
of 1100, 1250, and 1500 °F. Producer gas
contains about 5 to 10 percent steam by
volume, as excess steam is used to reduce
temperature in the combustion zone, and the
Commercial
laboratory 1
Commercial
laboratory 2
Mechanically
formed by
MERCC
Surface
area,b
m2/gram
6.5
4.2
4.2
Bed
temperature,
°F
1000
1250
1500
1000
1250
1500
1000
1500
g S removed/100 g sorbent
From sorp-
tion data
12.5
14.7
22.2
7.5
11.5
22.5
10.5
27.6
From regen-
eration data
12.4
13.9
22.0
6.7
11.0
17.4
10.9
25.6
aAII at 3 psig sorption pressure.
bBET nitrogen sorption method.
cMorgantown Energy Research Center, Morgantown, W. Va.
gas leaves the generator at a temperature
around 1200°F. Steam amounting to 7 percent
by volume was added to the gas for many of
the above tests to closely simulate producer
gas. Results obtained with gas containing
steam, Figure 7, indicate a reduction in
capacity when compared with capacities for
dry gas as shown in Table 2. This was
attributed to the lowering of the hydrogen
sulfide concentration at the gas-solid interface
by the added steam. Improving the mass
transfer coefficient by raising the bed
temperature was effective in increasing the
capacity from 6 g sulfur/100 g sorbent at
1100°F to 10 g at 1500°F.
Iron oxide catalyzes the water gas shift
reaction, HaO + CO = fib + CO2, and steam
in producer gas affected the composition of
the producer gas in passing through the
sorption bed. The composition change
resulting from the shift reaction was
determined at 300 psig and temperatures of
1100, 1300, and 1400° F by passing producer
gas containing 18 mole-percent steam through
a bed of iron oxide fly ash sorbent using 1000
space velocity. Heating value was decreased by
dilution from the carbon dioxide added to the
gas; increased hydrogen and decreased carbon
monoxide concentrations resulted in virtually
no net change in heating value because they
have nearly the same value, 319 and 316
Btu/scf, respectively. The shift would be
in-2-3
-------
beneficial if pipeline gas is the end use,
because additional shifting would be needed
to bring the hydrogen to carbon monoxide
ratio close to 3:1. Increasing temperature
favors higher carbon monoxide concentration
at equilibrium. Results are shown in Table 3.
Table3. CHANGE IN COMPOSITION AT 300 psig
AND 1000 SPACE VELOCITY
Feed gas
Effluent gas:
1100°F
(dry)
1100°F
lwet)a
1300°F
(wet)a
1400°F
(wet)8
Hydrogen,
%
17.0
16.4
22.9
21.3
19.9
Carbon
Dioxide.
%
6.7
7.9
13.2
11.0
9.7
Nitrogen,
%
50.9
51.0
51.1
51.0
51.0
Carbon
Monoxide,
%
25.4
24.7
12.8
16.7
19.4
Healing
Value.
Btu/scf
134
130
113
121
125
aSteam content 18 volume-percent. Composition and heating
value on dry basis.
Two sorption-regeneration cycles were
completed, and cleaning gas was generated in
the pressurized gas producer using Upper
Freeport coal; the results are shown in Figure
8. Gas from the producer was transferred to
the sorbent bed at system pressure of 120 psig
via a heated pipeline. Bed temperatures were
controlled to give 1100 and 1200°F, and flow
rates were adjusted to give hourly space
velocities of 710 and 940, respectively. Hydro-
gen sulfide concentration averaged 380 gr/100
ft3; the gas contained approximately 0.516
dust, 1 Ib tar, and 5 Ib steam/1000 ft3
Hydrogen • sulfide in the gas leaving the
sorbent bed had its concentration reduced to
10 and 20 gr/100 ft3 and did not increase until
after 6 hours on steam. Removal was 95 and
97 percent effective with respect to hydrogen
sulfide. Tar was not removed by the sorbent.
Reaction mechanism is chemisorption,
whereby hydrogen sulfide difusses throughout
the sorbent and reacts with Fe2O3 forming
FeS and FeS2. Analyzing the spent sorbent
indicated the empirical composition was
FeS1-3. Iron oxide, Fe2O3, was regenerated
and the sulfur released as SO 2'by passing air
or oxygen over the hot bed. With oxygen
regeneration, the; effluent gas was pure SO2
until some oxygen passed through unreacted
after regeneration was 90 percent complete.
Rather than recovering the SO.2 as sulfuric
acid or ammonium sulfate, it appears possible
to reduce SO^2 to elemental1 sulfur. This may
be done by regenerating two beds of saturated
sorbent at the same time. Oxygen or air is
supplied to one ;;bed where the sulfur is
oxidized,
and the SO2-rich effluent, free of oxygen, is
supplied to the second bed where oxidation-
reduction at 1500°F gives elemental sulfur,
3 FeS + 2 SO2 = Fe3 O4 + 5 S. (2)
Before returning the second bed to sorption
duty, magnetite is oxidized to hematite, as
follows:
CONCLUSION
Results indicate that hydrogen sulfide can
be removed from producer gas by chemisorp-
tion using sintered pellets of iron oxide-fly ash.
Long life is indicated for the sorbent used in a
fixed bed. Fluidized- or expanded-bed
operation may be possible if the pellets are
.reduced in size and shaped as spheres. The gas
is generated and cleaned at pressure and
temperature, thus conserving space and
energy expended in gas compression. The
initial results indicate that elemental sulfur
may be recovered.
ffl-2-4
-------
AGITATOR DRIVE
BED LEVEL
GRATE DRIVE
27 ft - 9 in.
STEAM
RUPTURE DISK
Figure 1. Schematic drawing of gas producer.
HI-2-5
-------
WATER OUT-*
ROTATION, time per revolution
7 min, 13 sec SLOWEST
1 min, 30 sec FASTEST
VERTICAL TRAVEL, ft/hr
1.9 SLOWEST
9.4 FASTEST
-1/4-1..—| •
•c pipe //
M/4-
CS PIPE
TUBING
3-in. OD
1-in. ID
WATER IN
PACKING GLAND
WATER IN
PRODUCER SHELL
cm
TUBING -
6-1/4 -in. OD
3-1/2-in. ID
BEARING ASSEMBLY
HIGHEST
OSITION
BOTTOM SURFACE
6 ft-4 in.
SECTION A-A
1
1
l_
01
14
J 1-
nr
23 in.
ilr i
1 71 »
in.
.J LOWEST POSITION i
12 in.
t
TOP OF GRATE
Figure 2. Stirrer arrangement.
III-2-6
-------
STIRRER
COAL FEED
COAL FEED
1
Bfc.^M^—^— ... .1-. n. i....i „ ii..i. i —••»••»
Figure 3. Nuclear density gauges applied to gas producer.
III-2-7
-------
i'
o
10
20 30
TIME, min
Figure 4. Torque applied to stirrer.
50
60
160
§
o
S 140
120
ILLINOIS NO. 6 COAL
(H.V. B BITUMINOUS)
3025
2350 LIMITING AIR FLOW, Ib/hr
20
40
60
80
PRESSURE, psig
Figure 5. Pressure raises limiting air flow.
100
I1I-2-8
-------
o
111
Q£
s.
12
10
s
m
o
CO
O
OC
I 4
60 90
NUMBER OF REGENERATIONS
Figure 6. Pore volume reaches constant value after 30 regenerations.
A.
C.
BED TEMPERATURE
A. 1500 °F
B. 1250 °F
C.1100°F
50
100
150
200
NUMBER OF REGENERATIONS
Figure 7. Sulfur sorpfion increases with bed temperature.
m-2-9
-------
400
300
"S:
exi
(A) INLET CONCENTRATION
(B) OUTLET CONCENTRATION
SPACE VELOCITY, 940
BED TEMPERATURE, 1200 ° F
(C) OUTLET CONCENTRATION
SPACE VELOCITY, 710
BED TEMPERATURE, 1100 °F
LU
O
I
200
100
TIME, hr
Figure 8. Removing H2$ from Producer gas.
HI-2-10
-------
3. COAL GASIFICATION FOR CLEAN
POWER GENERATION
D. H. ARCHER, E. J. VIDT, D. L. KEAIRNS,
J. P. MORRIS AND J. L. CHEN
Westinghouse Research Laboratories
ABSTRACT
The growing demand for electrical energy in the U.S. requires the construction of new coal-fired
power plants. Coal gasification, coupled with combined gas and steam turbine generation,
provides a basis for a low cost, high efficiency, non-polluting plant. A fluidized-bed coal
gasification process adapted to power generation has been devised. It uses air and steam for
gasification and limestone or dolomite sorbent for desulfurization. A development effort is
underway which includes the construction of a 1200 Ib/hr coal gasifier and the performance of a
supporting laboratory program.
INTRODUCTION
Need for Power Generation
In the next 20 years the quantity of
electrical energy generated in the U.S. is
expected to increase by a factor of almost 4, as
shown in Figure I.1 Efforts to reduce the rate
of growth in demand for electrical energy have
been proposed. On the other hand it has been
suggested that additional quantities of
electrical energy will be required as programs
are carried out to improve the environment
and to maximize the efficiency of energy
utilization. It seems prudent, therefore, to
determine how the projected demands can
best be met.
Both nuclear and fossil fuels—coal, oil,
and gas—will be needed to supply this
demand. Projections of fuel usage are made in
Figure 2.1 Nuclear fuel usage is limited by the
number of nuclear power plants which can
conceivably be constructed in coming decades.
Natural gas shortages in the U.S. have led to
the prediction that its use by utilities will be
severely curtailed in the coming decade. Coal
and oil, therefore, must provide the difference
between total fuel demand for electrical
generation and that portion supplied by
nuclear fuel. An upper limit may be placed on
imported oil to avoid problems resulting from
dependence on foreign nations and from
unbalance of payments in foreign trade. If so,
the use of coal in power generation must
increase by a factor of 3 in the next two
decades. But if coal is not available in
sufficient quantities to meet the demand, the
use of oil in power generation will of necessity
continue to increase.
Additional power plants must be
constructed to meet the demands for electrical
energy, as illustrated in Figure 3.1 The
generating capacity must increase from 325
GW in 1970 to 1400 GW in 1990. Fossil fuels
will be used primarily in intermediate and
peak load plants. Intermediate plants vary
their output during each day to match the
varying demand; their overall electrical energy
in-3-i
-------
output is about 40 to 60 percent of the seasonally or in emergencies to match the
maximum (if the plant operated continuously varying demands; their energy output is 2 to
at rated capacity). Peaking plants operate only 20 percent of the maximum.
m-3-2
-------
Criteria for Power Plant Concepts
In order to compete successfully with a
conventional coal burning steam power plant
with a stack gas scrubber for SO 2 removal, an
improved power generation system should
have lower capital costs, higher operating
efficiencies, and pollutant emissions which
meet established requirements. Targets have,
therefore, been established for the overall
economics and performance of a new power
plant concept:2
1. Capital costs for 250 to 600 MW plants
operating in 1975, less than $330/kW
2. Overall operating efficiencies greater than
39 percent
3. Sulfur dioxide emission less than 1.2, NOx
emission less than 0.7, and particulate
emission less than 0.1 lb/106 Btu heat of
combustion of the fuel; thermal
discharges prevented by the use of cooling
towers.
Proposed Power Plant Concept
An efficient, economic power plant burning
coal and providing low cost electrical energy
for intermediate or base loads can be provided
by coupling a coal gasification and gas
cleaning system with a combined gas- and
steam-turbine generation plant as shown in
Figure 4. A two-stage fluidized-bed process
gasifies coal using air and steam at tempera-
tures of 1400 to 2100°F and pressures of 10 to
20 atm. The process desulfurizes the fuel gases
at high temperature, 1400 to 1800°F, imng a
limestone or dolomite sorbent. The resulting
CaS is treated for disposal; or the sorbent is
regenerated for return to the process, and
sulfur is recovered. Particles are removed from
the hot fuel gases by cyclones, pebble bed
filters, or porous ceramic filters. Most of the
fuel gases, 80 to 90 percent, flow to gas turbine
combustors where they burn with excess air to
provide hot gases for expansion in a gas
turbine. The remaining 10 to 20 percent of the
fuel gases flow to a heat recovery boiler which
provides steam at 1200 psig and 950 °F to the
steam turbine. About half the electrical energy
output from the plant is produced by the gas
turbine generator; and half by the steam
turbine generator. The gas turbine also drives
the main compressor for air flowing to both
the combustor and the gasifier. A booster
compressor for the air flowing to the gasifier is
used to overcome pressure losses in the
gasification and gas cleaning systems.
Similar combinations of coal gasification,
gas cleaning, and combined gas and steam
turbine generation have been proposed and
explored.3-4 These differ in the type of gasifier
proposed, in the design of the gas cleaning
system, and in the configuration of the power
generation system. For example, STEAG5 has
built a plant employing fixed-bed coal
gasifiers, low temperature aqueous scrubbers
(for particulate removal — desulfurization is
not included in the plant), pressurized boilers
(for combustion of the fuel gases), and a gas
turbine expander.
GASIFICATION FOR POWER PRODUC-
TION
General Background
Coal gasification for power production
produces a clean fuel gas — with minimal
sulfur and ash — which can be utilized either
at atmospheric pressure in a conventional gas-
fired boiler or at elevated pressure in a gas
turbine combustor. Gas turbine generator or
combined gas and steam turbine generator
plants are preferred for new installations,
because estimated costs for such plants are
appreciably lower than those for conventional
steam power plants. Gasification may also be
useful in the future preparation of a clean fuel
for an MHD6 or fuel cell power plant.7
Tn gasification, air (or oxygen) is supplied
to fuel in a quantity insufficient to complete
the conversion of its carbon and hydrogen to
COa and HaO. A number of possible
sequential processes become important. Some
in-3-3
-------
of the oxygen added to the fuel reacts to form
COz and H2O:
O,
CO,
(1)
H2+1/2O2 -> H2O
(oxidation)
These reactions release large quantities of
heat. But unburned carbon from the fuel re-
mains, and it reacts with CO2 and H2O to
form CO and H2:
C02 CO
(gasification) '2'
These reactions absorb large quantities of
heat. Hydrogen can also react with carbon
from the fuel to form methane:
C + 2H.
CH,
(hydrogasification) (3)
This reaction is moderately exothermic.
Finally the fuel, when heated, can also
undergo
Fuel + heat •* C + CH4 + HC
(devolatilization)
where HC indicates higher hydrocarbons and
tars. This reaction may also yield heat.
In a gasification process all of these pro-
cesses can occur simultaneously throughout a
reactor, or each reaction may be localized in a
region of a reactor or in a separate vessel.
Most gasification processes, however, are
carried out so that the heat released by oxi-
dation, hydrogasification, and devolatilization
balances the heat required by gasification and
the sensible heat of the overall reaction
products. This overall heat balance can be
achieved by controlling the amount of air (or
oxygen), the amount of steam, or the amount
of an inert gas added to the gasifier. If the
reactions are carried out in separate regions or
reactors, some mechanism is required to
transfer heat between these regions.
The gas composition produced by a gasifi-
cation process depends primarily on the
nature of the fuel and on the temperature,
pressure, and gas composition in the regions
where gasification, hydrogasification, and
devolatilization occur. These quantities
determine the kinetic rates and the ther-
modynamic limits of the various processes —
oxidation, gasification, hydrogasification, and
devolatilization.
When H2O, CO, H2, and CO2 coexist at
high temperatures, they can also undergo the
shift reaction:
C0
- H20
C0.
(5)
This reaction has a negligible heat effect, but
its equilibrium does affect the gas composition
according to the relative quantities of the
gases involved in the shift.
In gasification, sulfur in the fuel is
converted primarily to H2S. A
limestone/dolomite sorbent can be utilized in
a fluidized bed to remove this pollutant from
the fuel gases:
CaCO3
CaO
H2S
CO,
CaS + H7O+ _2
Over the past ten years much effort has
been expended in the United States to develop
processes to produce pipeline gas, consisting
primarily of methane. Recently, interest in
gasification processes to produce a fuel for
power plants has increased. There are
important distinctions between the gasified
coal properties required for pipeline gas and
those required for power plant fuel. These
distinctions are both technical and economic;
they have an important effect on the nature of
the optimum gasification process for each of
these applications.
In general, pipeline gas processes employ
either pure O2 or H2 together, with H2O at
pressures well in excess of 20 atm to produce a
produ^ high in CH4. Fuel gas for power
processes uses air and H2O at 20 atm or below
to produce a lower-cost product. Table 1
summarizes the differences in the
characteristic properties of a pipeline gas and
a power plant fuel produced by a gasification
process.
in-3-4
-------
Table 1. FUEL GAS PROPERTIES REQUIRED FOR PIPELINE GAS AND FOR COMBINED
CYCLE POWER PLANT FUEL
Heat content, Btu/Sft3
Pressure, atm
Temperature,°F
Composition
Cleanliness
Sulfur
Particulates
Fuel cost target, $/106 Btu
Pipeline gas
~1000
>60
~70
Primarily CH4
<1ppmb
« 0.01 lb/106 Btu b
0.50-1.00
Power plant fuel
> 90
10-20
70-1800
CO, H2,N2,C02, H20,CH4
1.2 Ib SC-2/106 Btu (H,550 ppm)
0.1 lb/106 Btu
0.25-0.50
A high temperature is advantageous and may be necessary if the heating value of
the gas is low.
b Limits established by process requirements.
Fuel is currently processed in three reactor
types — fixed bed, suspended bed, and
fluidized bed. In a fixed-bed reactor, gases
pass through a bed of solids at a velocity
sufficiently low that the solid particles are not
blown from the bed and are not supported by
the flowing gases. The weight of the particles
rests primarily on other particles which make
up the bed. A boiler with a chain grate stoker
is one type of fixed-bed reactor.
In a suspended-bed reactor, gases flow at a
sufficiently high velocity that solid particles
are carried along with the gases; their weight
is supported by drag forces exerted by the
gases. Contact between particles is limited to
occasional collisions. A pulverized fuel boiler
is one type of suspended-bed reactor.
In a fluidized-bed reactor, the gases flow
through a bed of particles at a sufficiently high
velocity to support their weight but not high
enough to carry them out of the bed.
Fluidized-bed gasification reactors have not
yet been applied commercially to utility power
generation. But at least five fluidized-bed
gasification reactors are currently under
development to produce pipeline gas and/or
liquid fuels from coal. Other fluidized-bed
gasification reactors are currently under
development to produce pipeline gas from oil;
fluidized-bed reactors are now used com-
mercially in the catalytic cracking of oil,
roasting of sulfide ores, incineration of oily
wastes and sludges, production of organic
chemical monomers, making of cement,
conversion of nuclear materials for fuel
elements, etc.
Fluidized-bed reactors provide the
following features in processing solids and
gases:
1. Ease and versatility in solids flowing and
handling. Solid materials can readily be
added to or removed from fluidized-beds.
Gas velocities can be chosen to promote
particles mixing in the bed or to cause
separation between particles of different
size and density.
2. Rapid heat transfer. The free movement of
particles in a fluidized bed promotes rapid
heat transfer both within the bed and
between beds. Bed temperatures are
therefore uniform and easy to control.
Heat can be transferred between beds by
the circulation of solids.
3. Effective gas-solid contact. Because the
relative velocity between gas and solids is
high, exchange of mass and heat is rapid.
A fluidized bed also provides a large
amount of solid surface in contact with
flowing gas in a relatively small volume.
Table 2 summarizes various reactor types
for coal gasification — their applications,
advantages, and problem areas.
in-3-s
-------
Table 2. REACTOR TYPES FOR COAL GASIFICATION
Reactor type
Application to
gasification
Advantages
Problems
Fixed bed
Suspended bed
Ruidized bed
Lurgi, McDowell Wellman
gasifiers.
Texaco partial oxidation,
BCR two-stage gasifier.
Consol gasoline and ac-
ceptor, IGT hydrogasifica-
tion, FMC gasification,
Bureau of Mines synthane
gasifiers.
Developed technology;
countercurrent flow of gas
and solids possible.
High temperatures do not
lead to excessive agglom-
eration of coal or ash.
Versatility and ease of
solids handling, uniform
temperature; effective gas-
solids contact.
Maintaining uniform gas
and solids flow, adding
coal, removing ash, tem-
perature control, transfer of
heat.
Separating ash solids from
gases, temperature control;
co-current flow.
Pretreatment to prevent
coal agglomeration; multi-
stage beds required to
achieve counter-current
flow.
The process of gasifying coal involves a
number of process steps including:
Drying - The water content of the coal is
reduced so that the particles are free-
flowing and more readily transported and
introduced into the gasification
equipment.
Pretreatment - The coal is oxidized and/or
devolatilized superficially in order to
prevent sticking and agglomeration of
particles.
Desulfurization - The sulfur released from
the coal as H^S during the gasification
process is sorbed by limestone/dolomite
particles of the bed.
Hydrogasification and devolatilization -
Volatile products are driven off the coal in
an atmosphere containing hydrogen,
which reacts with the coal and char
forming methane and higher hydro-
carbons and releasing heat.
Gasification - Steam reacts with the char
(or devolatilized coal) absorbing heat while
forming fuel gases — Ha and CO.
Combustion - Air reacts with the carbon of
the char forming combustion products
and releasing heat.
The rate and extent of each of these
processes depends on temperature, pressure,
atmospheric composition, and time. It
appears advantageous and probably necessary
to perform groups of these processes in
individual reactors or in individual reaction
steps within a single reactor; in this case
provision must be made for the flow of
reactants and heat between reactors or
regions.
Proposed Coal Gasification Process for
Electric Power Generation
A proposed improved multi-stage
fluidized-bed coal gasification for power
production process concept is illustrated in
Figure 5. It comprises three process units — a
dryer, a recirculating bed devolatilizer-de-
sulfurizer, and a fluidized bed gasifier-
combustor.
Crushed coal is dried in a fluidized bed and
transported to the devolatilizer-desulfurizer
unit. Here the devolatilization, desulfuri-
zation, and partial hydrogasific'ation
m-s-6
-------
functions are combined in a single recir-
culating fluidized-bed reactor operating at
1300 to 1700°F. Dried coal is fed into a central
draft tube of this reactor. In this tube, the coal
feed and large quantities of recycled solids —
char and/or lime sorbent — are carried
upward by gases from the total gasifler flowing
at velocities greater than 15 ft/sec. The recycle
solids needed to dilute the coal feed and to
temper the hot inlet gases flow downward in a
downcomer — a fluidized bed surrounding the
draft tube. These solids, flowing at rates up to
100 times the coal feed rate, effectively prevent
or control agglomeration of the coal feed as it
devolatilizes and passes through the plastic
and sticky phase. A dense dry char is collected
in the fluidized bed at the top of the draft
tube. Lime sorbent is added to this bed in
order to remove sulfur which is present as FhS
in the fuel gases. (An alternative concept
would employ a separate desulfurization
process. The fuel gases could be cleaned at
high temperature in a fluidized bed of lime
sorbent; or they could be cleaned, after
cooling, at low temperature in a scrubber.)
Spent (sulfided) sorbent is withdrawn from the
reactor after stripping out the char either in
the transfer line or in a separator of special
design. Char is withdrawn from the top section
of the bed. Heat is primarily supplied to this
unit from the high temperature fuel gas
produced in the total gasifier. Additional heat
can be transported to the devolatilizer by
solids carry-over in the gases from the total
gasifier or by solids exchange between the two
process units. Alternatively, additional heat
can be generated in the devolatilizer by
supplying air to the downcomers surrounding
the draft tube.
The final gasification of the low sulfur char
is conducted in a fluidized bed with a lower leg
which serves as a combustor. In this section,
char obtained from the devolatilizer-
desulfurizer is burned with air at ~2100°F to
provide the gasification heat. Heat is
transported from the combustor to the gasifier
both by combustion gases and by solid? which
flow up and down between the combustor and
gasifier. The ash at this temperature agglome-
rates and segregates in the lower bed leg for
removal. Gasification occurs in the upper
section of the bed at 1800-2000° F with the
sensible heats of both gas and char being used
through solids exchange, to provide the heat
requirements for the devolatilizer-desul-
furizer.
This concept has the potential for
overcoming the limitations of other gasifica-
tion processes and providing a lower cost
gasification system.
Utilization of wide variation in fuels —
Caking coals and high ash coals can be
used without costly and inefficient
pretreatment. This feature is achieved by
employing the recirculating bed to prevent
agglomeration of coal particles.
Utilization of a wide variation in coal size -
The sizing of the coal to the system is not
critical. Coal with a size range of 1/8- to
1/4-inch x 0 can be used in the fluidized-
bed system.
High thermal efficiency - Good heat
economy is realized through the counter-
current movement of gases and solids
between stages. The multi-stage arrange-
ment provides the long residence time
required for high carbon conversions, with
good control over the temperature in both
stages of gasification. Fluidized-bed gasi-
fication systems also provide a means for
minimizing high carbon ash leaving the
bed. This is achieved in the proposed
design by using the agglomerating bed.
Fluidized-bed agglomeration of coal ash
with low carbon loss «1 to 2 percent) has
been demonstrated on both small and
large scale equipment.
Reduced heat losses - Clean fuel gas can be
produced without the heat loss as
occasioned in cooling the fuel gases. The
fluidized-bed concept permits fuel desul-
furization by limestone or dolomite at
elevated temperature.
in-3-7
-------
Although this advanced gasifier concept is
unique, it is composed of sub-systems which
have been successfully operated by others. For
instance, the recirculating bed has been
developed by the Gas Council in England 8"13
and utilized by others.14 The desulfurization
step employing fluidized char and dolomite
has been investigated by Consol Coal,15 who
embodied this idea predominantly for
producing a low sulfur char; and by Esso
(UK),16 who uses fluidized beds of lime for
gasifying and desulfurizing oil. Similarly,
FMC employs multiple fluidized stages to
produce char22 Fuller, Chicago Bridge,
Battelle,19 and others 1?,18,20,21 have used
agglomerating fluidized-bed combustors in
their processes. The use of multiple fluidized
stages, with countercurrent flow of product
gas, to achieve total gasification with desul-
furization is a logical but novel method for
utilizing all the inherent advantages of all
these systems.
DEVELOPMENT PROGRAM
To realize this gasification concept and to
achieve its potential benefits in power
generation a development program is
currently underway. It involves three parallel
efforts:
1. The design, construction, and operation of
a process development plant for gasifying
1200 Ib coal/hr and for cleaning the
resulting gases.
2. The planning and pursuit of laboratory
studies in fluidization; H2S sorption and
lime regeneration; coal devolatilization,
char gasification, and ash aggolmeration.
3. The conduct of systems studies on overall
power plant performance and economics.
Process Development Plant Design
The purpose of the process development
plant is to provide a means for investigating
the gasification system — the devolatilizer-de-
sulfurizer and the gasifier-combustor, both
individually and in combination. The investi-
gation is to:
1. Establish the operability of the equipment
over a suitably wide range of conditions —
flow rates, pressures, temperatures, and
types of coal and lime sorbent;
2. Verify the suitability of the fuel gas
produced for power production in a
combined gas and steam turbine
generator plant;
3. Produce the data required for engineering
scale up and economic evaluation of the
gasification equipment for a power plant.
Preliminary plans for the plant have been
completed. These plans include flow diagrams
(Figure 6) material and heat balances, and
dimensional sketches for the devolatilizer-de-
sulfurizer and the gasifier combustor. Special
propane burners are included with the devola-
tilizer-desulfurizer to supply hot reducing
gases so that this unit can be operated
independently of the gasifier-combustor. If a
supply of char is provided, the gasifier-
combustor can also be operated independently
of the devolatilizer-desulfurizer.
In addition to the devolatilizer-desulfurizer
and the gasifier-combustor reactors, the
process development plant includes:
1. Coal, char, and, limestone sorbent
receiving, storage, and feed systems;
2. Ash and spent sorbent removal, treating,
and discharge systems;
3. Primary cyclones for removing particles
from gases leaving the reactors; and
4. Gas scrubbing and quench systems for
cleaning fuel gases prior to incineration.
Provisions have also been made for the
addition of other features to the development
plant including:
ffl-3-8
-------
1. A secondary cyclone and filter to clean fuel
gases leaving the primary cyclone of the
devolatilizer-desulfurizer. These gases
must be sufficiently clean to pass through
a gas turbine with minimal corrosion,
erosion, or deposition.
2. A gas turbine combustor to burn the clean
fuel gases efficiently with a minimum
production of NO.
3. A turbine blade test unit to demonstrate
that the combustion gases have been
effectively cleaned.
4. A lime sorbent regenerator to convert the
spent (or sulfided) sorbent back to a form
(a carbonate or oxide) which will absorb
additional sulfur.
Operating conditions have been selected
for the plant. Initially a coal feed rate of 300
Ib/hr was selected as a basis for sizing the
plant. Heat losses from the plant, however,
were appreciable at this scale — amounting to
20 percent or more of the enthalpy exchanged
between hot gases and coal in the devola-
tilizer-desulfurizer. Electrical heating was
used as a means for minimizing this loss.
Some of the critical internal dimensions of the
reactors were also small — 2 inches or less. It
was decided, therefore, to increase the
capacity of the process development plant to
1200 Ib coal/hr. Electrical heating is not
required; minimum clearances of 3 to 4 inches
are achieved. Finally, no increased cost of the
plant is predicted. The reactor designs are
simplified; outside dimensions of the pressure
vessels remain unchanged. The solids feed and
discharge systems are adequate for the
increased capacity without modification.
The operating pressure for the gasification
system is that required to supply fuel to the
gas turbine combustor — a minimum of 10 to
16 atm. Current large industrial gas turbines
use a combustor pressure of about 9 atm
(gauge); advanced models in the next decade
will probably use increased pressures, around
16 atm (gauge).
Operating temperatures have been
estimated for the two reactors. The tempera-
ture for the devolatilizer has been chosen as
1600°F — high enough to crack higher hydro-
carbons and thus to minimize carryover of
tars.23 This temperature also is close to that
required to maximize the effectiveness of lime
sorbents in sulfur removal.16 A somewhat
lower temperature — 1400 °F — may still be
effective for sulfur removal and would
decrease the heat requirements for processing
the coal in this reactor. The temperature for
the combustor has been chosen as 2000-
2100°F as required to agglomerate the coal
ash.19 The temperature in the gasifier will
depend on the effectiveness of the solids in
transferring heat between the combustor and
the gasifier; the more effective the transfer,
the smaller the temperature difference.
Gas velocities in the fluidized bed of the
devolatilizer and gasifier (and thus bed
diameters) have been selected on the basis of
assumed particle size distributions and
densities to achieve high capacities without
excessive carry over of solids. Gas velocities in
the draft tube and downcomer of the devola-
tilizer are chosen to achieve a ratio of about
80 to 1 of recycle solids to coal feed.9"13
Finally, in the combustor gas velocities are
chosen to minimize the quantity of char in the
agglomerating section; it is expected that
txcessive char will inhibit agglomeration.
The depths of various bed sections have
been chosen on the basis of various criteria.
The combustor is deep enough to complete
combustion of the char24 and to capture 80 to
90 percent of the ash.19 The gasifier is deep
enough to react about half of the steam
fed.25'26 The devolatilizer is dimensioned to
circulate the solids at the desired rate, to
devolatilize the coal, and to remove 95 percent
of the sulfur from the fuel gases.
The compositions of gas streams
throughout the system have been estimated or
selected on the basis of prior results or
practice in coal gasification. Water gas equili-
brium has> been assumed to relate concen-
ra-3-9
-------
trations of H2O, CO2, H2 and CO. It seems
possible, however, that in this particular
system greater amounts of the H2O will react
with the char due to the fluidization of this
material27 and greater quantities of CH4 will
be produced due to rapid heating of the coal to
1600°F in the presence of H2.
Material and heat balances have been
carried out which indicate that heat input to
the devolatilizer-desulfurizer in addition to
that provided by the hot gases from the
gasifier may be required. This heat can be
provided by exchange of solids between the
gasifier and desulfurizer with additional heat
generated directly within the devolatilizer by
supplying air to the downcomer.
Detailed design of the process development
plant is now about 50 percent complete.
Figure 7 shows a model of the plant. Specifi-
cations for process vessels will shortly be ready
to send to suppliers. The current schedule
calls for mechanical completion of the plant in
October 1973 assuming that complete funding
is immediately available. The estimated cost of
the plant is $4.2 million.
Laboratory Studies of Coal Gasification for
Power Generation
To support the design of the process
development plant and the evaluations of
commercial performance and economics,
laboratory studies are now underway in two
areas:
1. Cold model studies of fluidized beds to
study bed circulation, jet penetration,
solids separation and elutriation, and
solids attrition.
2. Limestone sort/eat behavior studies to
study sulfur capture, sorbent regenera-
tion, and sorbent disposal.
Work will be initiated shortly in a third area:
Coal behavior studies to study devola-
tilization, char gasification, and ash
pgglomeration.
Cold model studies have been carried out in
a two dimensional fluidized bed (Figure 8) to
study the transport of particles upward in a
draft tube and downward in a downcomer.
The penetration of the jet of particles
emerging from a draft tube into the fluidized
bed at its upper end has also been studied.
The transfer of particles from the downcomer
to the draft tube at the base of the devolatilizer
is now being observed. Correlations relating
gas flow rates; particle flow rates, pressure
drops, and bed dimensions have been
produced. These correlations are useful in
designing the devolatilizer and in determining
recirculation rates within it from process
measurements. Observations are now being
carried out on bed slugging, solids separation
and elutriation, solids attrition, and solids
movement in fluidized beds simulating the
gasifier.
Limestone and dolomite behavior has been
studied in a thermogravimetric analyzer29
(Figures 9 and 10), which measures weight
changes of small samples of the solid as it is
exposed to various atmospheres. The sorption
of H2S and the regeneration of the resulting
CaS by H2O and CO 2 have been studied at
pressures, temperatures, and gas compositions
projected for the process development plant.
Both the degree of sorbent utilization and the
rate of sorption and regeneration appear
sufficiently high for the process to be practical
(Figures 11 and 12). The oxidation of the
sulfided CaS sorbent in air to CaSO4 has also
been studied as a means for rendering spent
sorbent inert for disposal. A problem has been
encountered in completing this conversion
(Figure 13). Tests of sorbent behavior are
continuing to determine optimum operating
conditions for sulfur removal and regenera-
tion, to estimate the number of sorption-
regeneration cycles the sorbent can effectively
undergo, and to develop an improved sorbent
disposal process.
Systems and Economic Studies
Preliminary studies have been carried out
to compare the cost of the gasification' process
HI-3-10
-------
proposed here with alternative fixed,
suspended, and fluidized bed processes. It
appears that cost reductions of 20 to 40
percent can be achieved by the development of
the process. Cost calculations have also been
made for overall plants comparing a coal
gasification, combined gas and steam turbine
cycle plant with a conventional steam plant
using a stack gas scrubber for sulfur removal.
An electric power plant using the coal
gasification process is estimated to cost 20 to
30 percent less than the conventional plant.
CONCLUSION
A comprehensive program is underway to
develop a coal gasification system for a
combined gas and steam turbine generator
plant. Such a power plant is expected to be
lower in capital costs, lower in pollutant
emissions, but equal in overall efficiency to a
conventional power plant. The goal of the
overall program is to complete the demon-
stration of such a plant before the end of this
decade. Further improvements in gas turbine
technology are expected which will improve
the efficiency and reduce the costs for com-
mercial power plants in the 1980's.
BIBLIOGRAPHY
l.Hauser, L.G., W.H. Comtois, and R.R.
Boyle. The Effect of Fuel Availability on
Future R & D Programs in Power
Generation. (Presented at the American
Power Conference. Chicago. April 18-20,
1972.)
2. Evaluation of the Fluidized-Bed
Combustion Process. Summary Report,
Vol. 1. Westinghouse Research
Laboratories. Pittsburgh, Pa. Prepared for
Office of Air Programs, Environmental
Protection Agency, Research Triangle
Park, N.C. under Contract Number CPA
70-9. October 1972.
3. Rudolph, P.F.H. New Fossil-Fueled Power
Plant Process Based on Lurgi Pressure
Gasification of Coal. (Presented at Joint
CIC-ACS Conference, American Chemical
Society, Toronto. May 24-29,1970. pp. 13-
38.)
4. Robson, F.L., A.J. Giramonti, G.P. Lewis,
and G. Gruber. Technological and
Economic Feasibility of Advanced Power
Cycles and Methods of Producing Non-
polluting Fuels for Utility Power Stations.
Final Report, United Aircraft Research
Laboratories. East Hartford, Conn.
Prepared for the National Air Pollution
Control Administration, Durham, N.C.
under Contract Number CPA 22-69-114.
December 1970.
5. Bund, K., K.A. Henney, and K.H. Krieb.
Combined Gas/Steam-Turbine
Generating Plant with Bituminous-Coal
High-Pressure Gasification Plant in the
Kellermann Power Station at Lunen.
(Presented at 8th World Energy
Conference. Bucharest. June 28-July 2,
1971.)
6. Way, S. New Directions in Power
Generation-MHD. In: Proceedings of
North American Fuel Technology
Conference, Ottawa, Canada, June 1970.
7. Keairns, D.L. and D.H. Archer. New
Directions in Power Generation — Fuel
Cells. In: Proceedings of North American
Fuel Technology Conference, Ottawa,
Canada, June 1970.
8. Dent, F.J. Methane from Coal. BCURA
Quarterly Gazette, 42:1-14, 1960.
9. Dent, F.J., R.F. Edge, D. Hebden, F.C.
Wood, and T.A. Yarwood. Experiments
on the Hydrogenation of Oils to Gaseous
Hydrocarbons. Midlands Research
Stations, The Gas Council, Research
Communication GC37, Inst. of Gas
Engineers Trans., pp. 594-643.
10. Murthy, P.S. and R.F. Edge. The Hydro-
genation of Oils to Gaseous
Hydrocarbons. Midlands Research
Station, The Gas Council, Gas Council
Research Communication GC88, IGE
Journal, August 1963, pp. 459-476.
IH-3.11
-------
11. Thompson, B.H., B.B. Majumdar, and
H.L. Conway. The Hydrogenation of Oils
to Gaseous Hydrocarbons. Midlands
Research Station, The Gas Council Gas
Council Research Communication GC122.
IGE Journal, pp. 415-428, June 1966.
12. Horsier, A.G. and B.H. Thompson.
Fluidization in the Development of Gas
Making Processes. Midlands Research
Station. The Gas Council pp. 51-59,
March 1967.
13. Horsier, A.G., J.A. Lacey, and B.H.
Thompson. High Pressure Fluidized Beds.
Chemical Engineering Progress.
65(10): 59-64, 1969.
14. Pilot-Scale Development of the CSF
Process. R & D Reports Number 39,
Volume IV, Book 3. July 1, 1968 -
December 31, 1970. Consolidation Coal
Company, Research Division, Library, Pa.
Prepared for Office of Coal Research,
Department of the Interior, Washington,
D.C. under Contract Number 14-01-0001-
310 (1).
15. Theodore, F.W. Low Sulfur Boiler Fuel
Using the Consol CO2 Acceptor Process.
Report Number 2. Consol Coal. Prepared
for Office of Coal Research, U.S.
Department of the Interior, Washington,
D.C. under Contract Number 14-01-0001-
415, November 1967.
16. Craig, J.W., G.L. Johnes, G. Moss, J.H.
Taylor, and D.E. Tisdall. Study of
Chemically Active Fluid Bed Gasifier for
Reduction of Sulphur Emissions. Final
Report. Esso Research Center, Abingdon,
Berkshire, England. Prepared for Office
of Air Programs, Environmental Protec-
tion Agency, Research Triangle Park,
N.C. under Contract Number 70-46. June
22, 1970 to March 1972.
17. Godel, A.A. Ten Years of Experience in
the Technique of Burning Coal in a
Fluidized Bed. Revue Generale de
Thermique. 5:348-358, 1966.
18. Godel, A.A. and P. Cosar. The Scale-Up of
a Fluidized Bed Combustion System to
Utility Boilers. American Institute of
Chemical Engineers Symposium Series.
67(116):210-218, 1971.
19. Goldberger, M.W. Collection of Fly Ash in
a Self-Agglomerating Fluidized-Bed Coal
Burner. (Presented at Winter Annual
American Society of Mechanical
Engineers Meeting. Pittsburgh. ASME
paper 67-WA/FU-3. November 1967.)
20. Jequier, L., L. Longchambon, and G. Van
de Putte. The Gasification of Coal Fines.
J. Inst. Fuel. 33:584-591, 1960.
21. Jequier, L. et al. Apparatus for Dense-
Phase Fluidization. U.S. Patent 2,906,
608, September 29, 1959.
22. Jones, J.F., F.H. Schoemann, J.A.
Hamshar, and R.T. Eddinger. Char Oil
Energy Development. Chemical Research
and Development Center, FMC Corpora-
tion. Prepared for Office of Coal
Research, U.S. Department of the
Interior, Washington, D.C. under
Contract Number 14-01-0001-498,
October 1966 - June 1971.
23. Consol Coal, Private communication from
E. Gorin and G. Curran.
24. Hoy, H.R. and A.G. Roberts. Fluidized
Combustion of Coal at High Pressure.
(Presented at Annual AIChE Meeting,
San Francisco. November 1971.)
25. von Fredersdorff, C.G. The Reaction
Rates Between Carbon-Carbon Dioxide
and Carbon-Steam at 2000 °F and 1 atm.
I.G.T. Research Bulletin No. 19, 1955.
26. Blackwood, J.D. and F. McGrory. The
Reaction Rates Between the Gas and the
Carbon at 1600°F and High Pressure (1 to
50 atm). Australian J. Chemistry. 11:16-
33, 1958.
27. Blackwood, J.D. and A.J. Ingerme. The
Reaction Rates Between the Gas and the
ra-3-12
-------
Carbon at 1600 °F and High Pressure (1 to
50 atm). Australian J. Chemistry. 13:194-
209, 1960.
28. Squire, A.M., M.J. Gluckman, R.A. Graff,
R. Shinnar, and J. Yerushalmi. Studies
Toward Improved Techniques for Gasi-
fying Coal. Part II: Technical Presen-
tation. Submitted to the National Science
Foundation (RANN) by The City College
and Research Foundation of The City
University of New York, June 1972.
29. O'Neill, E.P., D.L. Keairns, and W.F.
Kittle. Kinetic Studies Related to the Use
of Limestone and Dolomite as Sulfur
Removal Agents in Fuel Processing.
(Presented at 3rd International
Conference on Fluidized Bed Combustion.
Hueston Woods. October 29 - November
1, 1972.) (Session I, Paper 6, this
document.)
APPENDIX A
Heat and Material Balances for 1200 Ib/hr
Process Development Plant
The heat and material balances pertain to a
mode of operation in which air is fed to the
downcomer of the devolatilizer-desulfurizer to
generate part of the heat requirements of the
vessel. Exchange of hot solids between the two
reactors is another way to increase the heat
input to the devolatilizer-desulfurizer; this
mode is not compatible with the planned
independent operation of the reactors,
however.
Air fed to the downcomer reacts mainly
with carbon since fuel gas is kept out of the
zone by the flow of. fluidizing gas. The air for
combustion is introduced as part of the
fluidizing gas. Rapid, countercurrent flow of
solids and gas in the downcomer prevents
excessive temperature rise.
The material balance is given in Table Al
and the individual flow streams are identified
in Figure Al. Separate heat balances for the
two reactors are presented in Table A2 and
A3.
The heating value of the fuel gas product is
dependent on the assumptions made and, in
the present example, is 123 Btu/scf. Assump-
tions regarding stream temperatures, fines
elutriation rates, and transport and fluidizing
gas requirements are incorporated in Table
Al.
Other assumptions follow:
1. Ultimate analysis of coal:
c
H
N
S
O
Ash
74.0 wt%
5.0 wt'%
1,5 wt %
3.5 wt %
6.0 wt%
10.0 wt%
100.0 wt %
2. The products of devolatilization, including
decomposition of tar and oil, were estimated
as follows:
The volatile matter includes all of the
oxygen and hydrogen and half of the sulfur in
the coal.
Oxygen divides equally between qarbon
and hydrogen in the volatile matter, forming
CO and H2O.
Sulfur is evolved as H2S.
The remaining hydrogen forms Ha and
CH4 in the molecular ratio of 2 to 1.*
Carbon in the volatile matter in excess of
that producing CO and CH4 reverts to solid
carbon.
3. The heat of carbonization of the coal is
small and can be neglected.
4. Twenty-three percent of the total char
carbon derived from the coal is burned in the
*Bituminous Coal Research, Inc., Gas Generator Research
and Development. Survey and Evaluation, Phase Qne, Vol
Two. BCR Rept. L-156, pp. 223-5 (1965).
m-3-13
-------
Table A1. MATERIAL BALANCE FOR 1200 Ib/hr PROCESS DEVELOPMENT PLANT
Stream no.
Temperature, °F
Solids
Lb/hr
Composition, wt %
Fixed carbon
Volatile matter
Ash
MgO-CaCOa
MgO-CaS
Gas
Lb/hr
Comoosition, mole %
N2
CO
C02
H2
H20
CH4
H2S
Air
1
77
Coal
1200
55.0
35.0
10.0
Transport
480
53.7
19.0
9.2
14.3
1.0
2.8
2
1000
Calcined
dolomite
696
100
Transport
47
53.7
19.0
9.2
14.3
1.0
2.8
3'
1600
Char
440
82.3
16.7
—
4
1600
Fines
560
66.7
33.3
Fuel gas
7354
50.0
17.7
8.6
13.3
7.9
2.6
5
1000
Ash
120
100
—
6
1600
Spent
dolomite
659
Tin
22.3
Stripping
94
53.7
19.0
9.2
14.3
1.0
2.8
7
2000
Fines
280
50.0
50.0
Gasifier
product
4680
48.6
19.5
8.8
11.4
11.2
0.1
0.4
8
1000
Char
440
83.3
16.7
Transport
88
53.7
19.0
9.2
14.3
1.0
2.8
9
1000
Fines
560
66.7
33.3
Transport
112
53.7
19.0
9.2
14.3
1.0
2.8
10
1000
—
Air
3113
100
11
400
—
Steam
741
100
12
665
—
Fluid-
izing
1546
10.5
89.5
-------
Table A2. COMBUSTOR-GASIFIER HEAT
BALANCE
(Btu/hr)
Input
Air and steam preheat
Transport gas preheat
In char and fines feed
Reaction heat: S -* H2S
C -* CO
C -* C02
H20-> H2
Utilization
Heat losses
Out in ash
Out in fines
Out in product gas
827,400
55,900
369,600
5,700
1,633,300
2,649,500
-2,138,000
3,403,400
321,900
25,800
167,100
2,888,600
3,403,400
downcomer and produces CO and COa in the
molecular ratio of 2 to 1.
5. Recycled fuel gas, after drying to 1 percent
moisture, is used to transport solids and to
Table A3. DEVOLATILIZER-OESULFURIZER
HEAT BALANCE
(Btu/hr)
Input
In combustor/gasifier product gas
In fines from combustor/gasifier
In dolomite feed
In transport gas
In fluidizing gas
Combustion of carbon in downcomer
Water gas shift reaction
Utilization
Heat losses
Out in char and fines
Out in spent dolomite
Out in product gas
Desulfurization reactions
2,888,600
167,100
165,600
38,400
127,100
1,337,700
16,700
4,741,200
267,000
493,700
298,300
3,626,300
55,900
4,741,200
strip char from the spent dolomite.
6. The concentration of CO2, CO, H2O, and
H2 in the product streams of the two reactors
are related in accordance with the water gas
shift equilibrium.
m-3-15
-------
cs
>^
o
cr
o
LLl
LU
2.0
1.0
1960
1965
YEAR
Figure 1. Annual electric energy generation in the U.S.
III-3-16
-------
3.0
I960
1970
YEAR
1980
1990
Figure 2. Forecast of power generation by fuel in the U.S.
IU-3-17
-------
Figure 3. Generation additions by, fuel and type in the U.S., 1970-1990.
IH-3-18
-------
FLUIDIZED BED
DEVOLATILIZER/
DESULFURIZER
COAL-
LIMESTONE
(CaC03) '
10-15
aim
1600 °F
HOT FUEL GAS
CHAR
E
120-160 Btu/scf)
SPENT
-STONE
(CaS)
FUEL GAS
ASH
10-15
atm
1900 °F
FLUIDIZED BED
COMBUSTOR/
GASIFIER
STEAM
HEAT RECOVERY
BOILER
STACK GAS
*
PARTICIPATE
REMOVAL
HIGH TEMPERATURE
GAS COMBUSTOR
COMPRESSOR
TURBINE
BOOSTER COMPRESSOR
' GENERATOR
AIR
0.5 Et
0.5 Et
,.
TURBINE
Figure 4. Coal gasification - combined cycle plant.
-------
j*
CLEAN FUEL GAS
-»- TO GAS TURBINE
CRUSHED _
COAL
HOT
GASES
LIME
SORBENT ^^
CaO
DRY COAL
i
lp
•:-v.;?;
J
'•£.*&••
88
ii
"•:'
^,
1
/-•ii
I
•\
•— — .
\
i
^H4R
— — _
>• SPENT
SUKBtNl UBS
— *
V HOT FUEL GAS
T
I
RECIRCULATING BED
DEVOLATILIZER/DESULFURIZER
(
i*.
f
)
;.'.-'•'.'•-'..•/ IUIML
$#::•:'. GASIFIER
•:.'•'•::•>'.':••
i";-
J}
1
1
>| AGGLOMERATING
i COIWBUSTOR
I AIR
• _ fTFftM
ASH
COAL DRYER
Figure 5. Westinghouse multistage fluidized-bed gasification process.
III-3-20
-------
CONTACT COOLING AND
COOLER. FLUIDIZING
GAS
COOLING GAS QUENCH CYCLONE
COOLER SCRUBBER
GASIFIER
ASH AGGLOMERATION
CHAR
LOCK HOPPERS
(N) NITROGEN PURGE
PRESSURING HEADER
(A) COOLING GAS
RETURN HEADER
© DEPRESSURING
HEADER
CHAR FEED
CONVEYOR
PROCESS
PREHEATER
TCOOLEITFUEL
u,«-rctf / nrowr., r 600M Btii/hr
WATER RECYCLE WATER INCINERATOR CIRCULATING
CIRCULATING COMPRESSOR J,RCULAT»NG
COOLER
3800M Btu/hr
PROPANE C02 STEAM
PROCESS AIR CHAR
COMPRESSOR STORAGE
w
to
Figure 6. Westinghouse coal gasificatjon process development plant flow sheet.
-------
CO
DOLOMITE
CONVEYOR
DOLOMITE FEED
LOCK HOPPERS
DEVOLATILIZER
-DESULFURIZER
CYCLONE
QUENCH
SCRUBBERS
SECONDARY
SPE?ARATOR
COMBUSTION (FUTURE)
UNIT
^-v (FUTURE)
COAL FEED
CONVEYOR
COAL FEED
LOCK HOPPERS
CHAR
DRAWOFF - POT
SPENT
DOLOMITE
OXIDIZER
CHAR
LOCK HOPPERS
FINES
LOCK HOPPERS
DISENGAGING
POT
». STORAGE
DOLOMITE COAL
STORAGE STORAGE
OXIDIZED DOLOMITE
COOLER
DOLOMIE
REGENERATION
(FUTURE)
CHAR
COOLER CHAR CONVEYOR
QUENCH WATER
CIRCULATING
Figure 6. Westinghouse coal gasification process development plant flow sheet.
-------
ALUMINA/SAND
SEPARATOR
GAS TURBINE
(UNINSTALLED)
Figure 7. View of solid waste combustor and gas preparation subsystem.
-------
Figure 8. Two-dimensional cold model recirculating bed.
IH-3-24
-------
Figure 9. Thermogravimetric analyzer for high temperature and pressure reaction studies on
3nd ch3r.
-------
N
>.
I CO
Is
Q>
1
CO
O
•o
I
O)
L
IH-3-26
-------
0.9
T=760°C
( p=i°*»
0.7
S 0.5
u.
0.3
0.1
10
15
20
25
30
35 '40
TIME, min
Figure 11. Sulfidation of calcined limestone at H2S + CaO-»-CaS
IH-3-27
-------
o
LU
(3
UJ
CE
0.80
0.70
0.60
0.50
C 0.4
o
fe
0.30
0.20
0.10
0.0
T=700°C P=10atm
H20 = C0£ = 20%
-35+40 DOLOMITE
(86% Ca FRACTION SULFIDED)
0 24 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
TIME, min
Figure 12. Regeneration of carbonate from sulfided dolomite CaS+ h +
CaCOs + H2S.
IH-3-28
-------
N
O
X
o
o
o
ATMOSPHERIC PRESSURE
AIR
700 C
O DOLOMITE 1337
55% Ca SULFIDED
LIMESTONE 1359
99% Ca SULFIDED
100
TIME, sec
Figure 13. Oxidation of sulfided limestone CaS-f-202-»-CaS04.
HI-3-29
-------
FUEL GAS+FINES
1600°F
DEVOLATILIZER-
DESULFURIZER
100-1000° F
FLUIDIZING GAS —
| „
GAS+FINES 2000° F
CYCLONE
FUEL
GAS
CHAR
1600 °F
SPENT
DOLOMITE
1600°F
STRIPPING
GAS
1600°F
COMBUSTOR-
GASIFIER
FINES
T.G. | COOLERS
SPENT
DOLOMITE
CHAR
vv
1
FINES
v
1000°F
-*T.G.
1000°F
STEAM
1000 °F
T.G. TRANSPORT OR STRIPPING GAS
Figure A1. Flow diagram for 1200 Ib/hr process development plant.
III-3-30
-------
4. SULFUR RETENTION IN FLUIDISED BEDS
OF LIME UNDER REDUCING CONDITIONS
J. W. T. CRAIG, G. MOSS, J. H. TAYLOR, AND D. S. TISDALL
Esso Ltd., Abingdon, Berkshire, England
ABSTRACT
Work done at the Esso Research Centre at Abingdon under EPA, OAP Contract CPA 70-46 has
amply confirmed the early promise of the oil-fired fluidised-bed desulphurising gasifier. Since the
Second International Conference on Fluidised-Bed Combustion an extensive test programme has
been carried out on a batch basis. In addition, the 7 x 106 Btu/hr continuously regenerating
gasifier, which was previously described, has been constructed and successfully operated. The
results show that the gasifier will operate satisfactorily at lower stoichiometric ratios and at higher
temperatures than were previously established.
INTRODUCTION
Papers presented at the First and Second
International Conferences on Fluidised-Bed
Combustion dealt with the mechanisms of sul-
phur absorption in fluidised beds of lime,
under oxidising and reducing conditions.
information was also given for the gasifying
case, concerning the effect on desulphurising
efficiency of variations in stoichiometric ratio,
the mean particle size of the bed material, and
bed replacement rate in cyclic operation.
Work carried out under OAP contract CPA
70-46 has enabled this information to be sup-
plemented considerably. In this more recent
work an American limestone, BCR 1691, was
used in conjunction with a western hemisphere
fuel oil.
in-4-i
-------
Batch Test Equipment
Two batch gasifiers were specially
constructed for the programme but were
essentially similar to those used previously.
Figure 1 shows a drawing of the new reactor
vessel with its various connections. The reactor
is made of mild steel lined with a castable
refractory. The lower section containing the
fluid bed is 7 inches in diameter and 33 inches
high. Fuel oil enters the reactor through a
single 1/4-in. diameter nozzle which protrudes
1 inch from the reactor wall at a point 5 1/2
inches above the bottom. The upper, disen-
gaging, section of the reactor contains two
cyclones which can be drained externally. The
distributor is of radial form cast in refractory
cement, with 16 horizontal holes distributed
around its circumference. The units are
brought into operation by underfiring with gas
and kerosine. When gasifying, product gas
leaves through a cyclone outlet and is flared
outside the laboratory. A portion of the gas is
burned in a sample flare located above the
reactor; the combustion products are analysed
for SO2, O2, CO, and COa. Because of the
wide range of sulphur compounds which
might be present in the product gas itself, this
is the only practical method for measuring
desulphurising efficiency. During regenera-
tion similar analyses are made of the
undiluted gas.
Batch Test Methods
Two types of tests were made, fresh bed
tests and cyclic tests. A new batch of calcined
lime was used in each of the fresh bed gasi-
fication tests. These tests were used to rapidly
screen the effects of the following variables:
bed depth, gas velocity, particle size, and
air/fuel ratio. The cyclic tests are the nearest
simulation to continuous gasifier operation
that can be obtained in batch units. The same
charge of lime is subjected to repeated cycles
of sulphur adsorption and regeneration. After
each regeneration a portion of lime is removed
and is replaced by an equivalent amount of
fresh limestone. The added limestone is
calcined to lime during the early part of the
next gasification cycle. Without this replace-
ment the activity of the bed gradually declines.
With replacement, the activity falls initially,
but after a few cycles it reaches an equilibrium
level determined by the replacement rate.
Batch Test Results
Because the fuel handling capacity of a
gasifier of given dimensions increases as the
air/fuel ratio decreases, efforts have been
made to operate at the lowest possible air/fuel
ratios and to define the limitations that may
exist. Under adiabatic conditions there is a
relationship between air/fuel ratio and
operating temperature. In the batch units,
however, the operating temperature was in any
case lower than the adiabatic level because of
heat losses through the walls and could be
lowered still further by means of the bed
cooling heat exchanger.
The results plotted in Figure 2 show how
changes in operating temperature affect
desulphurising efficiency over a range of
air/fuel ratios. All of these tests were made
using different batches of the same bed
material and in each case there was 5 percent
by weight of sulphur in the bed material when
the plotted result was obtained. The figures in
parentheses by each point indicate the carbon
content of the bed material. It can be seen that
there is a tendency for desulphurising
efficiency to fall as the air/fuel ratio is
lowered, and that this tendency is increased
when the operating temperature is lowered
from between 840 and 870°C to 800°C. It can
also be seen that there is, as would be
expected, a tendency for the carbon content of
the stone to rise as the air/fuel ratio is
lowered; this tendency, too, is increased when
the operating temperature is lowered. It would
not seem unreasonable to infer that the
presence of a surface layer of carbon reduces
the reactivity of the stone. Further evidence
supporting this view is shown in Figure 3. In
this case desulphurising efficiency is plotted
against percent of calcium utilisation at four
different air/fuel ratios. The figures in paren-
theses again relate to the carbon content of the
ffl-4-2
-------
stone at the adjacent point. If the point at 0.38
percent by weight of carbon is compared with
the point at 4.68 percent by weight of carbon,
it will be seen that although the temperature
was about the same in the two cases, 850 as
against 845°C, the desulphurising efficiency at
the higher carbon content was much lower.
The results obtained at an air/fuel ratio of
14.8 percent of stoichiometric and a tempera-
ture of 780°C might also have been affected by
some degree of recarbonation of the stone; this
might account for the relatively poor desul-
phurising efficiency at very low calcium utili-
sations when the carbon content was also
probably quite low.
At this point it might be appropriate to
discuss the factors controlling stone carbon
content. It has been found that much more
carbon than hydrogen is oxidised in the
gasifier. The injected oil cracks on the surface
of the stone laying down carbon which is
oxidised when the stone reaches the vicinity of
the distributor. It follows that the amount of
carbon on the stone is a function of the rates of
deposition and removal; the rate of deposition
reflects the rate at which fuel is injected, and
the rate of removal reflects the availability of
oxygen and the relative proportions of CO 2
and CO which are produced.
It is evident that CO2 is the predominant
combustion product which is produced in the
vicinity of the distributor where the incoming
air meets the carbon coated bed material. As
the combustion products pass up through the
bed, however, there is a tendency for the CO 2
to be reduced to CO at a rate dependent upon
the temperature and the availability of carbon.
This point is aptly illustrated by the data
plotted in Figure 4. These data relating
CO/CO.2 ratio with temperature were
obtained during the regeneration of a bed
which initially contained 7 percent by weight
of carbon. Readings of temperature and gas
composition were taken at one minute
intervals, and no oxygen appeared during the
period under consideration. It may be seen
that during the first two intervals there was a
steep increase in CO/CO2 ratio as the
temperature rose. Subsequently, however, as
the carbon content of the bed fell so did the
CO/CO2 ratio despite a further rise in
temperature. As a matter of interest, the
thermodynamic equilibrium gives a CO/CO 2
ratio of 40 at 900°C.
An example of the working of this
mechanism during gasification is given by the
data plotted in Figure 5. In this figure, percent
weight of carbon on lime is plotted against
duration of exposure to gasifying conditions.
The bottom curve lines out at a very low
carbon content within the first 60 minutes of
operation, whereas the top two curves show a
progressive increase in carbon content over the
first 120 minutes; though both of these curves
show a tendency for the carbon content to
reach an equilibrium level. The figures in
parentheses are the desulphurising
effeciencies at 120 minutes gasification time;
the 865°C run gave a much better result than
the 845°C run, although the stoichiometric
ratios as well as the temperatures were very
close to each other, being 24.8 and 24.1
percent. It seems reasonable to attribute this
difference in performance to the difference in
carbon content, these being 0.2 and 6.0
percent by weight, respectively. It may be
deduced from these results that it is
advantageous to run the gasifier at as high a
temperature as possible at low stoichiometric
ratios. Unfortunately, due to the heat "loss
through the walls of the small batch reactors it
has not been possible to approach adiabatic
conditions; but experience with the
continuously operating gasifier has confirmed
that good results are obtainable at air/fuel
ratios in the region of 18 percent of stoichio-
metric and temperatures in the region of
870°C. This is an area which will be explored
in more detail.
The tests which gave the results which have
so far been discussed were made in beds 15-in.
deep and at superficial gas velocities in the
region of 4 ft/sec. In this work, the practice
has been to discuss gas velocity in terms of the
superficial air rate, i.e., the velocity which the
air supplied to the reactor would reach at the
m-4-3
-------
operating temperature when flowing through
the empty vessel. The actual superficial gas
velocity is higher, due to the presence of
cracked oil products; the deviation will
increase as the air/fuel ratio falls.
The curve plotted in Figure 6 shows the
basic effect of variations in superficial gas
velocity. In these tests the bed depth was 15
inches, the operating temperature was 870°C,
the air/fuel ratio was 25 percent of stoichio-
metric, and there was 5 percent by weight of
sulphur in the bed material. These data were
obtained using a U.S. fuel and fresh beds of
U.S. stone, BCR 1691. It can be seen that
satisfactory results were obtained at
superficial gas velocities up to 6 ft/sec, but at 8
ft/sec there was a marked deterioration in
performance. The next step was to check the
effect of varying the bed depth. Figure 7 shows
the results which were obtained at a gas
velocity of 6 ft/sec, with all variables other
than bed depth held at the levels used in the
previous set of tests. In this case satisfactory
results were obtained at bed depths greater
than 15 inches but a marked fall in desul-
phurising efficiency occurred when a 10 in.
bed was used. At a gas velocity of 8 ft/sec the
desulphurising efficiency was only 40 percent
with a 10-in. bed, but reached nearly 100 with
a 20-in. bed.
The size of the stone which was used to
obtain these results was in the range 300 to
3000 /xm. This size range is obtained by de-
dusting the 1/8-in. diameter tailings from a
normal limestone quarry screening operation.
As a matter of interest some comparative tests
were made with narrower size range fractions
sieved from this stone. The results of these
tests are shown in Figure 8. It can be seen that
the narrower sized fractions gave markedly
poorer results than the material from which
they were obtained. The results listed in Table
1, however, show that when the smaller of the
two narrow cut fractions was substituted for
the make-up feed of a full size range bed
under cyclic conditions, contrary to expecta-
tions, the desulphurising performance was
improved. There is no clear-cut explanation
Table 1. EFFECT OF LIMESTONE PARTICLE SIZE
ON DESULPHURIZING EFFICIENCY IN CYCLIC
TEST
Size range
of makeup, /urn
Replacement rate
CaO/wts
Lined out SRE, %
Sulfur Input
Test number
T-3
(Cycles 1-14)
300 to 3175
2.38
61
3.0
T-3
(Cycles 24-31)
600 to 1400
2.37
68
3.1
for this effect, but it is possible that the single
cycle results reflected quality of fluidisation as
well as particle size.
In the case of fresh bed tests, the extent to
which the stone is reacted is not important;
but when cyclic tests are made it is necessary
to choose realistic levels. It is possible to
deduce on thermodynamic grounds that under
continuously operating adiabatic conditions
the reaction in the sulphur content of the stone
per pass through the regenerator will be in the
region of 1.0 to 1.5 percent by weight. It
follows that it is not possible to operate a cyclic
test, in the absence of progressive coking,
which adds more fuel, unless a sulphur
content higher than about 2 percent by weight
is obtained at the end of each absorption cycle.
This somewhat higher sulphur content is
required because of the additional heat
demand which occurs in a batch regeneration.
In continuous operation, more or less adia-
batic conditions obtain; in a batch operation
the refractory lining must be raised in
temperature during regeneration as well as the
bed, and the heat losses through the wall of
the reactor are appreciable. Because a fair
amount of-oxidation occurs at relatively low
temperatures there is a tendency for more
calcium sulphate to be formed during batch
regenerations than would be expected under
continuous conditions. This sets the level of
the sulphur content of the bed at the
beginning of the subsequent test. This level is
generally found to be about 2 percent by
weight.
111-4-4
-------
The fuel which was used for the cyclic tests
had a sulphur content of 2.3 percent by
weight; for convenience, a standardised run
duration of 45 minutes was generally used.
When account is taken of the range of gas
velocities and stoichiometric ratios which were
covered this gave sulphur inputs ranging from
1.5-3.7 percent by weight on lime per run.
The first series of cyclic tests were aimed at
obtaining a direct comparison of the
reactivities of BCR 1691 and the U.K. stone
which had previously been tested. The U.K.
stone is about 98 percent CaCOa, whereas
BCR 1691 is of inferior quality, containing
only 88 percent CaCOa. Fresh bed tests had
given results indicating that the two stones
were equally effective when reacted to the
same degree. Cyclic tests, however, revealed
that BCR 1691 is so inferior in performance
under these conditions that it requires about
three times the stone replacement rate for
an equal desulphurising performance in a bed
15.5-in. deep. Fortunately, it was also found
that a modest increase in the depth of the bed
gave an improvement in performance which
was quite disproportionate, allowing the stone
replacement rate to be substantially reduced.
An indication of the nature of the relationship
between bed depth, stone replacement rate,
and desulphurising efficiency is given by the
data shown in Table 2. All of these tests were
made at 870°C with 25 percent of stoichio-
metric air and a superficial gas velocity of 6
ft/sec.
Table2. EFFECT OF BED DEPTH AND STONE
REPLACEMENT IN RATE DESULPHURIZING
EFFICIENCY IN CYCLIC TESTS
Bed depth.
inches
15.5
20.0
20.0
22.0
Moles CaO
per moles S
1.39
1.43
2.9
1.4
Lined out
S.R.E.,%
61
90
98
97
It will be seen that increasing the bed depth
from 15.5 to 20 inches improved the desul-
phurisation efficiency from 61 to 90 percent at
roughly the same stone replacement rate of
approximately 1.4 mole CaO/mole sulphur. In
order to improve the desulphurisation
efficiency to 98 percent with a bed depth of 20
inches, it was necessary to increase the stone
replacement rate to 2.9 mole CaO/mole
sulphur. When, however, the bed depth was
increased to 22 inches then a 1.4 mole
CaO/mole sulphur stone replacement rate
gave a desulphurisation efficiency of 97
percent. It remains to be seen whether BCR
1691 will give a similar performance in the
continuous gasifier at a superficial gas velocity
of 6 ft/sec.
Results Obtained Operating the Continuous
Gasifier
The continuously operating gasifier has
been described previously; a detailed
discussion of its construction and
commissioning is not within the scope of this
paper. A general view of the installation,
however, is shown in Figure 9. During the
commissioning period three runs were made
giving a total operating time under gasifying
conditions of about 460 hours. U.K. stone was
used in these tests, but with U.S. fuel
containing 2.5 percent by weight of sulphur
and 350 ppm of vanadium. The prime purpose
of these runs was . to demonstrate fuel
gasification with sulphur removal on a
continuous basis; the study also took a quick
look at the effects of some of the controllable
variables. The study showed that the gasified
fuel ignites readily and burns with a bright
luminous stable flame. Smoke free operation
was obtained with about 1.5 percent oxygen in
the flue gas over long periods. This is a better
performance than that given by the
conventional oil burner which the gasifier
replaced. When tested prior to conversion it
was found that the package boiler required
about 3 percent oxygen in the flue gas for
smokeless operation.
m-4-5
-------
The operating conditions covered during
the test runs are indicated in Table 3. In
Table 3. CONTINUOUS GASIFIER OPERATING
CONDITIONS
Table 4. GASIFIER PERFORMANCE
Programme item
Number of test
Test duration
Limestone used
Oil used
Gasifier temperature
Regenerator temperature
Bed depth
Superficial gas velocity
Lime particle size range
Lime replacement rate
Air fuel ratio
Oil feed rate
Pilot plant operation
3
91-202 hu-
ll.K., Denbighshire
Venezuelan 2.5% S
820-920 °C
1050-1100 °C
13-23 in.
2.8-4.3 ft/sec
300/3200-800-320% m
0.54-4.8 mole CaO/moleS
15-31 % of stoichiometric
61-82lb/hr-ft2
general the superficial gas velocity was about 4
ft/sec; when the fuel rate was varied the
operating temperature was controlled by
recycling flue gas. During a considerable
proportion of the operating time desul-
phurisation was virtually complete. Although
this result was highly gratifying, it did not
yield much information concerning the effects
of the independent variables. In one period of
19 hours duration in which virtually no SO2
was detected in the flue gas, the running
conditions were as shown in Table 4. The
gasifier temperature was about 900°C, the
pressure drop through the bed averaged 14.5
inches water gauge the air/fuel ratio was 23
percent of stoichiometric, and the stone
replacement rate was about 1.4 mole
CaO/mole sulphur entering the bed. The
superficial gas velocity averaged 3.7 ft/sec. In
another period of 25 hours duration at the end
of the test, the operating temperature was
about 880°C, the pressure drop through the
Duration of
experiment, hr
Gasifier temper-
ature, °C
Air/fuel ratio,
% stoichiometric
Bed pressure drop,
in. H2O
Gas velocity, ft/sec
Stone replacement
rate, moles CaO/moles S
Sulphur removal
efficiency, %
19
900
23
14.5
3.7
1.4
100
25
880
22
19.5
3.9
0.85
95
bed was about 19.5 inches water gauge the
air/fuel ratio was 22 percent of stoichiometric,
and the desulphurisation efficiency averaged
95 percent. During the first 16 hours of this
period the stone replacement rate was about 1
mole CaO/mole sulphur, during the
remaining 9 hours the Ca/S ratio was (X6 of
stoichiometric giving an average figure for the
25 hours of 0.85 mole CaO/mole sulphur. In
the last hour the air/fuel ratio fell to 18
percent of stoichiometric, but the desul-
phurising efficiency never fell below 91
percent.
The operational problems which were
encountered during these runs, which totalled
460 hours under gasifying conditions, were of
a minor nature and remedial action has since
been taken. In view of the results which were
obtained there is no doubt at all that the
process is a feasible proposition.
m-4-6
-------
WATER COOLANT
CONTROL VALVE
GAS SAMPLING POINT.
MANOMETER TAPPING
THERMOCOUPLE
FILL POINT
SAMPLING FLAME BURNER
FLARE
BED COOLING LOOP
NITROGEN
•••IMIM^
1
r**
BED SAMPLE POINTS
BED DRAIN
DISTRIBUTOR PLATE
MANOMETER/THERMOCOUPLE
TWIN GAS VENT PIPES
TWIN CYCLONE
PILOT BURNER
REFRACTORY LINING
MANOMETER TAPPING
MANOMETER TAPPING
THERMOCOUPLE
FUEL INJECTOR
•THERMOCOUPLE
MANOMETER TAPPING
IGNITER AND GAS INLET
AIR INLET
SECTION ON A-A
Figure 1. CAFB batch unit reactor.
IH-4-7
-------
on
ce.
5 % BY WT SULPHUR IN BED
D 840 to 870 °C
80
35
STOICHIOMETRIC AIR, %
Figure 2. Interaction between air/fuel ratio and bed temperature.
III-4-8
-------
100
80
Q£
Q£
E *
40
0.38)
(4.68)
, (11.88)
10
20
30
40
CALCIUM UTILISATION, %
Figure 3. Result at 6 ft/sec in number 4 units ( sulphur removal curves at different air/fuel
ratios).
III-4-9
-------
TEMPERATURE,^
Figure 4." CO/CO? profile during regeneration, fresh bed test no. 7.
IH-4-10
-------
o
00
0
(83.6)
(85.3)
T,°C
810
845
865
% STOIC.
20.5
24.1
24.8
0
40
80
200
240
120 160
GASIFICATION TIME, min
Figure 5. Change in carbon content of lime during batch gasification cycle 1-A unit.
280
ffl-4-11
-------
100
5 % BY WT SULPHUR IN BED
40
4.0
5.0 6.0 7.0 8.0
SUPERFICIAL GAS VELOCITY, ft/sec
Figure 6. Basic effect of superficial gas velocity.
9.0
III-4-12
-------
a
OC
£ 80
5 % BY WT SULPHUR IN BED
70
60
10.0
15.0
BED DEPTH (STATIC), in.
Figure 7. Basic effect of bed depth.
20.0
IH-4-13
-------
O
cc
Q-
40
20
D 300 to 3175 Ji
• 600 to 1400 JJ
O 1200 to 3175 ji
10
20
Ca UTILISED, %
Figure 8. Basic effect on particle size range.
30
m-4-14
-------
5. THE CO2 ACCEPTOR GASIFICATION PROCESS-
A STATUS REPORT-APPLICATION
TO BITUMINOUS COAL
G. P. CURRAN AND E. GORIN
Consolidation Coal Company
ABSTRACT
This paper discusses experience gained and problems encountered during startup operations of
the Rapid City pilot plant. A project schedule is given for completion of that phase of the work
aimed at production of pipeline gas from low-rank western coals.
Process revisions that must be made in application of the CO2 acceptor system to high-sulfur
bituminous coals are discussed. The major revisions are installation of pretreatment facilities to
handle caking coals and an increase in gasification temperature to accommodate the poorer
reaction kinetics.
Experimental work on the pretreatment of bituminous coals to render them suitable for
pressurized gasification by preoxidation. Highly fluid coals such as Pittsburgh seam do not.
Promising results are reported via a "Seeded Coal" type process.
A revised flow sheet and heat and material balance is given for application of the CO2 acceptor
process to the processing of bituminous coals. Recycle-of CO 2 is a key feature in this operation.
INTRODUCTION
The CO 2 acceptor process has been under
development for a number of years. The major
goal of this work has been the production of
pipeline gas from low-rank western coals. The
process has been extensively described in
numerous previous publications and no
description is deemed necessary here. A
relatively complete description of the technical
basis of the process and its economic potential
is available in reports to the Office of Coal
Research.
A pilot plant to test the process has been
constructed at Rapid City, South Dakota. The
project is financed jointly by the Office of Coal
Research and the American Gas Association.
The purpose of this paper is to give a brief
status report on the Rapid City operations and
of the contemplated development schedule. It
also discusses problems and opportunities
involved in the application of the process to
treatment of bituminous coal. The use of bitu-
minous £oals in the process is not only of
in-5-i
-------
interest for production of pipeline gas, but
more broadly for the production of low-sulfur
boiler fuel.
The production of low-sulfur boiler fuel
from bituminous coals by an adaptation of the
process to produce a low-Btu gas without2 and
with low-sulfur char3 as co-product has been
studied in the course of a research contract
between Consolidation Coal and the EPA.
This work is discussed in more detail in a suc-
ceeding paper at this conference. It also
formed the subject of a paper presented at the
Second International Conference on Fluid-
ized-Bed Combustion. This paper gives a brief
resume of some work now being conducted for
the EPA in the pretreatment of caking bitumi-
nous coals to establish operability in pres-
surized fluid-bed gasifiers.
ra-s-2
-------
STATUS OF PILOT PLANT DEVELOP-
MENT
The construction of the pilot plant was
completed with formal acceptance of the plant
on December 28, 1971. Mechanical testing of
the various plant components occupied the
next period through the end of March 1972.
During this test period, a number of unit
operations were successfully carried out.
These included operation of the gas
purification, char grinding, and lockhopper
systems. The fired heaters were put on stream
and the process vessels were successfully
heated to 1400-1600°F by hot gas circulation.
Pilot plant operations since April 1972 were
aimed at initiation of an actual gasification
run using lignite char as a feedstock. The
initial run was chosen to demonstrate a
simplified two vessel version of the process to
be conducted at 150 psig. The system to be
demonstrated is illustrated by two vessels
shown in Figure 1 — the gasifier and
regenerator. The details shown on the
remainder of the flow sheet should be
disregarded at this time since they refer
specifically to future operations with
bituminous coal which will be discussed later.
It should also be noted that in the pilot plant
runs the temperature in the gasifier was
programmed for lower temperatures than
indicated in Figure 1, i.e., at 1520°F.
One of the unique features, from the
engineering point of view, in the COa acceptor
process is the dual fluo-solids handling system
wherein acceptor, which originally is either
dolomite or limestone, is fed to fluidized beds
of char. The acceptor particles are bigger and
heavier than the char particles and shower
down through the char bed. They collect as a
separate and segregated fluid bed of acceptor
in the boot at the bottom of the gasifier from
which the acceptor is recirculated to the
regenerator.
The ability to maintain a segregated
fluidized acceptor bed reasonably free of char
is one of the key elements in achieving a
successful demonstration of the process.
This feature has been well demonstrated in
the prepilot scale work at Consolidation Coal
Company's Research Laboratories, but one of
the purposes of the Rapid City pilot plant is to
demonstrate that this operation can be
successfully scaled up.
Acceptor circulation tests were carried out
in April, May, and June 1972 between the
gasifier and regenerator preparatory to
initiation of the gasification tests. Operations
in June were hampered by the June 9 Rapid
City flood and its aftermath. Difficulty was
encountered in these tests due to chronic
plugging of the pressure probes used to
control the operation. The plugging was due in
part to an inadequate purge gas system.
The situation was rectified by increasing
the diameter of the pressure probes from 1/4
in. to 1/2 in., installing a rod out system to
break plugs, installing duplicate probes at
critical measuring points, and improving the
purge gas supply system.
Successful hot continuous circulation of
dolomite-based acceptor was demonstrated
for a period of 14 hours during the end of this
period. The MgCO3 portion of the stone was
calcined to MgO and the stone circulated in
the "half calcined" condition. The run was
terminated involuntarily due to a pressure
upset and corresponding loss of pressure
balance between the two vessels.
One of the problems encountered during
this period was a high rate of attrition of the
acceptor in its soft, half-calcined condition.
This led to excessive generation of fines which
caused plugging difficulties at the entrance to
the quench towers.
The softness of the stone is of a transitory
nature since it is known from our prepilot
experience that it hardens rapidly when it is
cycled through the actual process operations
which were not attempted during these
circulation tests.
During July a successful hot acceptor
circulation test was performed, and
preparations were made again to start an
in-s-3
-------
actual gasification run. Beds of half-calcined
acceptor and char were established in the
gasifier and regenerator, respectively. Circu-
lation of acceptor through the char bed was
initiated, but difficulty was experienced in
obtaining a distinct char-acceptor interface in
the gasifier boot.
The acceptor-char mixture was transferred
via the lift leg to the regenerator. The acceptor
was, however, rich in char, and since the lift
gas was nearly pure air combustion of the char
in the lift line was initiated. This caused
development of a hot spot in the line which
resulted in its rupture and termination of the
operation.
The above incident is not typical of normal
operation since the lift gas usually is recycle
regenerator offgas substantially free of
oxygen. In this case air was present because
circulation was started before char
combustion in the regenerator was initiated.
A fourth startup was initiated after the
ruptured lift line was repaired in August.
Difficulty occurred at all times, however, with
blockages in the acceptor and char transfer
lines. It was obvious from both the
temperature and pressure profiles that fluid-
ized beds of acceptor in the gasifier boot and
char in the gasifier was not being maintained.
In spite of this, sufficient acceptor was
transferred through the lift line to establish
the regenerator bed, and sufficient char was
fed to the regenerator to establish combustion
therein. The regenerator was increased to
1700°F. The unit was shut down on August 20
for inspection when the char feed line to the
gasifier plugged. This was done to determine
the cause for the failure to establish the
desired fluid beds in the gasifier.
Inspection of the gasifier revealed that the
fluidization difficulty was due to a failure of
the refractory in the gasifier boot, particularly
at the seam between the head and the gasifier
shell, which allowed gas to bypass the solids
bed through massive holes and cracks in the
refractory.
The gasifier was taken down and a new
refractory configuration installed. Specifically,
the outer layer of soft insulating refractory was
removed from the gasifier boot to be replaced
with harder castable refractory. The unit, at
the time of this writing, was scheduled to be
put back in operation again early in October
1972.
The initial operating difficulties have not
been related to fundamental deficiencies in
the process itself. It is expected that successful
demonstration of the COa acceptor process
will be achieved in the Rapid City pilot plant
in the near future.
The original schedule has been set back
about 6 months, first by some delays in
completion of construction and second by the
startup difficulties outlined above. A new
operating schedule has been drawn up and
submitted to the Office of Coal Research for
their approval. The new schedule calls for
completion of the original pilot plant program
in July of 1974. The contemplated program,
however, does not encompass testing of
Eastern bituminous coals.
The processing of bituminous coals is of
interest not only for production of pipeline gas
but also for production of low-sulfur boiler
fuel. A series of new problems are introduced
when the process is adapted to use of caking
bituminous coals. Extension of the pilot plant
operating period as well as some modification
of the equipment would be required to study
bituminous coal processing.
PROCESSING OF BITUMINOUS COALS
VIA CO2 ACCEPTOR PROCESS
Pretreatment via Preoxidation
The use of fluidized-bed technology for the
gasification of caking coals requires that the
feed coal be pretreated to render it non-caking
in order to sustain an operable bed. The
problem becomes more severe as the operating
pressure is increased and may also be a
function of hydrogen partial pressure. The
specific role of hydrogen partial pressure as
m-s-4
-------
distinct from total system pressure in
intensifying the pretreatment problem has not
been fully defined.
The effect of increasing total pressure is
illustrated by the two experimental
observations outlined below. A highly caking
Pittsburgh seam coal was successfully
processed in 1949 in an atmospheric pressure,
1 ton/hr fluid-bed gasification unit without
any pretreatment.5 The above admittedly was
accomplished at a relatively low coal through-
put rate of 25 lb/hr-ft2, but the effect of
higher rates was not explored.
The other observation was, that in
processing non-caking sub-bituminous coals
at 20 atmospheres pressure in the hydro-
devolatilizer of the CO 2 acceptor process,
agglomeration of the bed solids occurred
unless the coal feed was pretreated by mild
preoxidation.10
The work on the development of the
synthane process at the USBM6 again
illustrates the fact that successful operation of
a pressurized fluidized-bed gasification
process with bituminous coal requires that the
feed be pre-treated by preoxidation.
A study of the degree of pretreatment
required for pressurized fluid-bed gasification
of two types of bituminous coals has been
carried out for the EPA One coal was from
the Ireland Mine in northern West Virginia
and is typical of the highly fluid, high sulfur
Pittsburgh No. 8 seam. The other coal was
from the Hillsboro Mine in central Illinois and
is" typical of the more weakly caking, high
sulfur Illinois No. 6 coals. The experimental
investigation was carried out in the same 4-in.
ID reactor previously used in the development
of the CO2 acceptor process.1 c The experi-
mental method is described in detail in the
Annual Report to the EPA. Only a brief
summary of results will be presented here.
The gasification conditions chosen for
testing operability of the pretreated coals are
those outlined in Table 1 and correspond to
conditions selected for adaptation of the CO
acceptor process to produce low-Btu gas.
Under normal conditions in the CO2 acceptor
process, i.e., for production of pipeline gas, a
higher partial pressure of hydrogen prevails
and even more severe pretreatment may be
required.
The severity of preoxidation conditions
required to establish operability for the two
coals in the gasifier operated at the conditions
cited in Table 1 are given for the case of a 28 x
100 mesh feedstock in Table 2.
Table 1. TYPICAL CONDITIONS USED FOR
TESTING OPERABILITY OF PRETREATED
COALS IN GASIFIER
Temperature, °F
Pressure, psig
Feed rate,lb/hr
Feedstock
Fluidizing velocity, ft/sec
Input, scfh
Stearn
C02
N2
Air
Percent carbon burnoff
Mean particle density,
Ib/ft3
1700
206
4.83
28 x 100 mesh,
Pretreated,
Hillsboro Coal
0.34
77
38
103
105
55
43.2
1700
206
5.50
20x1 00 mesh
Pretreated,
Ireland Coal
0.33
78
35
111
105
Table 2. MINIMUM CONDITIONS OF SEVERITY
OF PREOXIDATION TO PROVIDE OPERABLE
GASIFIER FEEDSTOCK
Coal
Temperature of preoxidation, °F
Size consist of coal feed
Stages of preoxidation
Oxygen consumed, wt %
(referred to raw coal)
Stage 1
3 Stage 2
Total
Ireland Mine
750
28 x 100 mesh
2
18.6
9.3
27.9
Hillsboro
810
28x100 mesh
1
8.7
-
8.7
Two criteria are used to evaluate the
impact on economics of the results of the
preoxidation tests. The first is the percent pre-
oxidation required as compared with the
"adiabaticr" quantity, i.e., with the amount of
in-s-5
-------
oxidation by heat balance to sustain the
reaction at the desired temperature.
This relationship is illustrated in Figure 2.
It is readily seen that the amount of pre-
oxidation required for Ireland Mine coal is
four times the adiabatic quantity at 750°F.
This is a strong economic debit since, in order
to carry out this process in practice, large
amounts of heat must be removed from the
preoxidizer. Somewhat larger amounts of pre-
oxidation are permitted by adiabatic
operation if the temperature in the preoxidizer
is allowed to rise. However, in the case of
Ireland Mine coal, substantially higher
preoxidation temperatures are precluded since
the preoxidizer itself becomes inoperable.
The demonstrated preoxidation severity
required for the Illinois No. 6 coal, however, is
only slightly above the adiabatic level. As a
matter of fact, lower extents of preoxidation
may in fact be permissible in this case since an
investigation of the lower limit of preoxidation
was not carried out.
The other desired property of the pre-
oxidized coals relates to the fluidization
behavior. In order to operate the gasifier at a
practical throughput and for the preoxidized
coals to have a relatively high particle density,
it is necessary to use a relatively coarse feed.
These properties permit operation of the
gasifier at reasonable gas velocities without
excessive entrainment, and they maintain a
reasonably high bed inventory to satisfy the
demands of the gasification kinetics. It is seen
from the data in Table 3 that significant
particle swelling occurred in the preoxidation
treatment of both coals.
Reduction in the amount of preoxidation
required for Ireland Mine coal can be
accomplished by use of finer coal. For the
reasons cited above, however, the use of fine
coal is economically undesirable.
The conclusion from these preoxidation
studies was that Illinois coals may be
pretreated successfully with pressure
gasification by use of "adiabatic"
Table 3. PROPERTIES OF COAL AND PRE-
OXIDIZED COALS
Coal
Treatment
%Sulfur
Ash
Mean diameter, inch
Mean density, tb/ft3
Pittsburgh seam
Ireland Mine coal
Raw
4.52
11.36
0.0165
81.0
28.3%
Preoxidation
at750°F
3.75
13.76
0.0156
56.4
Illinois No. 6
Raw
4.93
21.86
0.0176
80.0
8.7%
Preoxidation
at810°F
3.40
15.39 a
0.0175
51.8
aAsh is low because of segregation and selective removal of
mineral matter in the preoxidizer.
preoxidation. Highly fluid Pittsburgh seam
coals, however, require economically excessive
amounts of preoxidation, unless an
impractically small size consist feed coal is
used.
Pretreatment via "Seeded Coal Process"
The principle of the preoxidation method
of pretreatment is to convert the coal to a more
rigid structure via oxidation, such that the
fluidity is severely reduced when the coal
undergoes pyrolysis.
The "Seeded Coal Process" would operate
on just the reverse principle and actually
utilize the natural fluidity of the coal. In the
process visualized, char would be circulated at
a high rate by means of a lift gas through a
draft tube immersed in a normal fluidized
bed. Coal and fine size seed char would be fed
into the draft tube. The external fluid bed
would be maintained at 1000 to 1400 °F either
by injection of air or hot fluidizing gas from a
gasification step, as shown in Figure 1.
The coal melts, smears out over the
surfaces of the seed char and external bed
material, and then solidifies on completion of
pyrolysis.
The demonstration of such a device7 was
successfully carried out in the low-
temperature carbonization section of the CSF
Coal Liquefaction Pilot Plant at Cresap, West
Virginia. The feed material, in this instance,
was somewhat different and constituted the
ra-s-6
-------
underflow from the hydrocyclone separation
of the extraction effluent. Coal extract in this
case was used instead of the fluid coal; the
extraction residue was used instead of the seed
char. Other differences were that the mixture
was sprayed into the draft tube as a slurry and
operating temperatures and pressures were
lower, i.e., 825 to 925°F and approximately 4
psig, respectively.
In this particular installation, a 36-in. ID
carbonizer was employed in which there was
installed a 6-in. ID x 11-ft high draft tube.
Solids were circulated through the draft tube
by injection of about 3500 to 4500 scfh of lift
gas into the bottom of the tube. The feed was
sprayed into the circulated char stream within
the draft tube by means of a nozzle 3 feet
above the lift gas injection point.
Solids circulation rates of the order of
100,000 Ib/hr were achieved in this device,
while complete operability and product size
consist control was maintained with extract
feed rates up to 200 Ib/hr. The ratio of
extraction residue solids to extract was in the
range of about 1.5:1 to 3:1. The above
throughput rates do not necessarily represent
the capacity of the system since high extract
feed rates were not available and consequently
were not tested.
The above results led to an attempt to
apply the same system to coal even though
coal is a less fluid material than extract and
the operating conditions, particularly the
pressure, required are higher.
Tests with an inert bed of 48 x 100 mesh
char at 1500 °F and 15 atm system pressure
showed that the external baffle was effective.
The solids circulation rate upward through the
tube was measured by substituting a known
amount of air for some of the N2 entering the
solids feed line.1C From the measured
temperature rise, the solids flow rate was
calculated as 900 Ib/hr by heat balance.
Calculations involving the pneumatic transfer
line model, devised in the course of develop-
ment of the CO2 acceptor project, showed that
without the external baffle about 270 of the
340 scfh of N2 fed to the bottom of the exter-
nal bed had entered the draft tube; with the
baffle, the flow was reduced to about 60 scfh.
Seven tests were made with the modified
draft tube, using an external bed of 48 x 100
mesh char at 15 atm system pressure.
Common conditions for the runs are listed
below:
Ireland mine coal, sized to 100 x 200 mesh
Coal feed rate: 2.0 Ib/hr
Duration of feeding: 3.3 minutes
Air to coal feed line equivalent to 100% of
adiabatic preoxidation level at the
temperature used.
Gas flows, scfh
Air + N2 to coal feed line 65
N2 to accelerating line 85
N2 to bottom of external bed 340
Tests were made at temperatures from 900
to 1500°F, in 100 degree increments. Temper-
ature limits of operability were established as
follows: (1) at 900dF little or no smearing
occurs as was shown by presence of coal-
derived material in the form of hollow spheres
in the bed after the run, and (2) at 1500°F
caking occurred in the draft tube.
Unfortunately, we were severely handicapped
by the small scale of the equipment available,
since the draft tube principle had to be
adapted to the existing 4-in. diameter gasifier
vessel.
The potential advantages of the process are
that it will supply a feedstock that is assuredly
operable with respect to agglomeration at
gasifier conditions; and it can produce a
dense, closely sized feedstock substantially
free of fines. This will permit a higher gasifier
throughput than otherwise.
A series of exploratory tests were carried
out with the configurations A, B, and C
indicated in Figure 3. Best results were
obtained with configuration C, but even here
two basic deficiencies were noted. From the
appearance of the agglomerates obtained, it
was apparent that insufficient mixing
occurred in the draft tube between the injected
ra-s-7
-------
coal and the circulating char. Part of the
difficulty is associated with the small scale of
the equipment, since calculations show that
the Reynolds number in the draft tube is
barely above the Stokes Law range. Also, it
was apparent that most of the'fluidizing gas
bypassed the main bed in favor of the draft
tube. The result was that a fluidized bed was
not maintained external to the draft tube.
To overcome these limitations, the
configuration C of Figure 3 was modified as
follows:
To allow installation of an external baffle
which would maintain fluid ization of the
external bed, the draft tube was raised 2
inches and the inlet lines were lengthened
accordingly. An elliptical baffle 3-5/8 x 1-3/4
x 1/16-in. thick was welded to the accelerating
gas line below the mouth of the tube at a slope
of 60° from the horizontal. To help promote
mixing a conical baffle was installed inside the
tube with the apex of the cone positioned 1/2-
in. above the end of the coal inlet tube.
The products from the runs at 1000 to
1400°F all showed more uniform smearing
than in any of the previous runs without the
internal baffle. At the end of each run, the
system was depressured and the bed was
drained by removing the coal inlet line. The
hot bed material was quenched rapidly by
contact with dry ice in the catchpot. The entire
bed material then was screened at 28 and 48
mesh. All the run products contained some
+48 mesh agglomerates which were external
bed particles cemented together by a thin film
of coal-derived-material. No agglomerates
larger than 28 mesh were found. The fewest
agglomerates occurred at 1300°F, indicating
that this may be the optimum temperature
with respect to uniformity of smearing. The
amounts of +48 mesh agglomerates which
formed are listed below:
Temperature,
°F
1000
1100
1200
1300
1400
+ 48 Mesh Agglomerates, wt%
of Bed Inventory
18.0
16.0
15.5
8.0
10.7
The particle density, measured in mercury, for
the +48 mesh agglomerates formed at 1300°F
had a high value of 85 lb/ft3.
An attempt was made to run for a
prolonged period at 1300°F and 15 arm
system pressure to determine the size distri-
bution of the "equilibrium" product. To
simulate the seed char in the commercial
embodiment (fines from the internal cyclones
in the gasifier) an initial external bed of -100
mesh precarbonized char was established.
Then, 100 x 200 mesh Ireland Mine coal and
additional -100 mesh char were fed to the
draft tube at rates of 2 and 4 Ib/hr,
respectively. The fine char contained a
considerable amount of -325 mesh material
which was elutriated from the reactor. The
outlet piping system of the present equipment
was not designed to handle large amounts of
solids. The run had to be terminated after 35
minutes of feeding coal because the outlet
system began to plug. Thus, an equilibrium
bed was not established. However, analysis of
the bed showed that it contained 50 weight
percent of +100 mesh agglomerates, with a
top size of 24 mesh.
The high particle density achieved is
favorable, in that smearing of liquid coal over
the seed particles apparently occurs as
desired.
The small size of the existing equipment
precludes any further meaningful studies of
the seeded coal process. The radial clearance
between the inlet line and the wall of the draft
tube is only 0.15 inch. The mouth of the tube
eventually would become choked by the larger
agglomera-vs which inevitably would be
formed.
ra-5-8
-------
The results of the exploratory studies
strongly indicate that future studies should be
made.
Several essential factors are required to
achieve success in such an operation. Intensive
mixing in the draft tube is required to achieve
smearing of the "liquid" coal over both the
seed and recirculating char. A sufficient
residence time in one pass through the unit of
the recirculating burden is needed to complete
the "drying out" or carbonization of the coal.
Finally, the draft tube must be large enough to
handle the largest size particles made in the
process without choking. All these factors
point to a need for a larger unit in which the
draft tube would be at least 2 inches in
diameter as opposed to the present 0.680 inch.
Such a unit, of course, would have a much
higher capacity for coal feed which would lie
approximately in the range of 30 to 300 Ib/hr.
Pretreatment via Pre-Extraction
This method would be a more direct
application of the draft tube pyrolysis method
already demonstrated at the CSF pilot plant at
Cresap, West Virginia. Two principal
differences would be required here. First of
all, the extraction slurry would be injected into
the draft tube unit operated at 15-20 atm
pressure instead of at substantial atmospheric
pressure, and secondly, the ratio of extract to
extraction residue normally would be greater
since little or no extract need to be recovered
as such. This technique would possibly prove
to be more operable.
Reaction Kinetics
The other limiting factor is the gasification
of bituminous coals via the CO2 acceptor
process is the poor reaction kinetics relative to
sub-bituminous coals and lignites. The
treatment of the gasification kinetics is
described in more detail in a companion
papers presented at this conference which
deals with gasification of bituminous coals to
produce low-Btu gas. Suffice to say, that data
available to us indicate that gasification rates
with bituminous coal chars are about l/15th
of lignite chars. This necessitates increasing
the gasification temperature to 1650°F to
achieve adequate rates. There is little incentive
from the kinetic point of view to increase the
temperature with the available size and
density of the char treated, i.e., 28 x 100 mesh
and 45 lb/ft3 particle density. The limiting
factor in throughput in this instance becomes
the fluid dynamics of the char particles rather
than kinetics of gasification. The use of
coarser feedstocks, of course, would remove
this limitation and would require higher gasi-
fication temperatures to achieve higher
outputs.
The acceptor process becomes deficient in
heat supplied to the gasifier at this high
temperature (1650°F) unless one or both of the
following expedients is employed. More
sensible heat as opposed to chemical heat may
be supplied by increasing the acceptor
circulation rate; or heat may be supplied by
recycle of carbon dioxide.
The amount of supplementary heat
required by either of the two above expedients
also may be lowered by increase in operating
pressure.
Process Description
An outline of the proposed process is given
in Figure 1 previously mentioned. This in-
corporates the system of COi recycle to supply
the heat deficiency in the gasifier, and the
"countercurrent" contacting of the feed coal
with the gasifier offgas in the draft tube
pretreater. This latter step not only "decakes"
the coal feed but also significantly increases
the Btu of the product gas.
The heat and material balance relationship
is given in Table 4 for the processing of
bituminous coal via the flow scheme of Figure
1. The heat and material balances were
derived by adaptation of the computer
program as previously devised for the OCR
project on the development of the CO2 ac-
ceptor process. The process assumptions used
are generally quite similar to those outlined in
the companion paper.8
m-5-9
-------
Table 4. HEAT AND MATERIAL BALANCE IN TREATMENT
Product
°F
moles
Ib
Composition,
mole %
CH/i
H2
CO
CO2
H20
H2S
N2
MgO-CaS
MgO-CaCOs
MgO-CaO
Hydrogen
Carbon
Stream number
1
Steam
1200
2.759
2
CO
1200
1.100
3
Recycle
gas
400
0.157
Same
as
10
4
Pre-
treater
char
1250
4.222
60.9
5.01
94.99
5
Calcined
acceptor
1 873
5.750
27.2
-
72.8
6
Casifier
except
recycle
1650
5.064
3.0T
46.43
21.30
7.52
21.74
-
-
7
Spent
acceptor
1650
0.230
27.2
25.5
47.3
8
Fuel
char
1650
2.124
37.3
2.61
97.39
aDry, H2S-free basis. HHV = 424 Btu/ft3.
r-»
Total sulfur content = 100 ppm.
"Water content only. Dust, tar, and phenols content not known.
System pressure: 14.84 atm (204 psig) .
Basis: 100-lb dry Ireland Mine coal
wt % (dry basis) mole
H
C
N
O
S
Ash
4.8
69.8
1.2
7.6
4.3
12.3
2.381
5.812
0.475
0.1341
6 percent moisture, as fed; 50 percent of coal sulfur removed in preheater;
95 percent carbon burnout in regenerator; cold efficiency, 79.9 percent;
total carbon gasified, 64.4 percent; and fixed carbon gasified, 56 percent.
ffl-5-10
-------
OF BITUMINOUS COAL VIA CO2 ACCEPTOR PROCESS (FIGURE 1)
Stream number
9
Pretreater
gas except
recycle
1250
7.120
17.15
40.41
19.92
9.89
11.70
0.93
10
Water
gas
100
6.279
19.63a
46.25
22.80
11.32
-
-
11
Air
550
9.051
0.44
12
Regenerator
gas includ-
ing lift
1873
13.548
-
1.98
27.27
0.78
b
69.97
13
Absorber
gas
-
4.538
Same
as
12
14
Makeup
acceptor
60
0.230
48.1
15
Dirty
liquor
200
0.841
100C
16
Lift
gas
1200
2.500
-
2.7
2.5
0.4
-
94.4
17
Excess
gas
-
0.938
as
16
in-5-ii
-------
The gasification temperature 1650°F is
compatible with an overall gasification rate of:
RX = 39 x 1(H Ib fixed carbon gasified/lb
carbon in bed/minute, and the other con-
ditions cited with respect to carbon and steam
conversions, etc.
Recent experiments under the EPA
contract with the seeded coal process show
that the pretreated coal from the draft tube
(actually, char, since the temperature is in the
range of 1.200 to 1300°F) will have a high
particle density of about 85 Ib/ft3. The
particle density of the gasifier bed then would
be about 45 Ib/ft3 after gasification of 56
percent of the fixed carbon (see Table 4).
Calculations involving the gas flow rates,
the above gasification rate, and our fluidized-
bed density correlation10 showed that one
train with the gasifier bed dimensions shown
below will be capable of processing 272,000
Ib/hr of MF coal.
Fluidized bed height:
Gasifier ID:
Bed density:
Fluidizing velocity:
50ft
25.4ft
14.7 Ib/ft3 char
0.81 ft/sec
The estimated Btu content of the dry
product gas is 424 Btu/ft3 and the overall
thermal efficiency on a cold gas basis is 79.7
percent. Thus, a single train is capable of
producing about 1.5 x 108 ft3 or 6.5 x 1010
Btu/day of raw gas. This compares favorably
with the 3.5 x 1010 Btu/day (22.5-ft-ID
gasifier) raw gas output projected for a single
train in the CO2 acceptor process when used
for lignite gasification.
There are a few features incorporated in
the process which should be mentioned.
A small amount of recycle gas (Stream 3) is
added to the gasifier boot to prevent oxidation
of CaS in the recirculating acceptor stream to
CaSC>4 by the incoming steam and CO 2-
The fate of the coal sulfur is not known at
present. It was assumed for the purposes of
the balances that are presented that half of the
sulfur is eliminated as H2S in the pretreater; it
will have to be removed by scrubbing the
product gas.
The regenerator is operated sufficiently on
the reducing side, such that the remaining
sulfur is largely retained and discarded in the
spent acceptor as CaS. It also will be recovered
as H2S via the Claus-Chance reaction as
proposed previously.1 d
The sulfur content in the regenerator off-
gases is low enough, such that it may be flared
without scrubbing; but prior incineration of
residual reduced sulfur forms to SC«2 is
required.
The system as shown here is certainly not
optimum and some improvements are
potentially possible as listed below:
1. Increase operating pressures from 18-20
atm. This will reduce the quantity of CC»2
recycle needed for heat balance and also
raise Btu content of the product gas. As
offsetting features, higher regenerator
temperatures and a system to recycle
product gas to the gasifier would be
required.
2. The CO 2 recycle requirement could be
derived from the product gas. Sufficient
CC>2 would be removed from the product
gas, after water-gas shift and prior to
methanation to provide the CO 2 recycle
requirements. As a matter of fact, the hot
pot absorbent could be regenerated in
such a way as to generate directly the
steam-CO2 mixture required for gasifica-
tion under full system pressure, thus
eliminating the CO2 compressor. A
difficulty here is the presence of H2S in
the product gas. Processes are available,
however, which afford at least partial
selectivity in H2S versus CO2 removal.
3. Recycle tar oil and tar to pretreater rather
than the gasifier. It would be used as a
vehicle to pressurize and transport the
coal to the pretreater. By preheating the
slurry to obtain partial extraction, im-
proved operability may be achieved.
ra-s-12
-------
4. By use of a catalytic afterburner to
combust the CO in the regenerator offgas,
it may be possible to generate surplus
power through the expansion turbine. An
afterburner will be required in any case to
convert traces of H2S, COS, and S2 in this
gas to SO 2-
ACKNOWLEDGMENT
Appreciation for financial support of the
work described in this paper and for per-
mission to publish the results is expressed to:
1. Office of Coal Research, Department of
the Interior, and the American Gas
Association for the portion describing the
Rapid City pilot plant.
2. Environmental Protection Agency for the
portion describing the bench-scale
processing of bituminous coals.
BIBLIOGRAPHY
1. Colsolidation Coal Co., Research &
Development Report No. 16 to the Office of
Coal Research, U.S. Dept. of the Interior,
Washington, D.C. under Contract Number
14-01-0001-415.
a. Interim Report No. 1. Pipeline Gas
from Lignite Gasification — A Feasi-
bility Study. U.S. Dept. of Commerce,
National Technical Information
Service,-PB-166817 (feasibility study),
PB-166818 (appendix). February 1965.
b. Interim Report No. 2. Low-Sulfur
Boiler Fuel Using the Consol CO
Acceptor Process. U.S. Dept. of Com-
merce, National Technical Information
Service PB-176910. November 1967.
c. Interim Report No. 3. Phase II —
Bench-Scale Research on CSG Process.
January 1970.
Book I, Studies on Mechanics of
Fluo-Solids Systems. U.S. Govern-
ment Printing Office Catalog
Number 163.10:16/INT3/Book 1.
Book 2, Laboratory Physioco-Chemi-
cal Studies. U.S. Government
Printing Office Catalog Number
163.10:16/INT3/Book 2.
Book 3, Operation of the Bench-
Scale Continuous Gasification
Unit. U.S. Government Printing
Office Catalog Number
163.10:16/INT3/Book 3.
d. Interim Report No. 4. Pipeline Gas
from Lignite Gasification — Current
Commercial Economics. U.S. Govern-
ment Printing Office Catalog Number
164.10:16/INT4.
2. Curran, G.P., J.T. Clancey, C.E. Fink, B.
Pasek, M. Pell, and E. Gorin. Annual
Report to Office of Air Programs, En-
vironmental Protection Agency under
Contract Number EHSD-71-15. September
1, 1970 to November 1, 1971.
3. Curan, G.P., W.E. Clark, and E. Gorin.
Low-Sulfur Char as a Co-Product in Coal
Gasification. Environmental Protection
Agency, Research Triangle Park, N.C.
EPA-R2-76-060, October 1972.
4. Curran, G.P., C.E. Fink, and E. Gorin.
Proceedings of Second International
Conference on Fluidized-Bed Combustion,
1970. Office of Air Programs, Environ-
mental Protection Agency, Research
Triangle Park, N.C. Publication Number
AP-109. pp. III-l-l to III-l-ll.
5. Unpublished work carried out jointly by
Consolidation Coal and Esso Research.
6. Forney, A.J., SJ. Gasior, R.F. Kenny, and
W.P. Haynes. Proceedings of the Second
International Conference of Fluidized-Bed
Combustion, October 1970. Office of Air
Programs, Environmental Protection
Agency, Research Triangle Park, N.C.
Publication Number AP-109. pp. III-3-1 to
III-3-21.
7. Pilot scale Development of the CSF
Process. Period July 1,1968 - December 31,
1970, R & D Report No. 39, Volume IV,
Book 3. Consolidation Coal Co. Prepared
for Office of Coal Research, U.S. Dept. of
the Interior, Washington, D.C. under
Contract Number 14-01-0001-310.
8. Curran, G.P., B. Pasek, M. Pell, and E.
Gorin. Low-Sulfur Producer Gas via an
m-s-13
-------
B
BFW
QUENCH
TOWER
TOH2S
REMOVAL
SPENT ACCEPTOR
TO GLAUS-CHANGE
STEAM
"0
2
1
Figure 1. Processing of'bituminous coal by the C02 acceptor process.
-------
22
21
20
19
18
17
16
15
G 14
o
i-
Q
X
O
IU
as
a.
12
11
10
Si PREOXIDATION -100 (Ib 02 CONSUMED/lb DRY COAL)
BASIS
COAL IN AT 60 °F
6% MOISTURE IN COAL
AIR IN AT 398 °F
ALL OXYGEN CONSUMED
HEAT OF REACTION:
200,000 Btu/lb mole 02 CONSUMED
400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600
TEMPERATURE, °F
Figure 2. Percent preoxidation versus temperature for adiabatic constraint.
III-5-15
-------
RECYCLE
4 in.
8 in.
24 in.
100 x 28 x
200m 100m
COAL CHAR
RECYCLE
RECYCLE
4 in.
AIR
8 in.
24 in.
100 x 28 x 48 x
200 m 100 m 100 m
COAL CHAR CHAR
4 in.
N2
NZ
6 in.
34 in.
• ex
o
cj
•a:
100 x 48 x
200 m 100 m
COAL CHAR
AIR
N2
CONFIGURATION A
CONFIGURATION B
CONFIGURATION C
DRAFT TUBE
COAL FEED LINE
TIP POSITION
ACCELERATING
GAS LINE
0.500 in. OD x 0.444 in. ID
0.250 in. OD x 0.180 in. ID
HALFWAY INTO SKIRT
0.75 in. OD x 0.680 in. ID
0.250 in OD x 0.180 in. ID
1 in. ABOVE TUBE BOTTOM
0.750 in. OD x 0.680 in. ID
0.250 in. OD x 0.180 in. ID
5 in. ABOVE TUBE BOTTOM
0.375 in. OD x 0.305 in. ID
(TIP POSITION 1 in. ABOVE TUBE BOTTOM)
Figure 3. Configuration of draft tubes used in seeded coal tests.
-------
6. LOW-SULFUR PRODUCER GAS VIA AN
IMPROVED FLUID-BED GASIFICATION PROCESS
G. P. CURRAN, B. PASEK, M. PELL AND E. GORIN
Consolidation Coal Company
ABSTRACT
This paper describes the evolution of the process concepts for generation of clean low-Btu gas
from bituminous coals via fluid-bed gasification. The improved process now under development
for the EPA does not involve the CO2 acceptor principle. Hot sulfur recovery from the gas is
achieved by the use of dolomite. The residual char from the gasifier is utilized in a carbon burn-up
cell. The heat sink utilized in this case is the sensible heat of the air and steam feed to the gasifier.
Dolomites show activity for hot sulfur cleanup via the reaction,
CaCO3 + H2S = CaS + H2O + CO2.
A single limestone was tested and was substantially inactive.
Various dolomites have been assessed and best results are obtained with pure crystalline type
stones.
Experimental background data around other key process steps are also briefly presented.
INTRODUCTION
The production of low-sulfur producer gas
via an adaptation of the CO 2 acceptor process
was described in a paper1 given at the Second
International Conference on Fluidized-Bed
Combustion. The system was described in
some detail along with some supporting back-
ground experimental data.
A detailed process design and feasibility
study of the sytem as well as an experimental
evaluation was carried out under a contract
with the EPA. The results are reported in
detail in the Annual Report.2 The economics
will be briefly summarized later in this report.
The experimental evaluation of the system
indicated feasibility of all steps in the process
with one exception. It was found that the sul-
fur recovery from the acceptor would be
incomplete from the regenerator. This neces-
sitated addition of another step in the process
in which sulfur is rejected by the reaction first
proposed by Squires,3
CaS + H2O + CO2 = CaCO3 + H2S.
(1)
With this added complication introduced, fur-
ther thought was given to refining and simpli-
fying the overall process.
It became apparent that there is no real
advantage in using the CO 2 acceptor reaction
simultaneously with the sulfur acceptor reac-
tion in the gasifier when low-Btu fuel gas is the
III-6-1
-------
desired product. Disposition of the residual thermic calcining reaction in the acceptor
char from the gasifier can be accomplished by regenerator.
use of a carbon burn-up cell which preheats all
the steam and air required for the gasifier.
The sensible heat duty involved in preheating Since the acceptor no longer needs to be in
serves as the "heat sink" .which is necessary to the gasifier, the sulfur acceptance reaction
prevent ash slagging during combustion of the now can be carried out in a separate, external
residual gasifier char. In the CO2 acceptor reactor containing a dense-phase fluidized bed
process the heat sink is provided by the endo- of dolomite in the form of CaCO3 «MgO.
ra-6-2
-------
IMPROVED FLUID-BED PROCESS
DESCRIPTION
A schematic diagram of the revised process
is shown in Figure 1. The simplest con-
figuration occurs when the burn-up cell is
integral with the gasifier. In this instance, it
would be analogous to the CO2 acceptor
gasifier in that the burn-up cell would be in
the form of a "boot" which would contain a
fluid bed of coarse inert solids such as "dead
burned lime." Combustion of the char .residue
from the gasifier would take place in the boot.
In figure 1, the burn-up cell is shown as a
separate reactor. This is a costlier configu-
ration but permits more selective rejection of
ash. The hot fuel gas is desulfurized in the
H2S sorption bed by the reaction,
+ CaCO=CaS
(2)
The bed temperature is held at a level below
which the acceptor can calcine by the reaction,
CaCO3 =CaO + CO2.
(3)
The low-sulfur hot gas is cooled to 1300°F by
heat exchange with the water needed to
generate the gasifier steam, and then is
cleaned of particulates and alkali by high
pressure drop cyclones.
The sulfided acceptor is conveyed to the
regenerator by continuously recirculating a
stream of CO2 and steam. In the regenerator
the "Squires" reaction takes place at about
1300°F,
CaS + CO2 + H2O = CaCO3 + H2 S. (4)
The regenerated acceptor is returned by
gravity to the sorption reactor.
A computer program has been devised to
evaluate the heat and material balance
relationships and overall thermal efficiency of
the scheme shown in Figure 1.
The program evaluates the interaction
between the various components of the system
consistent with the thermodynamic, fluo-
solids mechanic and kinetic restraints on the
system.
The entire system was represented by 27
simultaneous linear and non-linear equations
which represent the five basic process steps
given below:
1. Carbon burn-up cell,
2. Gasifier,
3. Sulfur reactor,
4. Steam-product gas exchanger,
5. Squires reaction - impact of temperature
only,
which also are interrelated by the following
quantities:
1. H, C, and O balance,
2. Heat balances around components 1
through 4, above,
3. Water-gas-shift equilibrium in compo-
nents 2 and 3, above,
4. Methane yield correlation,
5. Equilibrium in the reaction,
CaCO+HS =
(5)
The above equations were solved by an
iterative procedure for the moles of air and
steam fed to the burn-up cell as a function of
the following variables:
1. Burn-up cell temperature.
2. Gasifier temperature.
3. S/l Ca, mole ratio.
4. "Squires" reactor temperature.
Once having computed the input and
output flows and compositions by the method
outlined above, it was necessary to determine
the gasification reactor sizes. The vessel sizing
is determined by the interaction of the fluid i-
zation mechanics of the char particles and the
gasification kinetics.
The basis' used here was to provide for a
single train to process 120,800 Ib/hr of coal.
The fluidized-bed height was fixed at 50 feet.
The fluidized-bed density was then calculated
using the correlation developed during the
work on the CO2 acceptor process. It was
assumed here that a high-density, closely sized
char particle would be generated by the
"seeded coal process" from Ireland Mine coal
(the mean particle diameter was taken as 0.04
m-6-3
-------
inch and the initial particle density of 85
lb/ft3). The reduction in particle density as
affected by carbon burnoff was based on the
relationships developed during the work on
the CO 2 acceptor process.6
It is now necessary to compute the vessel
cross section and fluidizing velocities which
are compatible with the estimated bed inven-
tories by use of reaction kinetic data.
Extensive differential rate data7"9 were
obtained some time ago on the gasification of
bituminous coal chars as a function of tem-
perature, pressure and mole fraction of
hydrogen in hydrogen-steam mixtures. Subse-
quently, extensive kinetic data were obtained
on the gasification kinetics of lignite
char.6'10 In this work it was found that the
reaction rate was strongly inhibited by the
presence of CO as well as hydrogen. Thus, the
prior data on bituminous coal chars could not
be used since the inhibiting effect of CO was
not taken into account. However, under
comparable conditions (in the absence of CO)
it was found that on the average the
bituminous coal chars had about l/15th the
reactivity of the lignite chars.
Therefore, in developing the kinetic calcu-
lations the equations developed for lignite
char10 were used with introduction of a
correction factor of l/15th to account for the
lower reactivity of the bituminous coal chars.
The differential kinetics were translated into
integral kinetics; i.e., they were averaged over
the whole bed by the method given in Ap-
pendix D of the Annual Report.2
Having calculated the integral rate Rf, the
required bed inventory is calculated from the
equation, Jj
R _ lb fixed carbon gasified x JQ -4
T min/lb fixed carbon in bed
The fluidization calculations outlined above
are then used to determine the required vessel
cross section and fluidizing velocity.
A total of 14 cases were computed covering
gasification temperatures between 1650 and
Table 1. RANGE OF VARIABLES STUDIES IN
SYSTEM ANALYSIS
Range of independent
variables studied
Gasifier temperature, "F
Burn-up cell temperature, "F
S/I Ca,mole ratio
Squires reactor temperature, "F
Range of calculated quantities
Cold gas efficiency, %
HHV dry product gas, Btu/ft3
Sulfur removed, %
Steam conversion, %
Carbon to burn-up cell, wt % of C
in coal feed
Gasification lb C (gasified) x 104
rate lb C in bed, min
Gasifier, ID, ft
Char particle density
Gas fludizing velocity, ft/sec
Gasifier cross section index3
Constant parameters
System pressure, atm
Gas outlet temperature, "F
Base Case
1650-1750
1750-1950
0.1-0.4
1200-1300
79.1-81.2
143-149
93.1-97.5
49.4-55.7
14.3-16.9
45-78
24.9-25.3
28.7-35.8
1.46-1.52
392-415
1725
1883
0.2
1300
80.6
147
96.7
53.6
15.4
68
25.1
30.2
1.49
398
- 15 -
- 1300 -
JFt2/109 Btu-hr (HHV of product gas).
1750°F. Because of the kinetic and thermo-
dynamic limits the system is highly con-
strained, and the response of the system is
quite limited. This is illustrated by the ranges
given in Table 1.
The conditions for the base case, which is
felt to be close to a practical "optimum" for
the system, are also given in Table 1. A more
complete heat and material balance around
the base case is also given in Table 2.
The process concept given here has several
potential advantages over the original process
which utilized the CO2 acceptor process, as
outlined below:
Operability May Be Improved in the New
Process
1. The O2 partial pressure to the burn-up cell
is lower than to the previous regenerator.
Steam and air N2 serve as the heat sink.
There is less chance of ash slagging,
especially since the burn-up cell can be
operated at much lower temperatures than
are needed to regenerate CaCO3.
m-6-4
-------
Table 2. KEY STREAM FLOWS AND ANALYSES CASE 6
Identification
°F
moles
Ib
Composition, mole %
CH4
H2
CO
C02
H20
H2S
N2
02
NH3
MgO-CaS
MgO-CaCO3
Hydrogen (as Hj!
Carbon
Stream Nnmhpr
1
Air
398
11 .41
0.40
2
Steam
870
3.504
3
Fuel
char
1725
0.936
23.14
4.25
95.76
4
Burn-up
cell gas
1883
14.68
X
X
X
6.10
24.42
X
61.31
8.17
5
Raw
product
gas
1725
20.37
1.72
16.31
18.96
7.85
10.06
0.66
44.18
X
0.26
6
Spent
acceptor
1624
0.648
20
80
7
"Squires"
offgas
1300
3.522
X
X
X
48.16
48.16
3.68
X
X
8
Regenerated
acceptor
1300
0.648
0
100
9
Sulfur
320
0.1296
4.15
/
10
CO 2
200
0.1296
11
Clean
product
gas
1300
20.50
1.71
16.56
18.50
8.78
10.29
0.02
43.90
X
0.25
Basis: Ireland Mine Coal (100 Ib dry coal)
H
C
N
O
S
Ash
wt %, dry basis
4.8
69.8
1 .2
7.6
4.3
12.3
moles
2.381
5.812
-
0.475
0.1341
-
moisture as fed; system pressure: 15 atm (206 psig)
C/i
-------
2. The Oa partial pressure to the gasifier is
lower, since the air is diluted with all of the
input steam and all of the products of com-
bustion of the burn-up cell. Thus, one
possibly can raise the temperature to
1750°F to improve kinetics without
increasing the danger of ash slagging.
3. More positive contact of dirty gas with the
dense-phase bed of acceptor in sulfur
reactor is effected. Also, in certain circum-
stances it may be possible to use the sulfur
reactor as a fluid-bed filter to remove
"residual" particulate matter.
The Cold Gas Efficiency is Definitely Im-
proved
1. Lower duty to calcine make-up acceptor.
Circulation rate is about 10 percent that of
the original concept.
2. 100 percent burn up of carbon (versus 98
percent).
3. Improved gasification kinetics require less
steam. Thus, less latent heat in product
gas.
4. Less air required. Thus, less sensible heat is
lost with N2.
EXPERIMENTAL BASIS
Pretreatment
This was discussed in the preceding paper4
and only the conclusions need be reiterated
here. Pretreatment by preoxidation is a viable
procedure for the more weakly caking Illinois
No. 6 coals, but is not a desirable procedure
for use with the more highly fluid Pittsburgh
Seam coals. For the latter coals, pretreatment
by the seeded coal process appears promising
but further development to prove out the
method is required.
Gasification Operability and Kinetics
Studies were carried out to demonstrate
operability of the gasifier with respect to both
caking and ash fusion using pretreated Illinois
No. 6 coals. The conditions studied were those
that correspond to the original adaptation of
the CO2 acceptor process to low-Btu gas
production. Complete operability in the 4-in.
diameter gasifier was achieved on both points.
It is felt that the conditions in the present
system are less stringent, as was pointed out
above, such that operability problems due to
ash fusion are less likely to occur.
The background data on differential
kinetics which were used to calculate integral
gasification rates were outlined in the previous
section. Integral gasification rates were ob-
tained also in operation of the continuous unit
with both Disco char and Illinois No. 6 coal.4
The results were given in the Annual Report.2
The results are in approximate agreement
with the basis used for reactor design given
above, although the rates obtained with Disco
char tend to be somewhat lower than those
obtained with the pretreated Illinois coal.
Further data on the kinetics of gasification of
bituminous coals are required to provide a
firmer basis for reactor design.
Carbon Burn-up Cell
No data on the operability of this unit are
available at this time although it is planned to
obtain such data in the course of the present
EPA contract. The operation comprises
combustion at full system pressure of residue
char in the presence of a fluid bed of inert
solids, such as "dead burned lime." An
analogous operation is the regeneration of the
CO 2 acceptor process where residue char is
used as fuel for acceptor regeneration.
Operability of this process has been
demonstrated in prepilot scale work on the
process. A full-scale pilot test, of course, is
scheduled at the Rapid City pilot plant.
Desulfurization and Sulfur Recovery Steps
General
An experimental program to test these
steps is now under way in our library bench-
scale unit as specified in our present EPA
contract.
m-6-6
-------
Preliminary results are now available for
desulfurization of simulated producer gas at
about 1600°F by means of half calcined
dolomite and for regeneration at 1300°F by
means of the "Squires" reaction. These will be
discussed below.
The offgas from the Squires reaction at the
specified operating conditions (1300°F) is
relatively low in H2S content due to equili-
brium limitations. Special techniques are
required to recover sulfur from this gas
economically without condensation of steam
or removal of carbon dioxide. For this pur-
pose, a liquid-phase Claus reaction was
proposed using hot water under pressure as
the reaction medium. This is the so-called
"Wackenroder" reaction, and the system is
described in detail in the Annual Report.2
Laboratory equipment to test this process is
now being assembled but no results are as yet
available to report.
Description of Experimental System
A flow diagram of the new experimental
unit is shown in Figure 2. Acceptor in the form
of CaCOs'MgO is fed continuously at a known
rate to the top of the H2S-sorption reactor via
a pneumatic lift line. The carrier gas is
recycled product gas and it does not pass
through the fluidized bed in the sorption
reactor. Hot, H2S-laden, producer gas is fed to
the bottom of the bed. Steam, N2, CO2, H2,
and H2S are added to a stream of recycle gas
to simulated the partial pressures of the
various components of the product gas from
the gasifier shown in the process flow diagram
in Figure 1. The reactor, previously used as the
CO2 acceptor regenerator, is 3-in. ID with a
bed height of 18 inches.
The sulfided acceptor is fed by gravity to
the top of the Squires reactor and is regen-
erated while being fluidized in a stream of
steam-CO2. The reactor, previously used as
the gasifier vessel, has been necked down from
4-in. to 2-in. ID and has a bed height of 48
inches.
Continuity of acceptor recirculation is
maintained by withdrawal and feeding
through parallel lockhoppers as shown in
Figure 2. In both reactors the acceptor is fed
to the top of the bed and is withdrawn from
the bottom. The height of each fluidized bed is
held at the desired level by means of a AP cell
placed across the upper part of the bed, which
actuates a solids control valve located below
the acceptor stand leg.
The product gas from either reactor can be
monitored continuously for H2S content by
means of a dualprange infrared analyzer. The
continuously recirculating inventory of ac-
ceptor is sampled periodically from both
reactors and analyzed for CaS and
Note that the physical arrangement of the
two reactors is reversed from that shown in
Figure 1, for the process reactors. From an
experimental standpoint, it is immaterial
which reactor is the upper vessel. It is possible
that regeneration of the acceptor by the
Squires reaction will require a greater ac-
ceptor retention time than that in the H2S-
sorption reactor. The existing pressure shell
and electrical furnace for the lower reactor is
considerably larger than that for the upper
reactor. It was chosen to house the Squires
reactor because the bed volume can be easily
increased by a factor of 3 over that of the
present design if initial operations show the
need.
Preliminary Experimental Results
A series of experiments were carried out, all
with conditions similar to those listed in
Tables 3 and 4 for the gas desulfurizer and
regenerator, . respectively.
The first series of runs were conducted with
Tymochtee dolomite which was used in the
previous work on the CO2 acceptor process.6
A fundamental difficulty found here is the
very high attrition rate, which ran as high as
18 percent of the acceptor fed per pass.
Run A4 (Tables 3 and 4) was made with air
injection into the regenerator in an attempt to
harden the stone by partial oxidation to
in-6-7
-------
TableS. GAS DESULFURIZER
Table 4. REGENERATOR
Acceptor
Feed rate, Ib/hr (half calcined basis)
Solids residence time, min
Input, scfh
Recycle to bed
H2S
CO2
H2
H20
N2
Purges (CO2lto bed
Purges (N2) above bed
Recycle, acceptor lift gas, above b
Output, scfh in cycle
Exit gas rate, scfh (dry basis)
Composition, mole %
H2
CO
C02
Nl2 (by difference)
H2S
Outlet gas, top of bed
composition, mole %
H20
»2
CO
C02
N2
H2S
Flow rate at top of bed , scfh
Fl-uidizing velocity, ft/sec
Attrition, % of feed rate
Duration of circulation with H2S feed.hr
Removal of feed sulfur, %
% H2S in outlet/equilibrium %H2S
Conversion of acceptor/pass, mole %
Run number
A4
35x48 mesh
Tymochtee
dolomite
5.9
32
130
3.5
35
60
—
65
5
15
ed 92
1-3
148
18.3
17.3
12.1
52.1
0.09
10.9
16.3
15.4
10.8
46.5
0.08
275
0.40
5.6
7.1
97
2.3
23
Temperature = 1600"F, Pressure = 206psig
A7
28x35 mesh
Canaan
dolomite
6.5
33
175
3.5
54
73
—
96
5
15
71
1-2
215
17
18
12
53
0.05
9.7
15.4
15.9
11.2
47.7
0.04
416
0.60
0.7
25.2
97
1.4
19
CaSC>4 in situ. Prior work in the CO2 acceptor
process development showed that hardening
of the stone occurs when this is done at higher
temperatures due to the formation of a
transient liquid in the CaS-CaSO4 system.6
The attrition rate apparently was somewhat
reduced over the comparable run where air
injection was not used but was still unac-
ceptably high. The operating temperature
(1300°F) was apparently too low to achieve
hardening by the transient liquid mechanism.
A Tymochtee dolomite which had been
hardened by cycling through the CO 2 acceptor
process was also tested. This material showed
the expected good resistance to attrition. Only
preliminary results are available at this
writing, but the activity of the stone appears to
Solids residence time, min
Input, scfh
H2O
COa
Air
H2
Purges (N2) to bed
Purges (N2) above bed
Purges (C02) above bed
Output, scfh in cycle
Exit gas rate, scfh (dry basis)
Composition, mole %
C02
N2
H2S l
S2
Outlet gas, top of bed
Composition, mole %
H20
C02
N2
H2S
S2
Flow rate, top of bed^scfh
Fluidizing velocity, ft/sec
Regeneration of acceptor, mole %
Run number
A4
38
89
82
7.5
—
10
10
5
1-3
110
76.1
23.0
0.8
0.05
43.5
47.3
8.6
0.5
0.03
186
0.53
6.7
A7
37
110
110
—
—
8
10
5
1-2
133
85.7
13.5
0.7
—
48.3
47.8
3.5
0.4
—
228
0.64
5.1
Temperature = 1300°F, Pressure = 206 psig
be less than that of the fresh Tymochtee
dolomite.
A Nebraska limestone was also tested; it
shoi:"*H very low absorption of H2S as com-
pared with half calcined dolomites. This is in
accord with prior laboratory studies by Ruth,
et al. Other limestones such as BCR-1692
used in the Esso, Ltd. work will be tested to
determine whether the low activity is generic
to all limestones.
IH-6-8
-------
A pure dolomite from Canaan, Con-
necticut, was also tested under the conditions
indicated in Table 3 and 4. This material
showed excellent attrition resistance with an
average weight loss of only 0.8 weight percent
per pass through the system.
Both the Tymochtee and Canaan dolomites
showed excellent fresh activity for removal of
HaS in the gas desulfurizer. Ninety-seven
percent removal of H2S was achieved in both
cases. The removal of CaS in the regenerator
was highly incomplete in both cases. Thus, it
would appear that the kinetics of the Squires
reaction is a limiting factor in this system and
this factor requires more study.
The definition of results in terms of
number of cycles is difficult because of the
semi-continuous nature of the acceptor circu-
lation, loss of material due to attrition, and
sampling of material for analyses. An ap-
proximate, and "conservative" method for
calculating the number of cycles was used.
Each pass of a charge through the unit is
calculated as a fractional cycle An using the
relationship: An equals the actual amount fed
divided by the internal inventory plus the
actual external inventory.
The number of cycles that could be
achieved with the Tymochtee dolomite were
limited due to the high attrition losses. In the
case of the Canaan dolomite, attrition was not
the determining factor; the run was ter-
minated when breakthrough of H2S occurred
in the gas desulfurizer.
The composition of the offgases from the
gas desulfurizer and regenerator as a function
of the number of cycles for runs A4 and A7 are
given in Figures 3 and 4, respectively. In both
cases the H2S content of the regenerator off-
gases increases with the number of cycles.
This, as we will show below, is due to accumu-
lation of CaS on the acceptor and points to
poor kinetics in the Squires reactor. The E^S
content of the desulfurizer offgas remains
relatively steady until the acceptor becomes
heavily loaded with sulfur towards the end of
the run.
The sulfur content of the stones for the two
Runs A4 and A7 as a function of cycle number
are given in Figures 5 and 6, respectively. It is
immediately clear that in both cases there is
rapid buildup of CaS on the acceptor due to
incomplete regeneration of the CaS. The
efficiency of sulfur removal and recovery is
further illustrated by the tabular data
presented for Run A7 with Canaan dolomite
in Table 5.
At 19 percent conversion per pass to CaS
(Table 3, Run A7), the known percent of sulfur
rejected (Table 5) makes possible a rough
estimate of the Ca/S ratio required if fresh
dolomite feed were added continuously to the
system operated under the above conditions.
The Ca/S ratio would be approximately 0.45.
Tables. RUN A7 H2S CONTENT OF EXIT GASES (DRY BASIS)
CANAAN DOLOMITE
Gas desulfurizer
Cycle No.
0.3
1.4
5.5
7.5
8.6
12.6
H2S,
mole%
0.046
0.049
0.048
0.058
0.064
0.672
H2S
Removal, %
97
97
97
96
96
59
Regenerator
Cycle No.
0.5
1.5
5.7
7.8
9.2
12.5
H2S,
mole%
0.671
0.927
1.39
1.69
1.40
1.90
Recovery of
H2SFeed,%
25
35
53
64a
53a
72 a
The condensate contained about 3% of the feed «ulfur as elemental sulfur.
III-6-9
-------
The acceptor at the end of Run A7 was
nearly completely converted to CaS, hence the
break-through observed in the desulfurizer. At
the end of the run, 2 hours additional
residence time was given to the batch of ac-
ceptor remaining in the regenerator. The CaS
content of the acceptor was reduced from 85 to
76 mole percent. This indicates again that the
poor kinetics in the Squires reactor is con-
trolling. Better results should be achieved if
longer residence times in the regenerator are
used. It does appear that a considerable
amount of inactive CaS is inevitably formed;
at the present time, however, we are unable to
clearly distinguish between CaS of low
reactivity and the "dead" material.
It thus appears that high purity crystalline
dolomites have acceptable physical strength
and activity for use in the process.
Economically acceptable make-up rates can
be achieved at the proper operating con-
ditions. The geographical distribution of
dolomites with acceptable strength and ac-
tivity is now being studied. A variable study is
also planned with selected dolomites to
determine optimum conditions for their use.
ECONOMICS OF PROCESS
No economic figures are available for the
improved process. The potential economics of
the CO2 acceptor based process were given in
the Annual Report.2 Figure 7 is reproduced
from that report. It gives the cost of low-Btu
gas as a function of coal cost delivered to a
1200-MW boiler from a large-scale gasifi-
cation plant. The figures are based on 1976
operation, 15 percent capital charges and 7.5
percent/yr escalation on materials and labor,
and 7.5 percent/yr interest during con-
struction and an operating factor of 70 per-
cent.
The economics of the new process are
expected to be somewhat better.
ACKNOWLEDGMENT
Appreciation is expressed to the Environ-
mental Protection Agency for financial
support of the work presented in this paper
and for permission to publish the results
given.
BIBLIOGRAPHY
1. Curran, G.P., C.E. Fink, and E. Gorin.
Proceedings of Second International Con-
ference on Fluidized-Bed Combustion,
1970. Office of Air Programs, Environ-
mental Protection Agency, Research
Triangle Park, N.C. Publication Number
AP-109. pp. III-l-l to HI-1-11.
2. Curran G.P., J.T. Clancey, C.E. Fink, B.
Pasek, M. Pell, and E. Gorin. Annual
Report to Control Systems Division, Office
of Air Programs, Environmental Protec-
tion Agency, Research Triangle Park,
N.C., under Contract Number EHSD-71-
15. September 1, 1970-November 1, 1971.
3. Squires, A.M. Fuel Gasification. Advances
in Chemistry Series No. 69. American
Chemical Society. Washington, D.C.,
1967. pp. 205-229.
4. Curran, G.P. and Everett Gorin. The CO2
Acceptor Process — A Status Report.
(Presented at 3rd International Confer-
ence on Fluidized-Bed Combustion.
Hueston Woods. October 29-November 1,
1972.) See Session 3, Paper 5 this volume.
5. Book 1, Studies on Mechanics of Fluo-
Solids Systems. Consolidation Coal Co.,
Research & Development Report No. 16 to
the Office of Coal Research, U.S. Dept. of
the Interior, Washington, D.C., under
Contract Number 14-01-0001-415.
Government Printing Office Catalog
Number 163.10:16/INT3/Book 1.
6. Book 3, Operation of the Bench-Scale
Continuous Gasification Unit. Consolida-
tion Coal Co., Research & Development
Report No. 16 to the Office of Coal
Research, U.S. Dept. of the Interior,
Washington, D.C., under Contract Num-
ber 14-01-0001-415. Government Printing
Office Catalog Number 163.10:16/INT3/-
Book 3.
ra-6-io
-------
7. Goring, G.E., G.P. Curran, R.P. Tarbox, 10. Book 2, Laboratory Physico-Chemical
and E. Gorin. Ind. Eng. Chem. 44:1051, Studies. Consolidation Coal Co., Research
1057, 1952. & Development Report No. 16 to the
Office of Coal Research, U.S. Dept. of the
8. Goring, G.E., G.P. Curran, C.W. Zielke, Interior, Washington, D.C. Government
and E. Gorin, Ind. Eng. Chem. 45:2586, Printing Office Catalog Number
1953. 163.10:16/INT3/Book 2.
11. Ruth, L., A.M. Squires, and R.A. Graft.
9. Zielke, C.W. and E. Gorin. Ind. Eng. Environmental Science and Technology.
Chem. 47:820, 1955, and 49:396, 1957. November 1972.
m-6-ii
-------
O3
COAL
PRODUCT GAS
WITH H2S
GASIFIER
CHAR
STEAW
"SQUIRES"
REACTOR
SULFUR '
REACTOR
H2S
SORPTION
1 CaS- MgO
H2S
STEAM
C02
SULFUR
RECOVERY
I
SULFUR
CLEAN
PRODUCT
GAS
1300 °F
. H20
Figure 1. Two-stage fluidized-bed partial combustion process.
-------
DPCV
VENT
WATER
UDEMINER
11 ALIZER
CONDENSER
RECEIVER
DRY GAS
'METER
H2S SORPTION
REACTOR
ACCEPTOR
STAND LEG
TOGAS
ANALYZER VV
MAKE-UP
ACCEPTOR
RECYCLE
COMPRESSOR
ACCEPTORr*lF"-TERS
CHARGE POT
CONDENSER
RECEIVER COOLER
cu
, BALANCE
H, No CO, H,S GAS
HSOLIDS
CONTROL
.VALVE
VENT A,TFRTnpVENT EMERGENCY
• ACCEPTOR i SOLENOID
ACCEPTOR
VFNTSAMPLE
VEN.T TUBE ^
DEMINER-
ALIZER
DRY GAS
METER
ACCEPTOR
STAND LEG
ACCEPTOR
FEED
HOPPERS
WATER FEED
PUMP
BALANCE
GAS
FEEDER SYSTEM
"SQUIRES"
REACTOR
COIVTPRESS^R
ACCEPTOR WITHDRAWAL
HOPPERS
Figure 2. Flow diagram of new experimental unit.
-------
* 1-4
1.2
1.0
o4
0.6
REGENERATOR
1 2 3
CYCLE NUMBER
0.12
0.11
0.10
0.09
£• 0.0
0.07
C/9
CS1
0.06
0.05
0.04
GAS OESULFURIZER
4 0
Figure 3. Run A4-tymochtee dolomite.
1 2 3
CYCLE NUMBER
III-6-14
-------
0.07
0.06
* 0.05
o>
•I" 0.04
_c
6*
£ 0.03
E
IS)
* 0.02
0.01
0
14
/ TO 0.67 at 12.6 cycles
GAS DESULFURIZER
4 6
CYCLE NUMBER
8 10
Figure 4. Run A7-canaan dolomite.
12 14
III-6-15
-------
90
80
70
60
E
5
1 50
JD
O
<3
+
00
40
30
20
10
GAS DESULFURIZER
1 2
CYCLE NUMBER
90
80
70
60
S 50
8 40
s
TO
o
30
20
10
REGENERATOR,
1 2
CYCLE NUMBER
Figure 5. Run A4-molar sulfur content of exit solids.
III-6-16
-------
4 6 8 10
CYCLE NUMBER
12
4 6 8 10
CYCLE NUMBER
Figure 6. Run A7-CaS content of acceptor.
80
S
o
cx>
<
CJ
70
60
28 30 32 34 36 38 40
COAL COST,
-------
SESSION IV:
Conceptual Designs and Economics
SESSION CHAIRMAN:
Dr. D.H. Archer, Westinghouse
IV-0-1
-------
1. SMALL-SCALE APPLICATIONS OF FLUIDIZED-BED
COMBUSTION AND HEAT TRANSFER
D. E. ELLIOTT AND M. J. VIRR
University of Aston, Birmingham, England
Research on fluidized-bed combustion has
primarily been aimed at improved economics
and anti-pollution measures for coal- and oil-
fired power stations or for relatively large
packaged boilers. At the last Hueston Woods
Conference, however, the Keynote Address
hinted that advantages might be gained from
applying fluidized-bed combustion and heat
transfer techniques even on very small
systems.
In the two years since then, considerable
progress has been made in research to identify
and solve problems associated with small-scale
application; a small development company
has been started to exploit areas of likely com-
mercial interest.
As some of the work involved may give
feed-back to large boiler technology, this
paper reviews the state of the art and gives
data on combustion and heat transfer.
GAS-FIRED
COMBUSTION
FLUIDIZED-BED
Up to now, research at the University of
Aston Mechanical Engineering Department
has mainly concentrated on gas firing. That
gas can be burned successfully in deep fluid-
ized beds is well-known. Reference 1 describes
Russian work. Reference 2 cites French
research. The Coal Research Establishment
(NCB) used gas combustion in the early days
of fluid-bed coal burning work to investigate
aspects of volatile burning. Most of this work
was done in beds 1-ft deep or more; it
required the use of relatively high pressure
blowers, therefore necessitating either high-
speed motors or some form of multi-stage
rotors. In a practical plant the former solution
would result in noisy appliances, while the
latter incurs high manufacturing costs. A
further disadvantage with deep fluidized beds
when applied to small heat inputs is that the
total surface area from which heat can be lost
is high in relation to the throughput, so .that
beds with an L/D ratio of more than about 0.5
have to be surrounded with a high degree of
lagging to prevent undue heat losses during
startup.
Research was, therefore, initiated to find
the minimum bed depth which would give
stable and efficient combustion. Because of
the very poor lateral gas mixing in fluidized
beds, the idea of introducing separate gas jets
into an air fluidized bed was discarded; the
bed depth needed for complete mixing and
good combustion would be too great unless an
extremely large (and costly) number of gas jets
were used. The experiments were conducted
with pre-mixed gas/air mixtures fed through a
porous ceramic distributor into a bed of silica
sand (Figure 1). Provided that the gas/air
mixture is initially within the flammability
limits (2.2 to 9.5 percent by volume for the
propane normally used) and the initial fluid-
izing velocity is less than the flame propaga-
tion speed, then such a bed can be ignited by
simply lighting the gas/air mixture above the
surface of the bed. Whether or not the system
heats up to obtain controlled fluidized-bed
IV-1-1
-------
combustion, however, depends upon having
suitably sized solids particles, choosing an
appropriate heat input rate for the system,
and preventing undue heat loss at startup.
Various stages of startup on a bed with cor-
rectly chosen conditions (1-in. deep, 6-in.
diameter, 18,000 Btu/hr input, 10 percent
excess air) are as follows:
1. Air is blown through the bed at a rate
which is around the incipient fluidizing
velocity. The bed surface is hardly
disturbed. A near stoichiometric ratio of
gas is then admitted.
2. On ignition, the gas/air mixture burns at
the top of the bed with a blueish flame,
which dances around in an irregular pat-
tern as it is not attached to any particular
stabilizing system.
3. Particles which had been thrown up from
the top surface of the bed into the flame
and heated, now return to the bed and
carry the heat down.
4. The gas/air mixture now becomes
preheated as it passes into the bed and
burns more readily at the surface with a
distinct popping, roaring noise. The pre-
heating causes the fluidizing velocity auto-
matically to increase; it also causes more
particles to be thrown up, which in turn
maintains the rate of temperature rise in
the bed. The flame structure modifies, and
red-hot particles begin to tinge the colour
of the flame.
5. When the bed temperature reaches 600°C,
combustion begins which glows dull-red
behind the reddish-blue flame. Noise level
increases. By now the fluidizing velocity is
some three times greater than the initial
cold fluidizing velocity and more bubbling
takes place. Combustion in the bubbles
becomes more" violent; mild explo-
sions/detonations occur which tend to
throw up far more particles from the sur-
face of the bed.
6. When the bed temperature has reached
700° C, much of the combustion occurs in
the bed; by 800°C only a small proportion
of the gas burns at the top. The noise level
is now at a peak, probably indicating that
the combustion is occurring after the
gas/air mixture has formed itself into bub-
bles within the fluidized bed. The surface
splashing of particles has decreased, which
indicates that although it is suspected that
combustion occurs in the bubbles, these
bubbles are further below the surface when
combustion occurs than in the case of the
600°C bed.
7. By 900°C the noise level is reducing; at
1000° C the level has dropped several deci-
bels below its maximum level. The ultimate
temperature level reached depends mainly
on the gas and air input to the bed. Tem-
peratures up to 1200°C can be readily
maintained with silica sand, but particles
with higher fusion temperatures are needed
to go much beyond this temperature
because of sintering. The noise level is
reduced in deeper beds over 800° C and
disappear above 880° C.
Figure 2 shows a typical heating up rate for
a shallow bed combustor; Figure 3 shows the
noise level spectrum emitted from the bed at
various temperatures. It is believed that lower
noise levels occur at the higher temperatures
because combustion is extremely rapid in the
first few millimeters of the bed before the
gases have time to form into bubbles. Thus,
the likelihood of detonations occurring is
reduced because the combustion is quenched
by the large number of particles present.
STABILITY
Once a temperature above 800°C has been
reached, the system can be operated over a
wide range of gas/air mixture strengths well
outside the normally accepted limits set by
flame propagation phenomena. The fluidized
bed operates as a very effective pre-heater,
bringing the incoming gas up to bed tempera-
ture within the first few millimeters of the bed.
rv-i-2
-------
Provided the external heat losses from the bed
are small, achieved by lagging and by placing
reflecting surfaces above the bed to radiate the
heat back down to the bed, extremely weak
gas/air mixtures can be burned.
Combustion can be maintained with ex-
tremely shallow beds (below 0.5-in.), but it is
not yet clear whether or not complete temper-
ature equilibrium between the exit gases and
the solids is achieved. With very shallow beds,
some combustion probably takes place after
the gases leave the bed. For beds 0.5-in. deep
and above the exit gases appear to be more or
less in temperature equilibrium with the bed,
and combustion efficiency is excellent with
CO/CO2 ratios dropping below 0.002 (a factor
of ten better than the standards insisted upon
by the UK Gas Council for domestic appli-
ances).
Because of the quenched low temperature
combustion it was expected that NOX produc-
tion would be low. This has been borne out by
samples drawn through Dager tubes
measuring NO + NO2 which indicated less
than 5 ppm in the exhaust gases. These figures
represent a considerable reduction compared
with emission from normal flames.
It is interesting to note that the combustion
intensity of these beds is in the region of one
million Btu/ft3-hr based on the full depth of
the bed. As it is likely that most of the com-
bustion takes place in the bottom half of the
bed, the actual combustion intensity must be
at least twice this rate. A similar bed, but with
heat transfer by direct contact between the
particles and cooling surfaces, can be operated
at more than twice the above heat release rate;
figures of 3 x 106 Btu/ft3 -hr have already been
achieved.
The ease of operation of this form of gas
"burner" in the temperature range of 800 to
1200°C, combined with the excellent heat
transfer which occurs when small objects are
placed in the bed makes possible the use of
such furnaces for laboratory, workshop, and
factory metallurgical processes, e.g., har-
dening, annealing, heating small billets prior
to forging etc. Figure 4 shows a typical heating
curve for a 3-in. long x 1/4-in. diameter alloy
steel bolt immersed in a fluid-bed combustor
operating at 960°C (the extra heat absorbed
during the transformation zone is clearly
seen). With these possibilities in mind a small
company, Fluidfire Development Limited, has
designed and built a range of furnaces:
1. 6-in. diameter x 1-in. deep self-contained
units for laboratory investigations and
demonstration work.
2. 6-in. diameter x 6-in. deep units again for
laboratory work which contain a two-stage
blower or can be used from a shop air
supply.
3. An 8-in. diameter x 8-in. deep metallur-
gical furnace with automatic temperature
regulation for use in small quantity har-
dening and annealing work.
4, A larger 12-in. diameter x 12-in. deep
metallurgical unit, again fully temperature
controlled, shown in Figure 8.
The metallurgical furnaces which can
operate over a temperature range of from
about 700 to 1200°C can replace traditional
salt bath furnaces for hardening and tem-
pering a wide range of materials. The fluidized
furnace has the following advantages:
1. Higher operating efficiency allows lower
fuel costs per pound of metal processed.
2. The cost of special salts and the difficulty
of handling and disposing of the spent salt
are eliminated.
3. The furnaces have a short startup time and
therefore can be switched off overnight.
4. The atmosphere in the heating zone can be
adjusted to suit the requirements of treat-
ment. It is usually made to be reducing.
5. The furnaces can operate over a wide range
of temperatures and with slight modifi-
cations can be changed to carburising
duties in which case they eliminate the use
of cyanide with all its attendant safety and
disposal .problems.
IV-1-3
-------
6. All types of steels may be heat treated.
The performance of these units with
respect to operating cost, heating rates, and
oxidising rates at various temperatures and
the hardness achieved under various condi-
tions is described in reference 3.
A special unit has also been designed to
operate in line with automatic continuous pro-
duction of hardened steel components. This
unit uses a continuous wire mesh belt to
support and convey articles through two
separately controlled fiuidized- bed combus-
tors: the first higher temperature combustor is
the heating zone; the second combustor,
running at a precisely controlled temperature,
allows a short soaking period and ensures that
the articles are fed into the quenching system
at the correct temperature.
RADIANT HEATERS
Shallow-bed combustion systems of this
type are interesting in their own right since
they are a new way of making more effective
radiant gas heaters. Normally, radiant gas
heaters rely on convective heating of ceramic
plaques by very hot gases and the re-radiation
of the heat from the plaques to the
surroundings. The exit gases are at a substan-
tially higher temperature than the radiation
surfaces; the overall effectiveness is generally
such that only 25 to 35 percent of the input
heat is radiated.
The immense surface area exposed to the
combustion gases in a fiuidized bed allows the
temperature differential between the gases
and the solids to be negligible. Thus, a bed
operating stoichiometrically at 1000°C will
radiate approximately 50 percent of its heat
away from the bed. The operating tempera-
ture/radiation efficiency of an ideal bed (i.e.,
perfect burning of the air and gas before
leaving the bed and thermal equilibrium
between the gases and the solids) can be
readily calculated. The radiation efficiency is
given by the expression
n = radiation = 1 -
ili
H
where: ht is the enthalpy of the exhaust gases
at the bed operating temperature, and H is the
calorific value of the fuel burned.
There are only two sources of heat loss from
the bed, the enthalpy of the exhaust gases and
the radiation from the bed. The external
convective heat losses are negligible if the
height of the containment wall is low. Any
heat directed downwards to the distribution
plate is returned to the combustion bed as pre-
heat in the gases.
Figure 5 plots the radiation efficiency
versus bed temperature for various gases. It
will be noted that the radiation efficiency for
stoichiometric mixtures is not very different
for the various gases and is not a function of
the bed emissivhy which only affects the rating
of the bed per unit area. Figure 6 shows how
the heat input and the radiant output of-a-6-r
in. diameter shallow-bed combustor varies as
the temperature of the bed changes for stoi-
chiometric gas/air mixtures. The full curves
correspond to a bed of unit emissivity and the
dotted curves to a bed of emissivity 0.7.
It will be noted that as the bed operating
temperature is lowered by reducing the gas/air
volume fed to the bed, the radiation efficiency
increases. Above about 800°C the small fiuid-
ized beds which have been produced perform
very nearly as shown provided that due ac-
count is taken of the effective emissivity from
the bed surface. It was originally expected that
the emissivity of the granulated surface of a
gently bubbling fiuidized bed would act very
nearly as a black body for most solids. But this
does not appear to be the case; the emissivity
for silica sand is approxmately equal to that of
silica sand itself.
Below 800° C the radiation effectiveness is
not as good as predicted; either thermal
equilibrium is not established or some gas is
leaving the surface unreacted.
Although the emissivity of a gently bub-
bling bed falls into line with the individual
particle emissivity, the overall effective radia-
tion from a fluidized-bed combustor using
IV-1-4
-------
very fine particles may be significantly higher
than this. As might be expected, the influence
of a particle cloud above a fluidized bed
materially affects its radiation characteristics.
This effect exists because the particles during
their stay in the gas space above the bed
radiate their heat very rapidly, cool down to a
temperature below that of the off-gases, and
thereafter tend to act as a second-stage cooling
medium for the gases leaving the bed. Thus, it
should be possible to operate a fluidized-bed
combustor at a temperature which is signifi-
cantly more than the temperature of the gases
leaving the system. In this case, the cloud of
particles above the bed shows a duller colour
than one would normally expect from a bed
operating at the same temperature. Research
into this phenomena is underway at Aston.
It will be seen from Figure 6 that the heat
output of a 6-in. diameter bed operating at
1000°C is approximately the same as that of a
traditional British radiant gas fire. This has
led to the concept of using shallow fluidized-
bed combustors as room heating devices. The
constantly varying pattern of the fluidized
bed, coupled with the similarity to the open
coal fire, was thought to offer an attractive
alternative to the traditional fire. Thus, self-
contained units incorporating brushless
electric motors, fans, controls, and safety
devices are now being developed by Fluidfire
Development Limited in order to assess the
potential of such appliances. Although much
more work is still needed to satisfy stringent
safety requirements, many problems inherent
in such a radically new system have been over-
come, and there appears a fair chance of
success.
NOISE
One of the requirements for room heaters is
that they should not generate a great deal of
noise, and the UK Gas Council recently pro-
posed a standard of acceptable noise levels
shown by the dotted line on Figure 3. It will be
seen that the noise level from an open radiant
bed is slightly higher than the permitted noise
levels. Thus, to quiet the fire somewhat and
also to lessen the danger of having a com-
pletely open fire, a glass screen could be incor-
porated in front of the fire. Tests with ceramic
glass show that this reduces the noise level to
well below the acceptance levels and ensures
that clothing cannot be ignited by direct con-
tact with the fire.
BOILER APPLICATIONS
It is obvious that if a radiant fluidized-bed
combustor operating at a temperature level of
800 to 900°C is surrounded by a water jacket,
then 55-60 percent of the heat input will be
radiated to the water walls, even if the fluid-
bed particles do not come into contact with the
walls. However, the rating of such a boiler
would be relatively low—a 6-in. diameter bed
having an output of something like 15,000
Btu/hr. Hence, the unit would not be particu-
larly compact and could not be considered as a
viable commercial alternative to the highly
rated gas-fired boilers which are now being
produced.
If an attempt were made to place heat
transfer tubes in the fluidized-bed combustor
in a similar manner to the way in which large
fluidized boilers have been designed, then the
simple startup procedure described earlier
would not be effective, and compartmentation
of the bed for startup purposes would be
necessary. For small-scale appliances, this
would be prohibitively costly. An alternative
was sought; the idea of locating heat transfer
surfaces just above and around the settled
fluidized bed was formulated and has been
successfully developed. This solution relies on
the principle that the expansion of a shallow
fluidized bed (as a percentage) is very high
compared with a deep fluidized bed. Thus on
startup with cold air the bed expansion is very
small, and the particles do not contact the
heat transfer surfaces. If a gas/air mixture is
then lit above the bed, the bed heats up in a
similar fashion to the radiant bed described
above; but by the time a temperature of 700 to
800 °C has been reached, the fluidizing
velocity is some three to four times the initial
velocity and the bed has expanded to contact
IV-1-5
-------
the heat transfer surface. Further contact
takes place by virtue of splashing of particles
from the bed upwards and sideways.
For beds of small output (up to 100,000
Btu/hr it is sufficient to surround the bed with
a water-cooled wall which is insulated from
the settled bed.
The direct contact between the particles
and the cooling surfaces allow the heat input
to be two to three times that which could be
sustained in a bed which was only cooled by
radiation. The principle could still hold good
for larger outputs, but in this case some
additional heat transfer surface would have to
be placed in such a position that it contacted
the expanded bed and received splashing heat
transfer.
Because the fluid-bed combustor will not
operate very satisfactorily below 800 °C, and
even if we could exploit father the principle of
particle cloud radiation to cool the off-gases to
below bed temperature, the overall efficiency
of the combustion system in transferring heat
to water would be too low for domestic central
heating systems. Some form of second-stage
heat recovery is therefore necessary. The
incorporation of convective heat transfer
surfaces would leave the system with many of
the disadvantages of normal systems, e.g.,
large heat transfer volumes or the use of high
extended surfaces with the possibility of
condensation and corrosion troubles. A
second-stage shallow fluidized bed was there-
fore incorporated above the combustion bed of
a trial 40,000 Btu/hr laboratory unit which
was supplied by air from an external source.
With a 1-in. bed depth and an 8-in. diameter,
the heat transfer area of about 1/6 square foot
around the periphery of the bed allowed the
existing gases to reduce in temperature to
400°C, giving an overall efficiency of about 80
percent.
The addition of a number of thick fins to
the walls increased the surface area to 1/2
square foot and enabled the gas temperature
to be reduced to less than 250°C, giving an
overall efficiency of about 87.5 percent.
An important point is that these high effi-
ciencies are achieved without operating with
metal temperatures below about 110°C. This
is possible because of the very high heat
transfer coefficients between the fluidized
solids and the metal surfaces. These coeffi-
cients result in the significant advantage that
condensation of the exhaust gases on the fins
does not occur and corrosion is eliminated.
This is not the case if highly extended surfaces
are used in normal convective heat transfer.
Figure 9 shows a prototype domestic fluid-
bed boiler incorporating blowers, controls,
safety cut-outs, etc. This latter unit is now
carrying out endurance trials to determine the
rate of loss of particle mass during extended
running.
With regard to particle life, experiments
have already been conducted at room temper-
ature using sand in a fluidized bed operated at
2 ft/sec; no measurable attrition loss occurred
in 500 hours. Whether or not the continual
heating and cooling of particles in a combus-
tor with the consequent thermal shocking will
produce more severe attrition is as yet
unknown.
Apart from the very effective heat transfer
in the second-stage bed, a further significant
advantage is that it produces an extremely
effective silencer for the combustion system.
Provided, therefore, that particles to resist
degradation can be found, it appears likely
that this type of approach can provide central
heating units with the following characteristics:
1. Very low NO x emission « 5 ppm).
2. Very low CO/CO2 ratios.
3. Lower aldehyde formation than for normal
flames due to the low temperature combus-
tion.
4. Efficiencies as high as 90 percent without
significant extra cost and with no fear of
condensation or corrosion.
5. Very compact plant with overall ratings of
100,000 Btu/ft3 obtained (including fans,
motors and gas controls).
IV-1-6
-------
6. Use of well-established cast iron techniques
which are cheap and produce long life, low
maintenance units, to produce boilers.
7. Easily adjustable units for all gases and
Wobble irrelevant number and flame
speeds.
8. Good turn down ratio of units without
losing efficiency.
9. Units readily developed to burn oil and
with a little development could possibly
burn solid fuel, (but could still be switched
back to gas easily).
10. Application of techniques to visible
fire/back boiler systems providing a central
focus in the living area as well as full house
heating.
SHALLOW BED HEAT TRANSFER
Early studies by the Central Electricity
Generating Board showed that uneconomical
high pressure drops would be incurred if an
attempt were made to use plain tubes in a
fluidized bed as a straightforward heat
recovery system (i.e., one not using a combus-
tion reaction). Early work by Petri et al.4
showed that if a tube were provided with an
extended surface finning system with an area
15 times that of the base tube, then the overall
heat transfer would be increased approx-
imately six-fold, i.e., an effectiveness of 40
percent. As the overall tube diameter would
not be more than doubled by employing the
fins, the net effect would be a far greater heat
flux per unit volume of bed. Thus, the
pressure drop penalties when using extended
surfaces would be significantly reduced.
Following up this idea, a new form of
extended surface heat exchanger was built,
and some preliminary results were presented
to the Second International Conference on
Fluid-Bed Combustion. These preliminary
results were obtained with the extended sur-
face systems in a comparatively deep bed with
fairly large particles. Further work showed
that if this particular arrangement of vertical
fins was placed low down in a fluidized bed, its
performance was unexpectedly higher. Figure
7 shows the performance of vertical-finned
extended surface tubing operated in a very
shallow fluidized bed compared with that for
1-in. tubing and with the results of Petri et al.
The graph plots the bed-to-metal heat transfer
coefficients against particle size. It will be
noted that the shallow bed results lie above the
generally accepted heat transfer coefficient
line. These results do not necessarily imply
that the vertical surfaces have an effectiveness
of over 100 percent, although they may in
some instances, but do show that shallow-bed
performance is superior to that of deep beds, to
which most of the world's data on heat trans-
fer relate.
The reason for the superiority of the partic-
ular configuration of vertical surfaces in
shallow fluidized beds is believed to be the
absence of large bubbles in the system. The
absence is partly because the vertical surfaces
prevent lateral mixing of the gases which
again restricts bubble formation. It is
interesting to observe that when these
extended surface tube bundles are placed in a
shallow bed they do not appear to disturb the
bubbling pattern.
A further possible explanation for the
improved heat transfer is that the viscosity of
shallow fluidized beds varies almost directly as
the bed depth. Shear stress/shear strain data
derived from a Stormer-type viscometer with a
hollow cylindrical rotor is given in reference 5.
Because of the better fluidization in shallow
beds, the resistance to shear of the fluidized
solids is much less. As we know that fluidized-
bed heat transfer depends upon the rate of
exchange of particles at the heat transfer
surfaces, it would be logical to expect that heat
transfer would improve if particle mobility
improves. Thus, shallow beds would be
expected to be superior to deep beds from a
heat transfer aspect.
Coupling extended surfaces with extremely
shallow beds has been patented with the
concept that we need no longer regard
fluidized beds as isothermal devices. This
enables the overall thermal effectiveness of a
IV-1-7
-------
fluidized bed to be improved by the correct
design of the heat transfer bundle/distributor
unit.
As pointed out at the last conference here,
in contrast to conventional convective
extended-surface heat transfer systems where
the improvement in heat fluxes per unit
volume is accompanied by a higher pressure
drop, the use of fluid-bed extended surfaces
increases the heat flux per unit volume and
decreases the pressure drop. The pressure
drops in some of the shallow-bed units we have
investigated are so low that the system can
now compete favourably with normal convec-
tive heat transfer, even for gas turbine waste
heat recovery where low pressure drops are of
paramount importance. It is contemplated
that a two or three stage fluid bed can be
operated with an overall pressure drop of less
than 12 inches water gauge.
Theoretical and experimental studies of the
mechanism of heat transfer between the fluid-
ized solids and the vertical fins and along the
fins themselves are underway; a vast amount
of data has been obtained for various proprie-
tary extended surface tubing as well as for
specially designed fin/tube arrangements.
These data, which are at the moment being
written up for presentation in the near future,
are sufficiently complete to allow an economic
appraisal of various forms of heat exchangers
to be made. Further advances in performance
are expected when a better appreciation of the
various phenomena is acquired.
It is suggested that for fluidized-bed boilers
or for combined gas turbine steam cycles
where steam is generated in the high tempera-
ture exhaust from the gas turbine, there is
already a case for investigating the use of
shallow fluidized-bed extended-surface
systems instead of normal convective heat
transfer.
Preliminary trials of a single stage unit
picking up waste heat from a diesel engine
have been very encouraging. Heat transfer
coefficients of the bed which was completely
covered by vertical fins were just as good as for
the original laboratory unit. The fluidized
solids and the fins became coated with carbon,
suggesting that there may be some possibility
of enhancing this effect to reduce pollution
from the sub-micron carbon in the exhaust
gases. The unit ulso acted as a very effective
silencer.
FUTURE RESEARCH AND
DEVELOPMENT
Initial experiments into burning distilate
oil in fluidized beds have been successful; it
seems likely that a dual fuel gas/oil system can
be developed quite quickly. The unit would
have no higher pressure'drop than the existing
gas-fired unit and, therefore, would have
advantages over normal pressure jet burner
furnaces with regard to fan pressure and noise
levels. Its pollution control level would be far
superior. It is expected that the unburnt
hydrocarbons will be much reduced compared
to normal oil flames.
Research on small-scale solid fuel-fired
units has also started; there appears to be no
insuperable problems in producing a shallow
open-hearth solid fuel fire giving radiation
outputs of over 60 percent. A high-efficiency
domestic fluid-bed solid fuel-fired boiler also
appears to be a practical proposition.
The extension of these ideas into the field
of packaged boilers is underway; it is expected
that economically viable units can be
developed.
In addition to the low pressure, hot water
boiler developments, studies of high tempera-
ture, high pressure steam systems indicate
that such units can be designed to startup and
operate satisfactorily and economically. These
units would employ high temperature alloy
tubing capable of operating dry during
startup, thus avoiding heat losses to the
cooling system during startup.A unit capable
of producing steam for a 150-hp engine would
have a diameter of approximately 2 feet with a
combustion bed 6-in. deep followed by a
further 6-in. deep economiser. With shallow
rv-i-8
-------
fluidized-bed extended surface systems the
bed depths would be even smaller.
CONCLUSIONS
Domestic fluidized-bed combustion/heat
transfer systems which are no more expensive,
are just as compact, and have far less pollution
than their conventional counterparts have
been developed.
Endurance and reliability trials are in pro-
gress so that by the end of 1972 we should be
able to assess the full potential of gas-fired
fluidized-bed combustion as applied to small-
scale boilers. The results so far suggest that
additional work should be undertaken on oil-
and coal-fired systems.
Fluidized-bed combustion and heat
transfer lends itself well to metallurgical heat
treatment processes in which the antipollution
aspects combined with rapid processing have
been shown to lead to environmental and
economic benefits. Batch processing furnaces
are now available, and in-line continuous
furnaces show distinct promise.
The development of shallow bed, extended
surface heat transfer systems promises to open
up a completely new field in heat recovery and
could help to reduce costs and space require-
ments in many types of plant.
It appears, therefore, that fluidized-bed
combustion and heat transfer techniques can
be usefully employed for antipollution
measures over the whole range for units
having an output of a bunsen burner up to
extremely large power station sizes; ,in many
cases it will be accompanied by economic
benefits.
REFERENCES
1. Ukilov, V. M., G. K. Rubtsov, and A. P.
Baskakov. Gas Combustion in a Packing
under a Fluidized Bed. Gazovaja
Promyshlennost. 5, 1969.
2. Tamalet, A. J. Application of Fluid Bed
Heat Transfer to Metallurgical Processes.
Inst. Chemical Engineering Symposium
Series. In: Proceedings Symposium on
Chemical Engineering in Iron and Steel
Industry, pp. 105-114, 1968.
3. Virr, M. J. and R. Reynoldson. Heat Treat-
ment in Fluidized Beds. Industrial Process
Heating.
4. Petrie, J. C, W. A. Freeby, and J. H.
Buckham. Bed Heat Exchangers. Chemical
Engineering Progress, 64(7), 1968.
5. Botterrill, J. S. M., M. Van der Kolk, D. E.
Elliott, and S. McGuigan. The Flow of
Fluidized Solids. Powder Technology,
6:343-351, 1972.
IV-1-9
-------
POROUS CERAMIC
DISTRIBUTOR
GAS/AIR
0 40 80 120 160 200 240 280 320
TIME, sec
Figure 1. Radiant cooled shallow fluidized bed. Figure 2. Start-up temperature/time
sequence.
OQ
40
FREQUENCY, Hz
Figure 3. Noise levels 2 ft away from radiant bed (including unsilenced fan noise).
IV-1-10
-------
1000
o
o
LU
a.
20
40
30
TIME, sec
Figure 4. Heating of 1/4-in.-diameter alloy steel bolt.
50
60
IV-1-11
-------
§,
UJ
o
o
•=t
cc
- PROPANE STOICHIOMETRIC
METHANE STOICHIOMETRIC
TOWN'S GAS ( )
STOICHIOMETRIC MIXTURE STRENGTH
50
45
40
1000
1100
1200
1300
1400
TEMPERATURE.'K
Figure 5. Percent radiation from shallow fluidized bed combustors.
IV-1-12
-------
24,000
20,000
16,000
12,000
800
900
TEMPERATURE,0 C
1,000
Figure 6. Input and radiant output from a 6-in.- diameter shallow bed combustor (based
on gross calorific value).
IV-1-13
-------
600
500
300
200
100
• ASTON DATA ON EXTENDED SURFACE/SHALLOW BEOS
1-irt. x 2-in. dia. TUBE DATA (DEEP BEDS)
PETRIEETAL
EXTENDED SURFACE
(DEEP BED)
550 1000
PARTICLE DIAMETER, fin
Figure 7. Comparison of shallow/deep bed heat transfer.
1500
IV-M4
-------
2. FLUIDIZED-BED COMBUSTION UTILITY POWER
PLANTS-EFFECT OF OPERATING AND DESIGN
PARAMETERS ON PERFORMANCE AND ECONOMICS
D. L. KEAIRNS, W. C. YANG, J. R. HAMM, AND D. H. ARCHER
Westinghouse Research Laboratories
ABSTRACT
Pressurized fluidized-bed boiler power plants have the potential to meet SO2, NO, and
participate emission standards at energy costs 10 percent below conventional plants with wet
scrubbing. This paper analyzes the sensitivity of the operating and design parameters selected for
the plant design on plant performance and economics. Results show that the plant costs and per-
formance are essentially invariant with projected changes in operating and design parameters—2.5
percent change in energy cost. The concept has the potential for achieving plant efficiencies of
~45 percent.
INTRODUCTION
A pressurized fluidized-bed boiler power
plant has been designed using state-of-the-art
power generation equipment.' -2 Performance,
costs, and pollution abatement were projected
for the system. The results show the concept
has the potential to meet SO2, NO, and par-
ticulate emission standards and may reduce
energy costs 10 percent below a conventional
plant with stack gas scrubbing.
Operating conditions and design param-
eters for the pressurized boiler were selected
based on an evaluation of available data,
power cycles, and alternative boiler concepts.
It is important to know how sensitive the oper-
ating and design parameters selected for the
base design are to the plant economics. An
understanding of the effect of changes in the
proposed design on plant cost will provide a
basis for evaluating current pressurized fluid-
bed combustion pilot plant data, planning
experimental programs, designing the devel-
opment plant, and understanding the
economic margin for solving technological
problems.
The sensitivity analysis evaluates the effect
of the following variables on plant design,
cost, and performance:
Fluidized bed boiler operating conditions
Bed temperature
Fluidizing velocity
Excess air
Pressure
Fluidized bed boiler design
Heat transfer surface—configuration,
heat transfer coefficient, and materials
Module capacity
Particulate carry-over from the boiler
Loading *
Size distribution
Power plant equipment operating
conditions
Gas turbine inlet temperature
Steam temperature and pressure
The evaluation is performed by considering
each variable separately; it is general in order
to permit the coupling of different effects to
assess alternative designs. It is performed to
indicate relative effects of variable changes in
plant costs relative to each other and the total
plant cost.
IV-2-1
-------
BASIS FOR SENSITIVITY ANALYSIS
The basis for the sensitivity analysis is the
boiler and plant design developed by
Westinghouse under contract to EPA. * '2 The
power plant cycle is shown schematically in
Figure 1. The plant subsystems included in
this sensitivity analysis are enclosed within the
broken lines. The pressurized boiler was
designed by Westinghouse and Foster
Wheeler and is shown schematically in Figure
2. The preliminary boiler design was for a
nominal 300-MW plant. The boiler design
consists of four modules; the modularized
design provides for a maximum of shop fabri-
cation and turndown requirements. Each
module includes four primary fluidized-bed
combustors, each containing a separate boiler
function—one bed for the pre-evaporator, two
beds for the superheater, and one bed for the
reheater. Evaporation takes place in the water
walls. All of the boiler heat transfer surface is
immersed in the beds, except for baffle tubes
above the bed to minimize particle carry-over.
Each module contains a separate fluidized bed
or carbon burn-up cell to complete the com-
bustion of carbon elutriated from the primary
beds. The philosophy used to design the boiler
was to maximize shop fabrication. Thus, the
300-MW plant utilizes boiler modules which
can be completely shop fabricated. From
roughly 300 to 600 MW, the boiler can be
partially shop fabricated—the pressure shell
being too large for rail transport. Larger
plants, utilizing the four-module concept,
would be field erected.
The operating conditions and design para-
meters for the boiler and the power cycle are
summarized in Table 1. The power plant per-
formance and economics were based on these
specifications. The cost breakdown for the
fluidized-bed steam generator is summarized
in Table 2. The fluidized-bed boiler design
was scaled to 600-MW capacity and the costs
estimated. These costs are summarized in
Figure 3. A breakdown of the power plant
equipment costs for a 635-MW plant is
presented in Table 3. Limestone or dolomite
regeneration is not included in this analysis.
IV-2-2
Thus the costs presented are for a once-
through system. The energy costs used for this
analysis are also presented in Table 3. The
costs of a conventional plant with wet scrub-
bing on the same basis are also indicated.
The following assumptions are made for
the sensitivity analysis:
1. The plant concept maintains the four
module with two modules per gas turbine
concept.
2. Coal feed rate is maintained constant for
each variable analysis. Thus the coal
feeding and handling system is assumed to
remain unchanged. This may not be
completely true if the bed area is changed
significantly and the number of feed points
increased or decreased. The cost of the coal
feed system is considered if the bed design
is altered.
3. Structural and erection costs are constant.
The structural steel and concrete costs for
the boiler plant equipment are~$2/kW.
The maximum change in these costs for the
cases considered is~$0.20/kW and will be
significantly less in general. The cost was
thus assumed constant for this analysis.
Erection cost changes are negligible.
4. Coal and stone feed size are assumed
constant—1/4 in. x 0. This parameter is
important when considering particle carry-
over, but insufficient information is
available to permit a quantitative analysis.
5. The ash and dust handling system cost is
assumed constant. This cost will be
affected by the particulate carry-over, but
it is considered a second order effect for the
cases considered.
6. Stack and foundation, instruments and
controls, and other costs are constant.
7. All variables not being evaluated are
assumed constant unless stated otherwise.
-------
Table 1. PRESSURIZED FLUIDIZED BED BOILER POWER PLANT
OPE RATING AND DESIGN CONDITIONS
Cycle
Steam system
Gas turbine expander
Pressure ratio
Inlet temperature
Air cooling
Coal feed rate
Number of boiler modules
Boiler modules/gas turbine
Fuel/air ratio
2400 psia, 1000°F superheat, 1000°F reheat
10:1
1600°F
5%
53,910 Ib/hr/module for nominal 300-MW plant
design
4
2
0.0919
Boiler design
Bed area
Heat transfer surface
Walls
Bed
Gas side heat transfer
coefficient
Tube materials
Bed depth (expanded)
Gas temperature drop from
primary beds to gas
turbine expander
35 ft2 (5x7 ft) -for ~80-M W module
2-in. OD tubes on 3-1/2 in. welded wall spacing
1-1/2-in. OD tubes in pre-evaporator and super-
heater; 2-in. OD in reheater (details in text)
50Btu/hr-ft2-°F
SA-210-A1 —pre-evaporator
SA-213-T2—lower superheater
SA-213-T22—water walls; upper superheater
(lower loops); reheater
SA-213-TP304H —upper superheater (upper
loops)
11to14ft
150°F
Boiler operating conditions
Bed temperature (100% load) 1750°F
Fluidizing velocity
Excess air
Particle carry-over
carbon from primary beds
6 to 9 ft/sec
17.5%
~7 gr/scf
6% of carbon feed
Auxiliaries
Coal feed system
Primary particulate
removal
Secondary particulate
removal
Stack gas coolers
Petrocarb feed system
4 size 355 VM 8/0/150 Duclone per module—
nominal 300-MW design
2 model 18000 Type S collectors per module-
nominal 300-MW design (quoted by Aerodyne Dev.
Corp.)
Conventional heat exchanger design.
IV-2-3
-------
Table 2. COST OF A 318-MW PRESSURIZED
FLUID-BED BOILER
Pressure parts3
Shell
Subcontracted and contracted
equipment
Drafting
Home office
Sub-total
Erection
TOTAL
$1,777,000
935,000
435,000
185,000
685,000
$4,017,000
500,000
$4,517,000
a Pressure parts include tubing cost, headers,
downcomers, risers, tube bending, tube welding,
and water wall fabrication.
b Field erected ($3,856,000 shop assembled).
BOILER PLANT EQUIPMENT
Operating Conditions
Bed Temperature
The full load design temperature is 1750°F.
Lowering the design bed temperature
increases the total amount of heat transferred
in the bed and thus increases the steam
turbine power generation. At the same time,
lowering bed temperature decreases the gas
turbine inlet temperature—assuming no
burning above the bed—and thus decreases
the total gas turbine power generation. The
decrease in gas turbine power is larger than
the increase in steam power, resulting in an
overall decrease in plant power (Figure 4).
Lowering the bed temperature increases the
total amount of heat transferred in the bed
and thus requires more heat transfer surface.
Assuming the cross sectional area of the fluid
bed (5 x 7 ft for a 300-MW nominal plant size)
and tube size/tube pitch are constant, the
expanded bed depth for each functional bed
increases with decrease in bed temperature as
Table 3. PRESSURIZED FLUIDIZED-BED
BOILER POWER PLANT COSTS
Equipment Costs
Component
Cost,$/kW
Boiler plant equipment
Boiler
Particulate removal
Piping/ducts
Stack and foundation
Coal handling and feeding equipment
Ash and dust handling system
Instruments and controls
Miscellaneous equipment
Steam turbine—generator equipment
Gas turbine—generator equipment
Other: land, structures, electric plant
equipment, miscellaneous plant
equipment, undistributed costs
Subtotal
14.49
12.76
4.43
0.47
14.94
1.55
3.10
0.94
44.14
14.80
70.38
182.00
Total capital cost (inc. escalation, IDC etc.) 265.00
($340/kW for conventional plant with wet scrub-
bing on same basis)
Energy costs
Fixed charges
Fuel
Dolomite
Operating and maintenance
mills/kWhr
6.44
4.04
0.52
0.71
11.71
(13.45 for conventional plant)
shown in Figure 5. The bed depth of the two
superheater beds is assumed to be the same
for convenience. This will not affect the total
heat transfer surface requirement and the
resultant module height shown in Figure 6.
The bed depth and module height can be
reduced by enlarging the bed area and module
diameter. However, since the module diameter
of 12 feet is considered to be the largest ship-
pable railroad size, increase in module
diameter to accommodate additional heat
transfer surface may not be economic for a
300-MW plant.
IV-2-4
-------
The effect of changing design bed tempera-
ture on the steam generator cost is calculated
based on the boiler cost estimation shown in
Table 2 and on the assumptions that the
module diameter is constant at 12 feet and
that the total number of modules is four based
on turndown consideration. The cost (not
including erection) of the 4-module steam
generator as a function of bed temperature
with constant module diameter (12ft) is shown
as curves la and Ib in Figure 7. Curve la
assumes the maximum allowable bed depth to
be 20 feet. That means any bed with expanded
bed depth larger than 20 feet will have to be
split into two beds with their separate air
plenums and freeboards. Curve Ib assumes
that there is no restriction on maximum bed
depth. The choice of 20 feet as the maximum
allowable bed depth is arbitrary, just to show
the importance of this variable on the cost of a
steam generator. It is doubtful that the bed
depth of each fluid bed can be unrestricted
without creating undesirable bubble
formation and slugging, poor bedtube heat
transfer coefficient, and temperature
gradients in the bed at some bed depth. The
maximum allowable bed depth at specific
operating conditions will have to be experi-
mentally determined in a large unit. Without
the required experimental evidence, the steam
generator cost (not including erection) is
plotted against the maximum allowable bed
depth in Figure 8. The bed temperatures were
calculated by assuming the gas turbine tem-
peratures of 1600, 1500,1400, and 1300°F and
by assuming a linear temperature loss between
the boiler and the gas turbine inlet. The cost of
the steam generator designed for 1636°F
increases ~ 20 percent ( ~ $2.8/kW) over that
designed at 1750°F if the maximum allowable
bed depth is 15 feet. The cost increase is
primarily due to the splitting of beds with bed
depth higher than 15 feet. Bed splitting can be
avoided by either decreasing boiler tube
diameter and spacing in the bed or increasing
the module diameter. Decreasing tube
diameter and spacing will change the bed-tube
heat transfer coefficient, tube bending and
fabrication, and tube wall thickness.
Increasing the module diameter will not only
change the cost of the pressure shell but also
affect construction—complete shop-
assemblage versus partial field erection. All
these factors have to be taken into account in
designing an optimal boiler. These factors are
discussed in separate sections.
In the present cost estimation, the possi-
bility of using thinner wall tubes for the
designs at lower bed temperatures was also
taken into account by calculating the
minimum tube wall thickness requirement. No
allowance for corrosion is provided.
Change in the operating bed temperature
will also change the gas temperature to both
the primary and secondary cyclones and thus
change the actual volumetric gas flow rate.
This will change gas inlet velocity to the
cyclones which, in turn, affects cyclone
collection efficiency. This effect was estimated
to be small compared to the effect of the
change in pressure drop across the bed due to
a change in the design bed temperature.
Decreasing the design bed temperature
increases the heat transfer surface require-
ment in the bed, which requires an increase in
pressure drop due to an increase in bed depth
if bed area and boiler tube configuration are
constant. The decrease in net plant power
output as a function of the design bed
temperatures is presented in curve 1, Figure 9.
The effect is small—a decrease of only ~ 0.3
percent if the design bed temperature is
reduced to 1407°F.
Bed temperature is one of the primary
variables used for load turndown in the
present design. A 4:1 turndown can be met if
the design bed temperature is higher than
1600°F. The primary limitation on the
operating bed temperature is the sulfur
removal efficiency of the sorbents in the bed.
At bed temperatures higher than 1750°F or
lower than 1350°F, the sulfur removal
efficiency in the bed is too low. Thus it is
concluded from the above bed temperature
analysis that the design bed temperature
should be the highest temperature required
IV-2-5
-------
for desirable degree of load turndown and
sulfur removal in the bed.
Fluidizing Velocity
At constant fuel feed rate and excess air,
increasing the fluidizing velocity requires a
decrease in the bed area and in the module
diameter (Figures 10 and 11). For a constant
overall heat transfer coefficient and a specific
design bed temperature the total heat transfer
surface in the bed is constant; a decrease in
the bed area will require an increase in the bed
depth at constant tube size and tube spacing
and thus an increase in the module height. An
economic design will depend on the balance of
thse factors.
The bed area and bed depth requirements
with respect to change in fluidizing velocity at
different design bed temperatures are calcu-
lated. The corresponding module height and
module diameter are presented in Figure 11.
The cost of the pressure shell at different
inside diameters is estimated based on the
data from Foster Wheeler Corporation' and
on an independent estimation by Westing-
house (Figure 12). The discontinuity at a
module inside diameter of 12 feet is due to the
cost difference between the shop assembled
and the field erected shell.
One additional cost which has to be taken
into consideration is the fabrication cost. In
addition to the shell cost, the change of fabri-
cation cost of water walls and tube bending
cost are not to be ignored. Taking into
consideration the factors involved, the steam
generator cost is plotted against the superficial
fluidizing velocity in the pre-evaporator ash
shown in Fugure 13 for 318-MW and 635-MW
plant. The superficial fluidizing velocity in
the pre-evaporator is used here since it is the
largest velocity in all beds inside a single
module. The superficial fluidizing velocity in
the superheaters and reheater can be calcu-
lated accordingly, the results show that
increasing the fluidizing velocity tends to
increase rather than decrease the total steam
generator cost at a 318-MW plant size.
Decreasing the fluidizing velocity in the pre-
evaporator below ~ 8 ft/sec requires a shift
from the shop-assemblage to the field erection
and escalates suddenly the steam generator
cost. A minimum cost does exist for a 635-
MW plant. Figure 13 is for unrestricted max-
imum allowable bed depth. If the maximum
allowable bed depth is limited to say 10 or 20
feet, the disadvantage of increasing the fluid-
izing velocity would be even larger at 318-MW
size. At 635-MW size, the minimum would
shift to lower velocity.
The important thing here is to understand
why increasing the fluidizing velocity increases
the steam generator cost at 318-MW size and
produces a minimum at 635-MW. To better
illustrate the point, cost reduction due to
decrease in module diameter and cost
escalation due to increase in module height for
a four-module design are shown in Figure 14
for the design bed temperature at 1750°F.
Increasing the fluidizing velocity escalates the
steam generator cost almost linearly from the
basic design point due to increase in module
height (curve 2). At the same time, the cost
decreases due. to decrease in module diameter;
however, the decrease is much more gradual
and levels off at higher fluidizing velocity
(curve 1). This is because the bed area alone
occupies less than 40 percent of the total cross-
sectional area of a pressurized module. The
remaining area is required for piping and
headers; this space is relatively unchanged at a
specific plant size even though the bed area is
reduced to increase the fluidizing velocity. At
635-MW size, however, the cost reduction due
to decrease in module diameter is larger
during initial deviation from the basic design
point and thus creates a minimum (Figure 14).
Another approach for analyzing the effect
of fluidizing velocity would be to change the
number of modules as well as the module
diameter. It is preferred to have a 5-module
design based 0:1 turndown consideration;
however, if a 3-module design shows a sub-
stantial saving with negligible effect on turn-
down capability, it would be a better choice.
IV-2-6
-------
Change in fluid izing velocity will change the
total bed area required for each functional
bed, but the total bed volume for each func-
tional bed will remain constant once the tube
size and spacing are fixed. Thus a design with
smaller number of modules will require a
larger module diameter at the same design
fluidizing velocity. An estimation can usually
be performed to evaluate the relative economy
between these two designs. For example,
consider a 4-module and a 3-module design at
the same design fluidizing velocity and with
the same bed depths. Since the bed volume of
each functional bed is constant at fixed tube
size and spacing for both cases, we have
D
3_
4
(1)
if bed height is assumed constant for both
cases and the bed area occupies a fixed
percentage of the total cross-sectional area of
a module. 04 and DS are the respective
module diameters for the 4-module and 3-
module designs. Since the shell cost is
dependent on the vessel diameter (Figure 12),
the relative advantage of these two designs will
depend on the plant size in question. For
example, for D4=12.5 feet, D3 can be calcu-
lated from equation (1) to be 14.4 feet. The
shell cost can be found from Figure 12 to be
$0.94 x 106 for the 4-module design and $0.79
x 106 for the 3-module design. If fabrication
cost of the module internals is similar in both
cases, the 3-module design will have a slight
economic advantage. However, this advantage
becomes progressively smaller because of
rapidly increasing shell cost at large module
diameter and rapidly decreasing shop-fabrica-
table portion in the design. It is estimated that
the largest module diameter which is still
economic for the 3-module design is ~ 17 feet.
This conclusion is based on the assumption
that boiler turndown is not a problem. If the
module diameter is in the shop-fabricatable
and railroad -transportable range, i.e., < 12
feet, designing for the maximum shippable
module will have definite advantages provided
that turndown is not a problem.
It is concluded from this analysis that for a
plant size around 300 MW, the module
diameter should be the largest within shipping
limitations (12 feet for railroad transporta-
tion); at a 600-MW plant size, an optimum
fluidizing velocity exists, and it should be
found for each capacity. However, the cost
deviation from that of the optimum design is
less than Sl.OO/kW (Figure 13). Of course,
decreasing the bed area may reduce the
number of feed points but the saving is only ~
$0.10/kW. Although the effect of the
fluidizing velocity on the steam generator cost
is primarily based on the basic design
conditions, i.e., bed-tube heat transfer
coefficient = 50 Btu/ft2-hr-°F, the trends
would be the same for bed-tube heat transfer
coefficient at 35 or 75 Btu/ft2-hr-°F. Change
in tube size and spacing may change the slope
of the curves or alter the minimum in Figure
13; the conclusions will remain the same.
The effect of changing fluidizing velocity
on the bed-tube heat transfer coefficient,
combustion efficiency, and total particulate
carry-over is not taken into account in this
analysis due to lack in accurate quantitative
data. The effect of dust loading and particle
size distribution on the cost of the particulate
removal system is evaluated separately.
Excess Air
Change in design excess air will affect the
cycle efficiency and the cost of the boiler
module, the steam and gas turbine equipment,
and the particulate removal system. In order
to quantify the effect of excess air on total
boiler cost, the air/fuel ratio is allowed to vary
with the total fuel input kept constant. To
simplify the analysis, other parameters—bed
temperature, tube size and tube spacing, bed-
tube heat transfer coefficient, fluidizing
velocity, and number of boiler modules—are
held constant at the basic design values.
Thus, when excess air is increased beyond
the design value, bed area has to be increased
if the fluidizing velocity is kept constant.
When the bed area is increased, the module
diameter has to be increased as well; however,
IV-2-7
-------
the module height is decreased due to a
decrease in bed depth. Increasing air input
into the bed will also increase the amount of
heat carried out from the bed by air and
reduce the total heat transferred in the bed. At
constant bed-tube heat transfer coefficient,
the total heat transfer surface requirement is
reduced. At constant tube size and tube
spacing, the bed depth is also reduced as well
as the total module height. At 100 percent
excess air, a reduction of > 40 percent in heat
transfer surface and a reduction of >30
percent in module height is possible. The bed
depths for different functional beds are
reduced to ~4 feet which decreases pressure
drop through the beds and increases cycle
efficiency. The module diameter, in turn,
increases from the original 12 feet inside
diameter to more than 16 feet for an 80-MW
module. Transferring all these changes into
economics, an increase in excess air can
reduce the boiler cost up to ~$0.60/kW as
shown in Figure 15. Cost reduction due to heat
transfer surface increases continuously with
respect to excess air because of the decrease in
the total amount of heat transferred in the
bed. Cost reduction due to pressure shell first
increases because of reduction in module
height and then decreases because of increase
in module diameter.
Increasing excess air will increase the over-
all plant efficiency as shown in Table 4. Larger
gas turbines or additional gas turbines are
needed to handle the increased mass flow of
gas. If additional units are used, the increase
in gas turbine equipment cost is shown as a
cost adder in Figure 16. This does not account
for cost reductions which can be realized by
going to larger turbine capacities. This cost
increase in gas turbine equipment is partially
offset by a decrease in steam turbine equip-
ment cost also shown in Figure 16. The major
equipment items taken into consideration in
this analysis include gas turbines with external
manifolds, steam turbine system, circulating
water and condensing systems, feedwater
Table 4. EFFICIENCY CALCULATIONS AT VARIABLE AIR FLOW RATES
Excess
air,
%
17.5
50.0
90.0
Fuel/air
ratio
0.0861
0.0674
0.0532
Plant
size,
MW
644.1
648.0
649.7
Heat
rate,
Btu/kWhr
9026
8972
8948
Gas
turbine,
MW
127.9
156.9
193.5
Steam,
MW
532.2
506.3
470.4
Gas
turbine/
steam
ratio
0.240
0.309
0.411
No. of
gas
turbine
2.000
2.450
3.025
IV-2-8
-------
system including station piping, and stack gas
coolers.
Instead of keeping the fluidizing velocity
constant, the bed area and module diameter
can be kept constant and allow the fluidizing
velocity to increase with excess air. In this
case, the cost reduction in heat transfer
surface will be similar, but the cost reduction
in the shell will continuously increase with
increasing excess air and not g6 through a
maximum (Figure 15). The cost reduction in
heat transfer surface and pressure shell at 100
percent excess air in this case (with 10 to 15
ft/sec fluidizing velocity) is estimated to be
~$1.00/kW. However, increasing the fluid-
izing velocity to larger than 15 ft/sec may be
impractical in this design approach.
Increasing the excess air will decrease the
total heat transfer surface required in the fluid
bed until no boiler tube surface will be
required at an excess air of approximately 300
percent. In this case the power system would
become a combined cycle plant with the gas to
the turbine expanders supplied from a coal-
tired, adiabatic combustor. The heat recovery
boiler would probably be unfired. This system
concept has several significant differences
from a pressurized fluidized-bed boiler power
plant concept: for example, the boiler
becomes an adiabatic combustor, particulate
removal equipment costs increase significantly
due to the increased gas flow, gas piping costs
increase and the gas turbine power contribu-
tion increases from ~ 20 percent up to ~ 70
percent. An economic analysis of this system
has not been made as part of this evaluation.
The heat rate for the adiabatic combustor
plant is projected to be 100 to 500 Btu/kWhr
(depending on the gas turbine inlet tempera-
ture) greater than for the pressurized boiler
plant. Further evaluation of this high excess
air case is required to perform a comprehen-
sive assessment.
Increasing the excess air will provide more
flexibility in turndown. At 100 percent excess
air, an additional ^ 10 percent load reduction
is possible as compared to operation at 10
percent excess air. This means a boiler
designed at 1600°F and 100 percent excess air
will be able to meet a 4:1 turndown require-
ment. An adiabatic combustor system should
extend the turndown capabilities.
More discussion on excess air and gas flow
rate appears in the section on particulate
removal.
Operating Pressure
The full load design pressure is 10 atm.
When the design pressure level is changed and
the other operating parameters remain
constant, the gas density will change in
proportion to the pressure, and the volumetric
flow will vary accordingly. Therefore, the bed
cross-sectional area will have to be changed to
maintain constant fluidizing velocity and the
bed depth changed to maintain constant bed
volume. The heat transfer coefficient may
change because of changes in the quality of
fluidization.
The changes in volumetric flow and in gas
density will affect the design of the particulate
removal equipment. Gas turbine cycle
efficiency is also dependent on the operating
pressure. However, analysis of the high
pressure fluidized-bed coiler system indicates
a reverse direction in pressure level effect.
Auxiliary equipment such as the coal and dol-
omite feeding systems are pressure dependent
as well. The relative importance of these fac-
tors with respect to operating pressure is ana-
lyzed in the following paragraphs.
First, consider the boiler module alone. At
constant fuel feed rate and excess air,
increasing the operating pressure will decrease
the gas volumetric flow rate. There are two
approaches in designing the boiler modules:
(a) keep the bed area constant and let the
fluidizing velocity change with the operating
pressure, and (b) keep the fluidizing velocity
constant and change the bed area according to
the operating pressure. The incremental
module cost for the constant bed area case is
IV-2-9
-------
the cost of reinforcing the pressure shells.
Since for a constant overall heat transfer
coefficient and a specific design bed tempera-
ture, the total heat transfer surface in the bed
is constant. This amounts to $0.20/kW and
$0.30/kW for operating pressures of 15 and 20
atm, respectively, at 300-MW nominal plant
size. At 600-MW plant size, the respective cost
increments are $0.10/kW and $0.20/kW. The
credit of decreasing particulate carry-over by
operating pressure will require a decrease in
bed area. Since the total heat transfer surface
is constant, a decrease in bed area will require
an increase in bed depth at constant tube size
and tube spacing, and thus an increase"nr the
module height. In this case, the effect of
operating pressure on boiler design and boiler
cost while keeping the fluidizing velocity
constant is the same as the effect of changing
fluidizing velocity with the operating pressure
constant. Thus the Figures 10, 13, and 14 can
also be applied for this approach if the
coordinate for the fluidizing velocity is
changed to (fluidizing velocity at basic design)
x (new operating pressure/basic design
pressure). For a design bed temperature of
1750°F, the cost increments are $0.80/kW
and $1.90/kW for operating pressures of 15
and 20 atm, respectively, at 300-MW plant
size. At 600-MW plant size, the cost incre-
vents are $0.20/kW and $1.20/kW respec-
tively.
Comparing these two design approaches,
the constant area case is the less costly one.
Moreover, if basic design bed area and
fluidizing velocity are maintained, increasing
operating pressure will mean a higher capacity
shop-fabricatable module (i.e., module
diameter <12 ft). At 15 atm operating
pressure, a module of ^120-MW capacity
can be shop-fabricated; at 20 atm pressure,
the maximum shop-fabricable module is
~ 160 MW. However, the module height will
be considerably increased because of increase
in total heat transfer surface required. The
total increase in module height will depend on
the heat transfer surface arrangement in the
bed.
Increasing the operating pressure reduces
the size of the particulate removal equipment
because of the decrease in volumetric flow
rate. It reduces the particulate removal
efficiency as well because of changes in gas
density and viscosity. Increase in operating
pressure will also require reinforcement of the
containment vessel and the piping and
ducting. The savings in the particulate
removal equipment by operating at 15 and 20
atm are estimated to be $4.0/kW and
$6.0/kW, respectively, for a 300-MW plant
size. The major saving comes from the
secondary collectors where maximum single
unit capacity is assumed restricted to 30,000
acfm. An increase in operating pressure will
result in fewer units. Cost reduction by
operating at higher pressure .will be even
greater for larger plant sizes or at higher
design excess air.
Cycle optimization calculation was
performed to evaluate the effect of operating
pressure. The parameters studied are: inter-
cooled and non-intercooled compression; gas
turbine compressor pressure ratios from 10 to
30; and cycle gas side pressure drops of 3 to 8
percent. The results are summarized in Figure
17 which gives plant heat rate for the inter-
cooled and non-intercooled cases. For the non-
intercooled case, the best efficiency is obtained
at a pressure ratio of 10; for the intercooled
case, the optimum pressure ratio is 15 but with
a higher heat rate and a more complex gas
turbine. Thus an increase in operating
pressure higher than 10 atm decreases the
overall plant efficiency. This decrease is small
however: ~ 0.2 percent at 15 atm.
Weighing the above discussions, the only
distinct advantage for operating at pressures
higher than 10 atm is the capability of shop-
fabricating a large capacity plant, especially at
higher design excess air.
IV-2-10
-------
Boiler Design
Heat Transfer Surface
Configuration
In the 300-MW design, the heat transfer
surface is provided by serpentine tubes having
the horizontal sections spaced as shown in
Figure 18. Tubes of 2-in. OD are used at
waterwalls and are space 3-1/2 inches apart.
Tubes for pre-evaporator and superheaters are
1-1/2-in. OD and tubes for the reheater are 2-
in. OD. The tubes can usually be arranged in
staggered or rotated diamond arrays, or they
can be arranged in a square or rectangular
pitch (Figure 19).
The effect of tube pitch/diameter ratio on
the cost of the steam generator (not including
erection) for constant 12-ft module diameter
was evaluated for three different tube sizes
and tube spacings with respect to change in
the design bed temperatures. The results are
plotted in Figure 7. Curve 2 represents the
.estimated cost for a staggered arrangement of
1-in. OD tubes, where H = 4 inches and V = 2
inches (see Figure 19 for definition) in all beds.
Curves 3a and 3b are for staggered arrange-
ment of 2-in. OD tubes where H = 8 inches
and V = 4 inches in all beds. Some interesting
trends are present when these results are com-
pared to the steam generator cost for the basic
design. Ignore curves Ib and 3b for the time
being because the assumption of unrestricted
maximum allowable bed^ depth is not con-
sidered reasonable. Then:
1. Decreasing tube size and tube spacing
increases the steam generator cost at
design bed temperatures above ~1520°F
(curve 2). Further decrease in bed
temperature necessitates splitting the
reheater bed in the basic desig^ into two
beds and substantially increasing the
steam generator cost of the basic design.
2. Increasing tube size and tube spacing also
increases the steam generator cost (curve
3a).
The reasons for these results are as follows:
1. When tube size and tube spacing are
decreased, more heat transfer surface can
be immersed in a unit bed volume which
results in lower bed height and module
height; thinner wall tubes can be used
which results in lower tubing cost. These
are positive advantages.
2. However, smaller tube size and tube
spacing increase the amount of tube
bending and tube welding required. This
is because more tubes of smaller diameter
are required to carry the same
water/steam load at a constant flow rate in
the tube. Fabrication cost as a function of
tube wall thickness is not taken into
account because of not enough informa-
tion available. Pumping costs, a part of
total operating cost, are not included here.
The balance of these two factors determines
the total steam generator cost.
Since the tubing cost constitutes only about
20 percent of the cost of the pressure parts, the
increase in fabrication cost is a more
important cost. This can best be illustrated in
Figure 20 where the component costs are
plotted against the available heat transfer
surface per unit volume which relates to the
tube pitch/diameter ratio. Shell cost increases
almost linearly with decreasing heat transfer
surface per unit bed volume. The cost of
tubing, headers, downcomers, and risers is
almost constant at heat transfer surface larger
than 6 ft2/ft3 bed volume; it increases rapidly
at heat transfer surface lower than ~ 6 ftVft3
bed volume, which corresponds to the use of a
tube diameter of 1-1/2-in. OD or larger.
When larger tubes are used, the minimum
wall thickness increases rapidly and so does
the tubing cost which contributes most of the
cost escalation at lower heat transfer surface
IV-2-11
-------
per unit bed volume. However, the cost of tube
bending, tube welding, and water walls fabri-
cation increases steadily with increase in heat
transfer surface per unit bed volume.
However, the cost of tube bending, tube
welding, and water walls fabrication increases
steadily with increase in heat transfer surface
per unit bed volume. The balance of all these
factors creates a minimum in total steam
generator cost at about 6.5 ft2 heat transfer
surface per ft3 of bed volume. This can be
achieved by arranging 1-in. OD tubes where
H = 4 inches and V = 3 inches. Fortunately,
the tube size and tube spacing used in the
basic design is very close to this actual
minimum. Clearly, there are different
minimums at different design bed
temperatures. An optimum design for a
specific operating condition requires a
separate evaluation. However, this optimum
design point is not as critical as may be
generally conceived. For example, at a design
bed temperature of 1750°F (Figure 20) the.
difference in the steam generator cost between
the optimum design and other designs is
within $1.00/kW for available heat transfer
surface of 3 to 11 ftVft3 bed volume, which
covers the three tube sizes and tube spacings
in our current evaluation.
The maximum allowable bed depth is a
more important variable. The steam generator
costs are also plotted against the maximum
allowable bed depth at constant design bed
temperature and constant bed area in Figures
21 and 22. Although the steam generator of 1-
in. OD tubes (where H = 4 inches and V = 2
inches) is not as economical compared to the
basic design, it becomes progressively more
attractive at a lower maximum allowable bed
depth. At maximum bed depths lower than 14
feet (Figure 21), a steam generator using 1-in.
OD tubes is actually cheaper than the basic
design by up to
-------
Heat Transfer Coefficient
In the basic design, the overall heat
transfer coefficients assumed are 47 Btu/ft2-
°F-hr for the pre-evaporator, 45 Btu/ft2-°F-hr
for the superheater, and 43 Btu/ft2-°F-hr for
the reheater. The bed-tube heat transfer
coefficient is assumed to be 50 Btu/ft2-°F-hr
for all beds. When the bed-tube heat transfer
coefficient is changed, the overall heat transfer
coefficient will be changed as well. This leads
to a change in the total heat transfer surface
requirement and the bed depth.
Change in the heat transfer coefficient will
also change the tube metal temperature. A
change of tube material may be necessary for
some cases. The design metal temperature is
assumed to be the maximum outside tube wall
temperature based on the minimum
permissable wall thickness.
Taking into consideration the aforemen-
tioned factors, the total steam generator costs
at bed-tube heat transfer coefficients of 35 and
75 Btu/ft2-hr-°F were projected for different
tube pitch/diameter ratios at different design
bed temperatures. At a low bed-tube heat
transfer coefficient (35 Btu/ft2-hr-°F) where a
large amount of in-bed heat transfer surface is
required, a boiler with smaller tube size and
tube spacing is much more economical than
the one with larger tube size and tube spacing.
Saving up to 20 percent of the total steam
generator cost is feasible if bed area is
constant and maximum allowable bed depth is
20 feet. If the maximum allowable bed depth
is less than 20 feet, the saving will be even
larger. At a bed-tube heat transfer coefficient
of 75 Btu/ft2-hr-°F where the heat transfer
surface requirement is substantially reduced,
smaller boiler tubes and spacings do not have
a clear advantage. If the bed-tube heat
transfer coefficient is 35 Btu/ft2-hr-°F rather
than 50 Btu/ft2-hr- °F as assumed in the basic
design, the steam generator cost will increase
by $2.7/kW at 1750°F design bed temperature
and by $5.4/kW at 1600°F. If 1-in. OD tubes
are used where H = 4 inches and V = 2
inches, the cost escalation would be $2.0/kW
at 1750°F and $3.4/kW at 1600°F. If the heat
transfer coefficient is increased to 75 Btu/ft2-
hr-°F, the reduction in steam generator cost
from that of the basic design is only marginal,
~ $1.0/kW at 1750°F.
Figures 25 and 26 show the effect of the
bed-tube heat transfer coefficient on the steam
generator cost at constant bed temperature.
An increase of the bed-tube heat transfer co-
efficient from 50 to 75 Btu/ft2-hr-°F (a 50
percent increase) decreases the steam
generator cost by ~ 10 percent ( ~ $1.40/kW).
A decrease of bed-tube heat transfer co-
efficient from 50 to 35 Btu/ft2-hr-°F (a 30
percent decrease) increases the cost by ~20
percent (~ $2.80/kW) (Figure 25). The'curves
start to level off at higher bed-tube heat
transfer coefficients. Thus, further increase in
bed-tube heat transfer coefficient larger than
about 75 Btu/ft2-hr-°F does not affect the cost
substantially. However, further decrease in
bed-tube heat transfer coefficient lower than
50 Btu/ft2-hr-°F increases the steam
generator cost rapidly, especially for large
tube sizes and tube spacings and at lower
design bed temperatures (Figures 25 and 26).
A 40 percent increase in cost occurs when the
bed-tube heat transfer coefficient decreases
from 50 to 35 Btu/ft2-hr-°F at design bed
temperatures of 1636°F for 2-in. OD tubes
where H = 8 inches and V = 4 inches (curve 3,
Figure 26). It is recommended to design the
steam generator at a bed-tube heat transfer
coefficient about 75 Btu/ft2-hr-°F if it is at all
possible, and to avoid designing the steam
generator at a bed-tube heat transfer co-
efficient lower than 50 Btu/ft2-hr-°F,
especially if large tubes and spacings are used
and if lower bed temperatures are employed.
To complete the evaluation, the cost
information was also prepared for other cases
at different design and operating variables to
show the interacting effect of tube
pitch/diameter ratio, bed-tube heat transfer
coefficient, maximum allowable bed depth,
and the design bed temperature.
IV-2-13
-------
Again, the maximum allowable bed depth
turns out to be the limitation of the steam
generator design and cost, especially at low
bed-tube heat transfer coefficient where large
amount of heat transfer surface is required in
the bed. In this case, a smaller tube size and
tube spacing and a higher design bed temper-
ature are preferred. At a maximum allowable
bed depth of 10 feet, the boiler designed at 35
Btu/ft2-hr-°F costs $3.5/kW more than that
designed at 50 Btu/ft2-hr-°F and S8.6/kW
more than that designed at 75 Btu/ft2-hr-°F at
1750°F design bed temperature. At design bed
temperature of 1407°F, the figures are
$6.7/kW and $12.5/kW, respectively. With 1-
in. OD where H = 4 inches and V = 2 inches,
the figures are $3.3/kW and $4.4/kW at
1750°F; and $2.8/kW and $8.1/kW at 1407°F,
respectively.
Cost savings become smaller when the bed-
tube heat transfer coefficient is further
increased over 75 Btu/ft2-hr-°F. If the bed-
tube heat transfer coefficient is decreased to
lower than 50 Btu/ft2-hr-°F, the steam
generator cost increases rapidly. Thus it is
recommended that the bed-tube heat transfer
coefficient be kept at higher than 50 Btu/ft2-
hr-°F and preferable around 75 Btu/ft2-hr-°F
at the current design conditions.
Tube Materials
The tube materials used in the basic design
are conventional boiler tube material with SA-
210-Al for tubes in pre-evaporator; SA-213-
T2 for tubes in lower superheater; SA-213-T22
for tubes in water walls, upper superheater
(lower loops), and reheater; and SA-213-
TP304H for tubes in upper superheater (upper
loops). Changes in design bed temperature,
bed-tube heat transfer coefficient or steam
temperature may require higher grade tube
materials; however, these changes do not sub-
stantially affect the steam generator cost
because the tubing cost alone constitiutes only
~ 10 percent of the total steam generator cost
( ~ $1.40/kW). Higher fabrication cost for
higher alloy material may increase this cost
slightly. Nevertheless, the total boiler cost is
not expected to increase significantly due to
change of tube material unless the operating
bed temperature and heat transfer coefficient
are drastically changed.
Module Capacity
Boiler modules can be shop fabricated,
partially shop fabricated or field erected
depending on the size. Modules up to 12-ft
diameter can be shop fabricated. Modules up
to 17-ft diameter can be partially shop
fabricated—the boiler internals being shop
fabricated, the pressure shell being field
erected. The plant concept for a given capacity
can be either multiples of shop fabricated
modules, partially shop fabricated modules, or
field erected modules.
The boiler plant equipment cost is different
for each case. The steam generator cost will
depend on operating conditions and design
variables such as design bed temperature,
tube size and tube spacing, maximum
allowable bed depth, etc. Auxiliary equipment
will also be affected: coal feeding, limestone or
dolomite feed and withdrawal, particulate
removal, steam piping, boiler feed water
system, etc.
The evaluation of these approaches is
based on Figure 3, which presents the cost
variation of the pressure parts, . shell,
subcontrated and contracted equipment,
drafting and home office, and erection with
respect to the plant size. Figure 27 presents
the resulting costs (including erection) for
design bed temperature at 1750°F; Figure 28
presents those for design bed temperature at
1636°F at a maximum allowable bed depth of
20 feet. The results show that at a plant size
larger than ~ 340 MW, partially shop
fabricated 4-module plants with maximum
shop fabrication of the pressure parts are
more economic than collective multiples of
largest shop-fabricated modules of the same
plant capacity. Changing the tube size and
tube spacing does not affect this conclusion
(Figures 27 and 28). Change in heat transfer
coefficient should not produce a different
conclusion because the cost escalation' due to
IV-2-14
-------
addition of a single module is more expansive
than simple enlargement of the existing
modules. However, this is no longer true when
the shell size is larger than ~ 17 feet, because
the degree of shop fabrication of the pressure
parts decreases and the steam generator cost
again increases at a much higher rate.
When the maximum allowable bed depth is
decreased from 20 feet, the cost saving of the
4-module partially shop-fabricated plant is
expected to increase. The splitting of the beds
in the shop fabricated module means an
increase in module height; however, in the
partially shop-fabricated module the splitting
of beds can be avoided by simply enlarging the
module diameter. From the discussion on
Figure 14, the latter means a more economic
alternative except when the module diameter
is increased beyond ~ 17 feet.
The optimum design variables (fluidizing
velocity, excess air, and pressure) in relation to
module capacity are discussed in their
respective sections.
Participate Removal System Economics
The primary variables taken into consider-
ation in analyzing particulate removal system
economics are dust loading, particle size
distribution entering and leaving the system,
and gas flow rate. The effects of boiler
operating variables — bed temperature, free-
board height, and superficial velocity — and
design variables — tube pitch/diameter ratio
and bed-tube heat transfer coefficient — were
also evaluated.
Range of Dust Loadings and Particle Size Dis-
tributions Considered
The cases evaluated are outlined in Table
5.
Gas Turbine Specification
A review of operating experience and
assessment of erosion in gas turbines was
prepared by Westinghouse under contract to
the Office of Air Programs.1 Specifications for
the fluidized-bed combustion system based on
that study are:
•Dust loading less than 0.15 gr/scf.
•Concentration of particles greater than 2
less than 0.01 gr/scf.
These design requirements will be updated
as additional laboratory test data and opera-
, ting experience become available.
TableB. CASES EVALUATED FOR DETERMINING EFFECT
OF DUST LOADING AND PARTICLE SIZE DISTRIBUTION
Cases
evaluated
Group
Casel
Case 2
CaseS
Case 4
CaseS
Group 2
Casel
Case 2
CaseS
Dust loading
leaving FBC
Basic design
(6.7 gr/scf)
Double design value
Triple design value
Triple design value
Triple design value
Basic design
(6.7 gr/scf)
Double design value
Triple design value
Particle size
distribution
(Refer to Figure 29)
Curve 1 for particles
elutriated from FBC
Curve 2 for particles
elutriated fromCBC
Curve 2 for particles
elutriated from FBC
Curve 3 for particles
elutriated from CBC
Cyclone system
design
Figure 30
Figure 30
Figure 30
Figure 31
Figure 32
Figure 30
Figure 30
Figure 30
IV-2-15
-------
Effect of Dust Loading and Particle Size Dis-
tribution on Particulate Removal Equipment
Cost
The effect of dust loading and particle size
distribution leaving the boiler on the par-
ticulates going to the gas turbine for the
particle removal systems selected are shown in
Table 6. The gas turbine specification is
Table 6. DUST LOADINGS LEAVING THE
SECONDARY CYCLONE FOR DIFFERENT
CASES EVALUATED
Cases
Evaluated
Group 1
Case 1
Case 2
CaseS
Case 4a
Case 5
Group 2
Casel
Case 2
Case 3
Total dust
loading leaving
secondary cyclone,
gr/scf
0.14
0.27
0.40
0.16
0.40.
0.27
0.46
0.67
Dust loading
for particles
> 2 /im, gr/scf
0.007
0.014
0.020
0.0006
0.019
0.013
0.022
0.032
Assuming same fractional collection efficiency for
the second secondary cyclone as that for the first
one. This assumption is too optimistic.
exceeded in several cases. In order to meet the
specification, alternative particulate removal
systems could be considered — granular bed
filters, electrostatic precipitators, ceramic
filters, etc.; additional mechanical collectors
could be used in series; or boiler operation
could be altered. The additional use of
mechanical collectors is the approach used to
evaluate additional costs. This approach was
selected since the equipment is available and
thus provides the best cost data, and because
the effect of boiler operating conditions on
particulate emission is difficult to project.
Table 7 presents a summary of the economic
implications.
If the gas turbine dust loading requirement
is defined as < 0.01 gr/scf without reference
to particle size, no centrifugal separator
presently available can meet this requirement
within reasonable cost in all cases discussed
above. If this were the case, high temperature
ceramic filters may have to be used. The
possibility of this application is being
evaluated. In this respect, the particle size dis-
tribution is a far more important parameter
than the dust loading, since the collection
efficiency for the particle size smaller than 2
l*m decreases rapidly. Thus, the particle size
distribution curves assumed for the present
evaluation (as shown in Figure 29) are con-
servative because a large amount of fines is
assumed to be present.
Effect of Total Gas Flow Rate on Particulate
Removal Equipment Cost
The effect of increasing gas flow rate was
evaluated for four cases which correspond to
the basic design flow rate, 30, 50, and 100
percent air. The selection of first stage
cyclones is based on the criterion of maximum
efficiency at minimum cost with a minimum
cyclone efficiency of 85 percent. The cost
increment for higher gas flow rate is shown in
Figure 33.
The first stage cyclone cost includes not
only the cost of the first stage separators
supplied by Ducon but also the cost of the
separator pressure vessel and all of the gas
piping from the steam generator outlet to the
secondary separator inlets. The costs for
pressure vessel and gas piping were estimated
for two cases. In one case, the gas piping from
the steam generator to the first stage separator
is lined with hard refractory, but the pressure
vessel and the gas piping from it are lined with
stainless steel. In the other case, hard refrac-
tory without an alloy liner was used through-
out. At 100 percent excess air, the increase is
$1.70/kW and $2.40/kW for hard refractory
liner and stainless steel liner respectively. The
major cost increase is from enlargement of
pressure vessels and gas piping due to higher
gas flow rate. The increase in separator cost
alone constitutes only about 18 percent of the
total cost increment.
IV-2-16
-------
Table 7. PARTICULATE REMOVAL SYSTEM COST
Case
Group 1
Casel
Case 2
Case 3
Case 4
CaseS
Group 2
Casel
Case 2
CaseS
Need for further clean-up to
achieve gas turbine specification
No
May not be required3
Yes, a second secondary cyclone in series
or granular bed filter
No
Yes, a second secondary cyclone in series
or granular bed filter
May not be required3
Yes, a second secondary cyclone in series
or granular bed filter
Yes, a second secondary cyclone in series
or granular bed filter
'articulate removal system
cost increase, $/kW
-
6.30
8.50
-
6.30
8.50
-
6.30
8.50
6.30
8.50
3 Dust loading is larger than specification but particle size is close to
specification.
The cost increase for the second stage is
presented in Figure 34. The incremental cost
for the second stage is more than 2 times that
of the first stage ($5.00/kW versus $2.40/kW
at 100 percent excess air). This is because
model 18,000 is the largest cyclone now
supplied by Aerodyne. The capacity of model
18,000 with dirty gas as the secondary gas is
30,000 ftVmin. Any gas flow rate higher than
that will require multiple units with their
individual pressure vessel. This tends to
increase the incremental cost of the second
stage; however, the rate of cost increase slows
down at excess air larger than 100 percent.. At
excess air larger than 100 percent, the rate of
cost increase for the first stage speeds up. This
is because increases in pressure vessel size and
gas piping diameter increase the incremental
cost rapidly at very high gas flow rate.
Taking into account the cost of gas piping
from the second stage cyclones to gas turbine,
the total incremental cost at different gas flow
rate is shown in Figure 35.
Combining cost figures from Figures 15,
16, and 35, an increase in total boiler cost of
$9.6 to S10.2/kW is required to operate the
boiler at 100 percent excess air. This, however,
does not take into consideration that the
combustion efficiency in the primary beds will
approach 100 percent at 100 percent excess air
and that the carbon burn-up cell can be
eliminated. In addition, the particulate
removal system would be much simpler and
the high excess air may also provide the
necessary flexibility to achieve plant turndown
if the boiler is designed at lower bed
temperatures. An additional ~ 10 percent
turndown capability is obtained by operating
at 100 percent excess air. The lower bed
depths would also permit combining the two
superheater beds into one bed which would
also reduce module height and cost. Thus it is
concluded that if the carbon burn-up cell can
be eliminated, operating at 100 percent excess
air may result in a lower energy cost with
increased cycle efficiency and flexibility. This
conclusion is true at 300-MW nominal plant
IV-2-17
-------
size, but is not necessarily true at 600-MW
plant size. This is because at 600-MW plant
size, the module has a 17-ft diameter at the
basic design. Further increase in excess air
requires either an increase in module diameter
or in the total number of modules. At 100
percent excess air, four modules with 23-ft
diameter are required. Since fabrication cost
of the internals increases rapidly at module
diameters larger than 17 feet because of
rapidly decreasing shop-fabricable portion, an
increase of $5.0/kW in boiler module cost
alone is conceivable. Adding the cost increase
in boiler cost is about $15.0/kW. In this case,
operating at higher pressures may be
beneficial.
Instead of increasing the module diameter
from 17 to 23 feet, the number of modules
could be increased keeping the module
diameter constant. At 100 percent excess air,
seven modules are required. In this case, in
addition to module cost individual particle
removal equipment has to be provided for
each module; ducting and piping manifolds
have to be increased; coal feeding systems
become more complicated; and above all,
instrumentation and control have to be more
sophisticated. The total increase in boiler cost
is estimated to be $20.0 to $25.0/kW in this
case. Consequently, even if the carbon burn-
up cell could be eliminated, the increase in
cost and complication in control do not clearly
favor the operation at high excess air for large
plant size. A careful evaluation of overall
design and control philosophy should be done
if operation at high excess air is to be
attemped for large plant size.
POWER GENERATION EQUIPMENT
Alternative boiler design and operating
conditions will have three primary affects on
the power generation equipment:
1. Capacity of gas and steam turbine
equipment.
2. Gas turbine inlet temperature.
3. Ability to achieve higher steam
temperatures and pressures.
IV-2-18
Since this analysis assumes a constant fuel
rate, the capacity changes are considered
small except for the high excess air case. The
effect of capacity was considered with the
excess air analysis.
Gas Turbine Inlet Temperatures
The design value for the gas turbine inlet
temperature was 1600°F for the base design of
the pressurized fluid-bed boiler, which is well
below the current state-of-the-art tempera-
tures of 1800-1900°F for utility intermediate
load applications. The 1600°F level was
established by assuming the flue gas leaves the
bed at the 1750 °F bed temperature, and that
the temperature difference between the bed
and the gas turbine inlet would be 150°F.
Several factors may alter the gas turbine inlet
temperature. These include:
1. Temperature drop between boiler and
turbine expander—150°F was assumed
which is probably excessive. A drop as low
as 50 to 75 °F may be achieved.
2. Bed temperature—if the boiler design
temperature is changed, the turbine inlet
temperature will change. Sulfur removal
considerations and ash agglomeration will
determine the maximum temperature.
3. Combustion above the bed—combustion
has been observed above the bed of a
fluidized-bed boiler which increases the
gas temperature 200-300°F. Any combus-
tion above the bed would increase the gas
turbine inlet temperature. No combustion
was assumed in the base design.
4. Modification of the cycle to provide for
reheat of the product gas from the boiler
prior to the gas turbine. One concept for
doing this is shown in Figure 36. Carbon
carried out of the primary beds would be
gasified to produce a low-Btu gas. The gas
would be used in the second stage
combustor.
Performance calculations were made to
determine the effect of a change in the gas
turbine inlet temperature on plant
-------
performance. The results are summarized in
Table 8. These results are for 17.5 percent
excess air and a boiler efficiency of 88.6
percent. A 200° F change in the turbine inlet
temperature will change the plant heat rate
~1 percent using current technology.
Table 8. PERFORMANCE OF PRESSURIZED
FLUID-BED BOILER POWER PLANT AS A
FUNCTION OF GAS TURBINE INLET
TEMPERATURE
Gas turbine
inlet temperature.
°F
1400
1500
1600
1700
1800
1900
2000
Plant
output,3
MW
625.6
634.9
644.1
644.8
645.3
641.8
636.2
Plant
heat rate,3
Btu/kWhr
9293
9157
9026
8921
8820
8773
8753
Fuel burned
after primary
beds,b
%
-
2.5
5.0
7.6
10.1
3The decrease in performance at high gas turbine
inlet temperatures is the result of increased bleed
air required for turbine cooling and the increase in
gas turbine waste heat which reduces steam cycle
extraction for regenerative heating. The heat rate
at 2000°F would be reduced ~ 270 Btu/kWhr if
turbine blade cooling was not required.
b Assumes any increase above the design value of
1600°F .has to result from burning fuel either
above the bed or separately.
Steam Temperature
Plant performance can be increased by
increasing steam temperature and pressure.
The effect of higher steam temperatures on
the performance of the plant is shown in Table
9.
Table 9. EFFECT OF HIGHER STEAM
TEMPERATURES ON PLANT PERFORMANCE
Steam
temperature,
°F
1000/1000
1100/1100
1200/1200
Steam
pressure,
psi
2400
3300
4500
Gas turbine
inlet temperature,
°F
1600
1600
1600
Power.
MW
644.1
673.9
690.8
Heat rate,
Btu/kWhr
9026
8627
8417
An increase of 100°F in both superheat and
reheat temperatures will give a reduction of
about 400 Btu/kWhr in plant heat rate. The
increased performance and inherently less
severe boiler tube corrosion in fluidized-bed
boilers make high steam temperatures attrac-
tive.
ASSESSMENT
A summary of the sensitivity analysis is
presented in Table 10. Each parameter is
indicated with a projection of what change
might be required as the result of experi-
mental data. For example, the bed
temperature was set at 1750°F for 100 percent
load. This temperature may prove to be too
high for economic sulfur removal and have to
be reduced to 1600 or 1650°F which may be
more favorable. The summary table indicates
how such a change would affect plant cost and
performance assuming no other variable
restrictions. In this case, the plant cost would
increase < $3/kW, plant performance would
decrease < 0.5 percent, and plant turndown
to 25 percent load could still be achieved. This
would result in an energy cost reduction of
< 0.3 mills/kWhr. Since the projected advan-
tage over a conventional plant with stack gas
scrubbing is greater than 1.5 mills/kWhr,1 the
penalty for lowering the temperature is not
significant. This conclusion holds for all of the
variables considered. The conclusion is
generally valid for plant capacities of 200 to
700 MW. The results of the sensitivity analysis
indicate that the base plant design is relatively
insensitive to changes in operating conditions
or design parameters.
Operating conditions and design
parameter changes can occur which would
result in a significant cost increase for the
system. This would most likely occur as the
result of additive problems. For example,
suppose the bed temperature had to be
decreased to 1600°F, the heat transfer
coefficient was only 35 Btu/hr-ft2-°F, and the
dust loading from the boiler was 3 times the
design value. The plant cost could increase by
7 to 8 percent and the efficiency decrease by
IV-2-19
-------
Table 10. SENSITIVITY ANALYSIS SUMMARY
Parameter
Boiler operating conditions
Bed temperature
Fluidizing velocity
Excess air
Pressure
Paniculate carry-over
Loading
Panicle size
Boiler design
Heat transfer surface
Configuration
Change
Reduction from
1750°Fto16rjO°F
Decrease to 5 ft/sec
Increase to 15ft/sec
Increase to 100%
Increase to 15 atm
Increase loading
to 3 times the
design value
(
-------
Table 10 (continued). SENSUIVrTY ANALYSIS SUMMARY
Parameter
Heat transfer
coefficient
Materials
Bed depth i
Power generation equipmentk
Gas turbine inlet
temperature
Steam temperature
Change
Decrease from
5.5ft2/ft3to
3 ft2/ft3
Increase from
SO to 75
Btu/hr-ft2.°F
Decrease from
50 to 35
Btu/hr-ft2-°F
Assume tubing
cost 50% greater
than base design
Reduced to 10 ft
±200°Ffrom
1600°F
100°F increase
in superheat
and reheat
Boiler
oj$1/kW increase9
$l/kW decrease
""vSS/kW increase
<$2/kW increase
~$3/kW increase
<$2/kW increase
PJantcost
Auxiliaries
Effect
Power generation
equipment
Negligible
Performance
± 1% in efficiency
2% increase in
efficiency
aIncrease will depend on bed depth restrictions; $3/kW would correspond to a maximum allowable height of ~15 ft.
"The fluidizing velocity will affect the particulate removal equipment — see participate emission parameter for costs.
Considerable savings may be realized for large capacity (>600 MW) plants since higher velocities avoid the need for field erection.
"Additional savings would be realized for large capacity (>600 MW) plants in order to avoid field erection. The $1/kW does not include a cost reduction
which may result from elimination of the carbon burn up bed due to increased efficiency.
e$4/kW assumes the larger capacity gas turbine has the same unit cost as the base machine. Actual cost $1/kW of a larger machine would be lower.
'Based on projected gas turbine requirement of <0.01 gr/scf of particles <2 (im. ,
9Assumes maximum allowable bed depth of 20 ft.
"Includes effect on fabrication.
'Assumes constant freeboard. (Change in freeboard requirement would have similar effect.)
'If the temperature is the result of equipment modification, the capital cost would be altered. Gas piping will be affected in any case.
kThe plant cost projections are based on the base case plant capacity and performance. An increase in efficiency will reduce the specific plant cost. Water
cooling has not been included. A higher efficiency will result in lower cooling equipment cost.
-------
~ 0.5 percent. This would result in an
increase in the energy cost of ^ 0.7
mills/kWhr. This is a significant increase but
still within the economic margin. Caution
must be exercised in interpreting multiple
changes in the variables. In the case above, the
cost effects were added. However, decreasing
the bed temperature and increasing the heat
transfer coefficient both increase the heat
transfer surface and the bed depth if the bed
area is maintained constant. Any restrictions
on bed depth would also have to be considered
in the evaluation. Parametric curves have been
prepared to enable this type of evaluation to
be made.
The effect of boiler plant pressure drop and
the steam turbine condenser pressure on plant
performance has been presented.2 The boiler
plant pressure drop has a small effect on plant
capacity and heat rate: 1 percent increase in
pressure drop results in a 0.1 percent increase
in plant heat rate. An increase in the steam
turbine condenser pressure from 1-1/2-in. Hg
to 3-in. Hg for a cooling tower results in a 2.5
percent increase in heat rate.
The potential performance of the plant was
evaluated by assessing the effect of higher gas
turbine inlet temperatures and higher steam
temperatures and pressures. The results
indicate that plant efficiencies of ~ 45 percent
can be achieved with gas turbine inlet temper-
atures greater than 2000°F, with high
temperature blade material to minimize
cooling requirements, and with steam
temperatures of 1200°F.
Advances in boiler plant subsystem con-
cepts have not been considered, in this
analysis. Cost reductions may be achieved by
using alternative concepts. The following
components have been studied for potential
savings:
Particulate Removal—The projected
system utilizes four secondary collectors
for final particle removal before the gas
turbine. Alternative systems are being
considered which may reduce the number
and size of the units. Cost estimates
indicate that reductions of $1 to $5/kW
for the total particulate removal system
may be possible.
High-Temperature Gas Piping—The high-
temperature gas piping cost would be
reduced if the particulate removal system
were simplified by using fewer units per
module. Additional savings might be
realized if refractory lined pipe could be
used between the secondary collectors and
the gas turbine. The present design uses a
high-alloy steel to assure protection of the
gas turbine from additional particulates.
Coal Feeding System—The coal feeding
system design is 'based on systems which
have been built and operated. The design
provides a separate coal feeding system for
each fluidized bed in order to assure
control of the coal feed rate to each bed. It
may be possible, however, to reduce the
number of coal feed systems from 16 to 4 if
independent control of solids flow to each
bed in a module can be achieved from a
single pressurized injector. The potential
cost reduction is estimated to be >
$2/kW.
Stack Gas Cooler Design—Cost estimates
were obtained for the stack gas coolers,
but no attempt was made to optimize the
design or consider nonconventional
designs, such as those using fluidized
beds. Preliminary conceptual evaluation
indicates that the cost might be reduced
$4.40 to $3.40/kW.
CONCLUSIONS
1. Base plant design is near optimal.
2. Pressurized fluidized-bed boiler power
plant maintains greater than 5 to 10 percent
energy .--ost advantage over conventional plant
with stack gas scrubbing.
Effect of a potential change in bed
temperature, fluidizing velocity, heat
transfer surface configuration, gas side
heat transfer coefficient, boiler tube
IV-2-22
-------
materials, bed depth limitations or
pressure will result in: no significant
change in plant operability or
performance, < 2 percent increase in plant
cost, and < 0.2 mills/kWhr increase in
energy cost. Effect of increasing the dust
loading to three times the design value will
increase the plant cost ~ 4 percent, energy
cost < 0.3 mills/kWhr.
Effect of increasing excess air will result
in: increased turndown capability and per-
formance, 3 to 6 percent increase in plant
cost, and no significant change in energy
cost.
3. Plant efficiencies of ~ 45 percent may be
achieved for gas turbine inlet temperatures
above 2000°F and steam temperatures of
1200°F.
ACKNOWLEDGMENT
This work was performed under contract to
the Environmental Protection Agency, Office
of Research and Monitoring. P. P. Turner
served as contract officer. N. E. Weeks,
Westinghouse Power Generation Systems
Division, performed power system cycle
analyses.
REFERENCES
1. Evaluation of the Fluidized Bed
Combustion Process. Volumes I-III.
Westinghouse Research Laboratories,
Pittsburgh, Pa. Submitted to the Office of
Air Programs, Environmental Protection
Agency, Research Triangle Park, N. C.
November 1971.
2. Keairns, D. L., J. R. Hamm, and D. H.
Archer. (Presented at Annual AIChE
Meeting. San Francisco. November 1971.
AIChE Symposium Series, Volume "Air",
1972.
3. Highley, J., D. Chandrasekeva, and D. F.
Williams. Fluidized Combustion Section,
Coal Research Establishment, National
Coal Board, London, England. Report
Number 20, April 1969.
4. McLaren, J. and D. F. Williams. J. of
Institute of Fuel. 303, August 1969.
5. Smith, S. Private communication, Boiler
Tube Division, The Babcock & Wilcox Co.,
Barberton, Ohio.
6. Genetti, J. W. and W. E. Bartel. (Presented
at American Institute of Chemical
Engineers 72nd National Meeting. St.
Louis. May 22-24, 1972.)
IV-2-23
-------
to
•u
COMPRESSOR POWER
1 COMPRESS
"R-fL>
GAS FLOW 1
== SOLIDS FLOW j
STEAM CYCLE '
1
i SULFUR I
[ RICH GAS 1
SPENT ,
' DESULFURIZING
OR TURBINE TURBINE
i
PRIMARY
REMOVAL
1
tf
\J \l
^ SECONDARY
REMOVAL
DESULFURIZING AGENT FLUJ!?!PD CARBON
AGENT REGENERATED rnwiRiKTnR BURN"UP
REGENERATOR STONE „ ngSjl^gfe, CELL
1 ft
1
1
IHttSTF ! MAKE-UP
SToJll STONE
1 COAL
1 (OIL)
i
1
1
1.
~TT
S\ 1 .XI ^
-------
REHEATED STEAM
REHEATER BED
SUPERHEATED STEAM
SUPERHEATER BED
SUPERHEATER BED
PRE-EVAPORATOR BED
FEED WATER
PLANT
SIZE
320 MW
635 MW
VESSEL
DIAMETER
12 ft
17ft
GRADE ELEVATION
ELEVATION
Figure 2. Pressurized fluidized-bed steam generator for combined cycle plant (four required).
IV-2-25
-------
2.0
1.5
^SHELL
81.0
HOME OFFICE (ENGINEERING,
CONTRACT RESERVE, ETC.
ERECTION
0.5
300
CONTRACTED AND
•SUBCONTRACTED
EQUIPMENT
' DRAFTING
3.5
3.0
"&
i—i
x
? 5
•> *•••*
2.0
1.5
' PRESSURE PARTS
500
600
400
500
600
700 800 300
PLANT SIZE, MW
Figure 3. Steam generator cost breakdown for the W-FW basic design.
700 800
IV-2-26
-------
300
PLANT POWER
STEAM TURBINE POWER
200
ASSUME BED DEPTHS ARE
CONSTANT AT BASIC DESIGN VALUES
I
o
100
GAS TURBINE POWER
1300
1400
1500
1600
1700
BED TEMPERATURE, °F
Figure 4. Effect of bed temperature on power generation.
1800
25
Q_
a
LLJ
oa
o
UJ
20
15
10
1300
1400
1700
1500 1600
BED TEMPERATURE, °F
Figure 5- Bed depth requirement at different bed temperatures.
1800
IV-2-27
-------
150
100
LU
31
LU
O
O
so
1300 1400 1500 1600 1700
BED TEMPERATURE, °F
Figure 6. Effect of bed temperature on total module height.
1800
IV-2-28
-------
I
ce
UJ
9
C3
i
LU
0
lb
BASIC DESIGN POINT
BED-TUBE HEAT TRANSFER COEFFICIENT: 50 Btu/ft2 -hr-°F
CURVE MAX. ALLOWABLE BED DEPTH
la
lb
2
3a
3b
20ft
UNRESTRICTED
20ft
20ft
UNRESTRICTED
TUBE ARRANGEMENT
I'/Hn. OD AT H=7 in. AND V=3 in. IN
PRE-EVAPORATOR AND SUPERHEATER
2-in. OD AT H=7 in. AND V=4 in. IN
REHEATER
SAME
Hn. OD AT H=4 in. AND V=2 in. IN
ALL BEDS
2-in. OD AT H=8 in. AND V=4 in. IN
ALL BEDS
SAME
1300
1400
1700
1500 1600
BED TEMPERATURE, «F
Figure 7. Effect of bed temperature on the steam generator cost of 318-MW plant
(not including erection).
1800
IV-2-29
-------
10
O
X
GO
O
O
(XL
O
DC
1470 "F
BED-TUBE HEAT TRANSFER COEFFICIENT: 50 Btu/ft2-hr- ° F
TUBE PITCH/DIAMETER RATIO: BASIC DESIGN
0
10
20
30
MAXIM ALLOWABLE BED DEPTH, ft
Figure 8. Dependence of the steam generator (318-MW) cost on the maximum allowable
bed depth ( not including erection).
IV-2-30
-------
100.5
UJ
V)
01
D-
O
o
Q.
100.0
99.5
99.0
TUBE PITCH/DIAMETER RATIO
BASIC DESIGN
J-in. OD AT H=4 in., V=2 in.
2-in. OD AT H=8 in., V=4 in.
THIS IS AN ADDER APPLIED TO FIGURE 4
BED-TUBE HEAT TRANSFER COEFFICIENT.
50 Btu/ftz-hr-°F
50 Btu/ft2-hr-°F
50Btu/ft2-hr-°F
1300
1400
1700
1800
1500 1600
BED TEMPERATURE, "F
Figure 9. Change in net plant power output due to change in pressure drop across the bed.
IV-2-31
-------
20
15
O
3
UJ
O
LLl
D.
10
BED TEMPERATURE: 1750 °F
TUBE PITCH/DIAMETER RATIO: BASIC DESIGN
BED-TUBE HEAT TRANSFER COEFFICIENT: 50 Btu/ft2 - hr -°f
SUPERFICIAL
VELOCITY
SUPERHEATER
PRE-EVAPORATOR
SUPERHEATER
REHEATER
PRE-EVAPORATOR
60
UJ
a
o
UJ
00
40
20
REHEATER
?
I
10
30
20
BED AREA, ft2
Figure 10. Dependence of bed depth on the fluidizing velocity.
40
IV-2-32
-------
300
250
200
C9
LU
150
100
50
BED TEMPERATURE
1407°F
1522 °F
MAXIMUM ALLOWABLE BED DEPTH: UNLIMITED
TUBE PITCH/DIAMETER RATIO: BASIC DESIGN
BED-TUBE HEAT TRANSFER COEFFICIENT:
50 Btu/ftZ-hr-°F
16
15
14
13
12
11
10
30
40
20
BED AREA, ft2
Figure 11. Dependence of module diameter and module height on the bed area.
10
IV-2-33
-------
•€,
i—I
x
§
3.0
2.5
2.0
1.0
COST ESTIMATION BY FW
.—..-._ INDEPENDENT ESTIMATION BY WESTINGHOUSE (WITHOUT /
INTERNAL SUPPORT ALLOWANCE) '
INDEPENDENT ESTIMATION BY WESTINGHOUSE (WITH S
INTERNAL SUPPORT ALLOWANCE) /
! 4 6 8 10 12 14 16
MODULE DIAMETER, ft
Figure 12. Shell cost (four modules) at basic design height.
18
20 22
IV-2-34
-------
'1750°F
635 MW
O
CJ
COST INCREASES K160.000 FOR BASIC
DESIGN CASE ) CAUSED BY SHIFT FROM
SHOP-ASSEMBLED TO FIELD-ERECTED
VESSEL AT VELOCITIES <8 ft/sec
318 MW
1407 °F
1522 °F
1636 °F
1750 °F
CURVE TUBE PITCH/DIAMETER RATIO
BASIC DESIGN
1-in. 00 TUBES AT H= 4 in. AND V=2 in.
BED-TUBE HEAT TRANSFER COEFFICIENT: 50 Btu/ft2.hr- °F
MAXIMUM ALLOWABLE BED DEPTH: UNLIMITED
10
15
FLUIDIZING VELOCITY IN PRE-EVAPORATOR, ft/sec
Figure 13. Dependence of the steam generator cost on the fluidizing velocity.
IV-2-35
-------
2.5
BED TEMPERATURE: 1750 °F
•
CURVE TUBE PITCH /DIAMETER RATIO
BASIC DESIGN
2.0
lin.-OD TUBES AT H = 4 in.
AND V= 2 in.
CURVE 1: COST REDUCTION DUE TO DECREASE
IN MODULE DIAMETER.
CURVE 2: COST ESCALATION DUE TO
INCREASE IN MODULE HEIGHT
o
1.5
1.0
635 MW
318 MW
0.5
5 10 15
FLUIDIZING VELOCITY IN PRE-EVAPORATOR, It/sec
Figure 14. Dependence of the shell cost on the fluidizing velocity.
IV-2-36
-------
s
Q£
o
0.7
0.6
0.5
0.4
0.2
0.1
0
CONSTANT FLUIDIZING
VELOCITY
•— CONSTANT BED
AREA
10
20
80
90
100 110
30 40 50 60 70
EXCESS AIR IN PRIMARY BEDS, %
Figure 15. Cost reduction in heat transfer surface and pressure shell at different air
flow rates.
o
THESE WOULD NOT BE CONTINUOUS
CURVES IN PRACTICE, BECAUSE TURBINES
ARE ONLY MANUFACTURED IN SELECTED
SIZE RANGES
5
a1
a
UJ
on
g
o
IB 20 10 40 so §0 n so so 100 no 120
OVERALL IXCESS AIR (INCLUDING AIR FROM CBC), %
18, 0§§t §§&§l§tl@n in gas turbine equipment and cost reduction in steam
turbifli equipminl at different air flow rates.
IV-2-3?
-------
9500
BOILER PRESSURE =2400psi
BOILER EFFICIENCY =88.
8900
10
25
30
OVERALL PRESSURE RATIO
9400
9300
9200
2 9100
8900
15 20 25
OVERALL PRESSURE RATIO
Figure 17. Plant heat rate versus compressor pressure ratio.
30
IV-2-38
-------
4 in.
-*»
X
3'/2in.
=> p.
1
^r
$
f"^
^
r
— -
"*•
^
,
T
O
/^
O
Q
s
2-in. OD WATER
^ WALL TUBES ~\
3 in.
^>- 2-in. OD TUBES
<
<
y/2 in.
c
'
P
6
o
^
C
V.
•* —
I
)
s
O
O
o
Q
tf
o
^
2-in. OD WATER
^ WALL TUBES
>-l'/2-in.OD TUBES
REHEATER PRE-EVAPORATOR AND SUPERHEATERS
Figure 18. Tube arrangement in basic design.
r)
O ^
^—(£)
O
O
^
O O O
O O
O O O
O O
STAGGERED OR ROTATED DIAMOND
ARRANGEMENT
© ©
1 ' ^
0— -6
o o
o o
o
o
o
o
SQUARE OR RECTANGULAR
ARRANGEMENT
Figure 19. Definition for different tube arrangements.
IV-2-39
-------
o
BED TEMPERATURE: 1750 °F, MAXIMUM ALLOWABLE BED DEPTH: 20ft,
BED TUBE HEAT TRANSFER COEFFICIENT: 50 Btu/ft2 . hr. -p
TUBE BENDING, WELDING,
' WATER WALLS
FABRICATION
DRAFTING,
HOME OFFICE, ETC.
SHELL
TUBING, HEADERS,
COMERS, RISERS
<§>
t—I
X
(s>
H-"
O
I I I
TOTAL STEAM
GENERATOR COST
6
12 14
8 10 12 14 2 4 6 8 10
AVAILABLE HEAT TRANSFER SURFACE/UNIT BED VOLUME, ft2/ft3
Figure 20. Change of the steam generator cost with heat transfer surface per unit bed
volume (not including erection).
IV-2-40
-------
10
x
,_ e
fr Q
uj 4
UJ
o
TUBE PITCH/DIAMETER RATIO
1-1/2 in. OD AT H»7 in., V-3 in.
IN PRE-EVAPORATOR AND SUPER-
HEATER _
2-in. OD AT H=7 in., V=4 in.
IN REHEATER
1-in. OD AT H«4 in., V*2 In.
IN ALL BEDS
2-in. OD AT H»8 In., V»4 in.
IN ALL BEDS —
BED TEMPERATURE: 1750° F
BED-TUBE HEAT TRANSFER
COEFFICIENT: 50 Btu/ft2-hr-°F
10
o
g
5
LU
BED TEMPERATURE: 1636 °F
BED-TUBE HEAT TRANSFER
COEFFICIENT: 50 Btu/ft2-hr-°F
SEE FIGURE 21 FOR CURVE DESIGNATION
10 20
MAXIMUM ALLOWABLE BED DEPTH, ft
30
Figure 21. Dependence of the steam
generator (318-MW) cost on the maxi-
mum allowable bed depth (not includ-
ing erection),
10 20
MAXIMUM ALLOWABLE BED DEPTH, ft.
30
Figure 22. Dependence of the steam
generator (318-MW) cost on the maxi-
mum allowable bed depth (not includ-
ing erection),
IV-2-41
-------
50
40
£
E 30
o
o
CO
o
UJ
ca
20
10
60mm
TUBES
A
O
35 mm
TUBES
V
A
NO. OF
ROWS
1
2
3
3
4
PITCH/DIAMETER
RATIO
2
4
4
6
PITCH/DIAMETER RATIO
Figure 23. Change of bed-tube heat transfer coefficient with pitch/diameter ratio.3
50
". 40
UJ
o
o
30
20
UJ
LU
m
a
UJ
ca
10
60mm
TUBES
A
O
35mm
TUBES
V
A
NO. OF
ROWS
2
2
3
3
PITCH/DIAMETER
RATIO
2
4
4
6
8
24 6 8 10
NARROWEST GAP BETWEEN TUBES, in.
Figure 24. Change of heat transfer coefficient with tube spacing.3-4
12
IV-2-42
-------
te.
is
1
CURVE TUBE PITCH/DIAMETER RATIO
1 BASIC DESIGN
2 1-in. OD AT H=4 in., V=2 in.
3 2-in. OD AT H=8 in., V=4 in.
BED TEMPERATURE: 1750°F
MAXIMUM ALLOWABLE BED DEPTH: 20 ft
BED AREA: 35 f(2
10
20
30
40
50
60
70
80
90
100
110 120
BED-TUBE HEAT TRANSFER COEFFICIENT, Btu/ft -hr- °F
Figure 25. Effect of bed-tube heat transfer coefficient on the steam generator cost
(not including erection).
IV-2-43
-------
8
o
TUBE PITCH/DIAMETER RATIO
BASIC DESIGN
1-in. OD AT H.4 in., V=2 in.
2-in. OD AT H=8 in., V=2 in.
BED TEMPERATURE: 1636° F
MAXIMUM ALLOWABLE BED DEPTH: 20 ft
BED AREA: 35 ft2-
10
20
30
40
50
60
70
80
90
100
110 120
BED-TUBE HEAT TRANSFER COEFFICIENT, Btu/ft -hr- °F
Figure 26. Effect of bed-tube heat transfer coefficient on the steam generator cost
(not including erection).
IV-2-44
-------
10
8
O
LU
tu
O
1
te
CURVE TUBE PITCH/DIAMETER RATIO
1 BASIC DESIGN
_ 2 1-in. OD TUBES AT H=4 in., V=2 in.
BED-TUBE HEAT TRANSFER
COEFFICIENT: 50 Btu/ft2-hr-°F
BED TEMPERATURE: 1636 "F
MAXIMUM ALLOWABLE
—BED DEPTH: 20ft
FIELD ERECTED
WITH MAXIMUM —I
SHOP FABRICATION
(FOUR MODULES)
// MUTIPLES OF LARGEST
'' SHOP FABRICATED
MODULES
10
i-H
X
S
k
0 100 200 300 400 500 600 700
PLANT SIZE, MW
Figure 27. Dependence of the steam generator
cost on the plant size.
CURVE TUBE PITCH/DIAMETER RATIO
1 BASIC DESIGN
_ 2 1-in. OD TUBES AT H-4 in., V-2 in.
BED-TUBE HEAT TRANSFER
COEFFICIENT: 50 Btu/ft2-hr-°F
BED TEMPERATURE: 1750°F
MAXIMUM ALLOWABLE
r~ BED DEPTH: 20 ft
0
FIELD ERECTED
WITH MAXIMUM -
SHOP FABRICATION
(FOUR MODULES
- y
,
/i MUTIPLE OF LARGEST
SHOP FABRICATED
MODULES
0 100 200 300 400 500 600 7 0
PLANT SIZE, MW
Figure 28. Dependence of the steam
generator cost on the plant size.
IV-2-45
-------
TYPE SCREEN NO.
0.1
1.0
5.0
15.0
25.0
35.0
45.0
55.0
65.0
400 200 100 60 40
20
10
« 75.0
g
oE 85.0
93.0
96.0
98.0
99.0
URVES 1,2,&3 REPRESENT PROJECTED PARTICLE
SIZE DISTRIBUTION LEAVING FLUID BED
BOILER SYSTEM (SEE TABLE 5)
JJJ
I
CURVE 4 REPRESENTS PROJECTED PARTICLE SIZE
DISTRIBUTION LEAVING PRIMARY COLLECTORS—1*
(GROUP 1)
CURVE 5 REPRESENTS PROJECTED PARTICLE SIZE
DISTRIBUTION LEAVING SECONDARY
COLLECTORS(GROUP 1)
I I I I MI 11 !
I !
1 10 100 1000
PARTICLE SIZE, urn
Figure 29. Particle size distribution for different gas streams.
0.5
3.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
.0
97.0
98.5
IV-2-46
-------
s.c.
FLUIDIZED
BED
COWI8USTOR
(FBC)
CARBON
BURNUP
CELL
(CBC)
f
P. C.= PRIMARY CYCLONE
S. C.= SECONDARY CYCLONE
Figure 30. Flow diagram for particulate removal system.
P.C.
FLUIDIZED
BED
COMBUSTOR
(FBC)
T
CARBON
BURNUP
CELL
(CBC)
P.C.= PRIMARY CYCLONE
S.C.= SECONDARY CYCLONES OF SAME
FRACTIONAL COLLECTION EFFICIENCY
Figure 31. Flow chart for .group 1, case 4.
IV-2-47
-------
FLUIDIZED
BED
COMBUSTOR
(FBC)
\
CARBON
BURNUP
CELL
(CBC)
•• PRIMARY CYCLONES OF SAME
FRACTIONAL COLLECTION EFFICIENCY
S. C,=SECONDARY CYCLONE
Figure 32. Flow chart for Group 1, Case 5.
<
•«o-
cc
Ul
a
a
o
Of
01
UJ
cc
cc
a
o
LU
O
§
cc.
1
^ 20 30 40 50 60 70 80 90 100 110
cc
2: EXCESS AIR IN BED, %
Figure 33. Increase in first stage cyclone
cost due to increase in gas flow rate (for
318-MW plant).
20 30 40
a:
£
50 60 70 80 90 100 110
EXCESS AIR IN BED, %
Figure 34. Increase in second stage cyclone
cost due to increase in gas flow rate (for 318-
MW plant).
IV-2-48
-------
o
a
O
O
O
UJ
oc
3
o
0
1: FIRST STAGE PRESSURE VESSELS
AND GAS PIPING ARE LINED WITH
STAINLESS STEEL
2: ALL LINED WITH HARD REFRACTORY
10
20
30
70
80
90
100
40 50 60
EXCESS AIR IN BED, %
Figure 35. Cost increments for participate removal system at different gas flow rates.
110
IV-2-49
-------
s
SORBENT OUT
AIR
COMB. PROD.
FUEL GAS
— •—STEAM AND WATER
SOLIDS
SORBENTIN
COAL
TO STACK
Figure 36. Pressurized fluid-bed combustion power plant with secondary combustor.
-------
3. APPLICATION TO COMBINED CYCLE POWER
PRODUCTION OF FLUID-BED TECHNOLOGY USED
IN NUCLEAR FUEL REPROCESSING
B. R. DICKEY AND J. A. BUCKHAM
Allied Chemical Corp.
ABSTRACT
Fluid-bed processing is used extensively at the Idaho Chemical Processing Plant (ICPP) in the
recovery of uranium from spent nuclear fuel elements. Fluid-bed denitration of uranyl nitrate
solutions and fluid-bed solidification of radioactive waste solutions are used routinely in plant
operations. In addition, current pilot-plant development of processes for recovering uranium from
graphite-based fuels depends largely on fluidized-bed combustion.
During the course of fluidized-bed operations and development at ICPP, advantageous appli-
cation of fluid-bed technology in areas other than nuclear fuel reprocessing has become apparent.
Current and near-term fluidized-bed technology appears directly applicable to the conversion of
the energy in wood wastes and municipal refuse to electric power. Fluidized-bed operations and
process development at the ICPP and a concept for combined-cycle power production based
largely on fluidized-bed combustion are discussed herein.
INTRODUCTION
Allied Chemical Corporation's Idaho
Chemical Programs-Operations Office
operates the Idaho Chemical Processing plant
for the Atomic Energy Commission. The pri-
mary mission of the facility is to recover
uranium from spent nuclear fuel elements in
an economic and safe manner. Metallic clad
nuclear fuels are dissolved in inorganic acids,
and the uranium and fission product solutions
are separated by solvent extraction. Uranium
rich solutions are denitrated and solidified to
uranium oxide; fission product solutions are
calcined to a mixture of metallic and fission
product oxides.
Fluidized-bed processing is used in the
solidification of fission product solutions and
in denitration of uranium rich solutions.
Fluid-bed solidification of radioactive waste
solutions has been used since December 1963.
Denitration by thermal decomposition in a
fluidized bed was started in 1971. Both waste
calcination and denitration are endothermic
processes. In-bed combustion of kerosene
supplies the heat for waste calcination; heat
for fluid-bed denitration is obtained from
wall-mounted electrical heaters.
The capacity to recover enriched uranium
from graphite-based fuels will be required in
the near future. Pilot-plant development of
processes for separation and recovery of
uranium from these fuels is well underway.
The heart of these processes is the fluidized-
bed combustion of the graphite matrix.
IV-3-1
-------
In the course of plant operation and wastes and municipal refuse through the con-
process development using fluid-bed tech- version to electrical power. The proposed
nology at the ICPP, concepts based on the process is based on a combined gas turbine-
application of fluid-bed technology in non- steam turbine cycle; the key to .producing
nuclear areas have evolved. A proposed appli- power at high thermal efficiencies in the pro-
cation of current interest is the recycle of wood posed concept is a fluidized-bed combustor.
IV-3-2
-------
OPERATIONAL FLUID-BED PROCESSES
AT ICPP
Waste Calcining Facility
Radioactive wastes resulting from nuclear
fuel reprocessing can be solidified by various
methods that have been under investigation
over the past 20 years; however, only one pro-
cess has been demonstrated on a production
basis. The Waste Calcining Facility (WCF) at
the Idaho Chemical Processing Plant is the
first production facility in the world for
converting aqueous radioactive wastes to
solids using the fluidized-bed calcination
process.
A schematic flow sheet for the WCF is
shown in Figure 1. The heart of the process is
the 4-ft diameter calciner vessel where the
radioactive aqueous wastes are continuously
atomized into the fluid ized bed of oxide par-
ticles by an external-mix pneumatic atomizing
nozzle. In the fluidized bed, which is main-
tained at 500 °C, water evaporates, the
metallic salts are converted to their
corresponding oxides or fluorides which are
deposited layerwise on the spherical bed par-
ticles. The solid particles are withdrawn
continuously from the calciner vessel to main-
tain a constant bed height and are transported
pneumatically to stainless steel solid storage
bins adjacent to the calcining facility.
The off-gas leaving the calciner vessel
passes through a dry cyclone where the
majority of elutriated fines are removed and
pneumatically transported to the solids
storage bins. The off-gas then passes through
a wet scrubbing system consisting of a quench
tank, venturi scrubber, cyclone separator, and
a demister. Here, the off-gas is contacted with
a nitric acid scrubbing solution which removes
the majority of the remaining solids. The
scrubbing solution is recirculated
continuously, and any accumulation is
recycled to the waste feed tank.
Heat for the endothermic calcination
reactions is supplied by in-bed combustion.
In-bed combustion consists of atomizing a
hydrocarbon fuel (kerosene) with pure oxygen
directly in the fluidized bed. Startup of the
process is achieved by heating the fluidized
bed to temperatures in the range of 360 to
400°C using preheated fluidizing air. A
nitrate-containing waste is then injected
through a separate waste atomizing nozzle,
followed immediately by the injection of the
fuel-oxygen mixture through the fuel
atomizing nozzle. Ignition of the fuel-oxygen
mixture is spontaneous at temperatures above
335°C in the presence of nitrates. After the
startup, the wastes are calcined at
temperatures in the range of 400 to 500_°C
during routine operation. The advantage of
this method of heating is that no heat transfer
surfaces are involved that can foul or limit
capacity; the heat flow paths of in-bed
combustion heating and an in-bed heat
exchanger are compared in Figure 2.
Denitration facility
The denitration process is based on the
thermal decomposition of uranyl nitrate solu-
tion to uranium trioxide. In fluidized-bed
denitration, solution is continuously sprayed
through an air atomizing nozzle into heated
fluidized bed of UO3. The bed temperature is
300°C; the pressure immediately above the
support plate is approximately atmospheric.
Granular product is continually withdrawn
into a product collection vessel. The process
off-gas~consisting mainly of fluidizing air,
water vapor, and oxides of nitrogen-flows
through a filtering section consisting of three
sintered metal filters which remove over 99.9
percent of the entrained and elutriated UC*3
dust particles. The filters are blown back
intermittently, and the fine particles serve as
seed particles for particle growth. A schematic
flow-sheet of the fluid-bed denitration process
is shown in Figure 3.
FLUID-BED BURNING PROCESSES
Pilot-plant development of a fluidized-bed
combustion process for separating uranium
from spent graphite-matrix nuclear fuels has
been in progress since January 1966 at ICPP.
This unique combustion process is required
IV-3-3
-------
because graphite, unlike the metallic cladding
of the more conventional nuclear fuels, is not
readily dissolved in common inorganic acids.
As the graphite matrix is removed by combus-
tion, the uranium and other metals are
converted to their oxides. Dissolution of the
resulting uranium oxide, U3O8, completes the
burn-leach head-end process. The uranium is
separated from fission products and other
impurities by conventional solvent extraction.
The fuel consists of uranium dicarbide
microspheres coated with pyrolytic carbon
dispersed in a graphite matrix. A protective
coating of niobium carbide is present on some
of the surfaces of the fuel elements. Although
aluminum or stainless steel orifices are present
in some elements, there are only three
constituents of consequence to the combustion
process: uranium, niobium, and carbon.
The fluidized-bed combustion process
developed at ICPP for nuclear rocket fuels
involves the following concepts: (1) charging of
whole fuel elements to a fluidized bed of inert
alumina particles, (2) combustion of
essentially all carbon, (3) oxidation of uranium
and niobium carbides, and (4) elutriation of
uranium and niobium oxides from the burner.
Requirements for the fluidized-bed
burning and elutriation processes are: (1)
combustion of at least 95 percent of the matrix
graphite and pyrolytic carbon, (2) conversion
of the uranium dicarbide microspheres and
niobium carbide to elutriable particles by
oxidation and attrition, (3) negligible attrition
and elutriation of the inert bed material (a-
alumina), and (4) adequate heat dissipation
for control of bed temperatures. If "steady-
state" conditions can be achieved, all of the
uranium and niobium charged to the burner
will be elutriated in the burner off-gas. The
amount of alumina and unburned carbon
carried overhead with the U3O8 product must
be minimized.
The fluidized-bed burner segment of the
Graphite Fuels Pilot Plant (GFPP) (Figure 4)
was constructed to permit studies of the
combustion process. The burner and leaching
equipment, located on two adjacent modules,
can be operated independently or simultan-
eously.
Major components of the burner module
are a fluidized bed for burning the fuel and a
dry product collection system for filtering and
retaining particulates from the burner off-gas.
The dry product collection system is bypassed
when direct introduction of the burner
product into the leaching equipment is
desired.
Several burner-designs were used in the
course of pilot-plant development. The final
burner design, the concentric fluidized-bed
burner, is shown schematically in Figure 5. A
4-in. diameter, 4-1/2-ft long first-stage
burner is located concentrically inside a 6-in.
diameter vessel which extends 8-1/2 feet above
the top of the first stage. Fluidized beds are
contained in the 4-in. diameter first stage, the
surrounding annular space, and in the 6-in.
diameter section above the first stage. The
wall of the upper 2 feet of the first-stage
burner is slotted to allow particle mixing
between the annular and inner beds.
The concentric fluidized-bed design was
proposed on the basis of potential increased
heat transfer rates from the wall of the inner
first-stage burner. The outer wall of the first-
stage burner in an earlier two-stage fluidized-
bed burner was cooled by forced-air convec-
tion; this proved to be adequate under normal
conditions but inadequate in the event of a
temperature excursion. Substitution of the
annular bed for the forced convection system
was expected to increase the heat transfer
coefficients at the first-stage wall by an order
of magnitude. The annular fluidized bed
provides the added advantage of secondary
containment should melt-through of the inner
vessel occur.
Tests conducted with a 6-in. diameter glass
column containing the actual 4-in. diameter
inner bed further showed that increased heat
transfer could also be expected from particle
IV-3-4
-------
mixing between the inner and annular beds.
The intensity of slugging in the inner bed was
greatly reduced in the upper 2 feet due to the
transfer of gas and particles between the inner
and annular beds.
The anticipated improvement in heat
transfer using the concentric fluidized-bed
burner has been realized. The concentric-bed
design has proved superior to an original two-
stage burner with respect to heat transfer and
temperature control. No temperature
excursions have occurred in any of the
experiments. Temperature differentials within
the first stage have ranged from a normal
spread of 25 to 45° F to a maximum spread of
75°F.
Combustion efficiencies equal to or greater
than the required 95 weight percent have been
obtained in the two-stage concentric fluidized
bed over the following ranges of operating
variables:
1. Nominal bed temperature—1400 to 1500°F.
2. Fluidizing-gas composition—80 to 100
percent oxygen to both stages.
3. Mean fluidizing velocities (average of inner
and annular fluidizing velocities) — 1.05
to 1.50 ft/sec.
4. Fuel • charging rates—up to 33 kg
graphite/hr-ft2.
APPLICATION OF FLUID-BED
COMBUSTION TO COMBINED-CYCLE
POWER PRODUCTION FROM WOOD
WASTES AND MUNICIPAL REFUSE
Problem Definition
At present, approximately 250 x 10 _ tons
(190 x 10'6 tons of which are collected) of
residential, commercial, and institutional
wastes are produced in the United States each
year. The per capita generation of such wastes
is rapidly increasing; combined with an
expanding population, the magnitude of the
problem may double by the turn of the
century. The disposal of refuse is -becoming a
critical problem for communities, especially
those in the heavily populated parts of the
country.
The lumber industry, particularly the
small-to-medium size mills, has a similar
waste disposal problem. Approximately 5 x
106 tons of wood wastes are produced in the
United States per year. The volume of waste
makes landfill disposal impractical; the wastes
are usually burned in inefficient teepee
burners. While economical, such systems
cause localized air pollution by emitting
smoke, particulate matter, and partially-
oxidized chemicals. Failure to meet the
stringent air pollution standards (now being
introduced) will prevent future operation of
teepee burners.
Recycle of Solid Waste By Energy Production
Through regional planning and cooper-
ation, wood waste materials and municipal
refuse would be transported to a central
location and burned in a highly efficient
pressurized fluid-bed burner. Heat released in
the bed would be used to generate steam
within an in-bed heat exchanger; both the
steam and high temperature off-gases (1400 to
1500°F) would generate electrical power using
steam and gas turbine cycles, respectively.
Removal of heat by generating steam in an in-
bed heat exchanger would also minimize the
amount of excess air normally required to
control the bed temperature; thus, the volume
of off-gas requiring cleanup would be reduced.
Technology required for plant-scale process
demonstration is either already established or
in the latter stages of development.
Advantages of the proposed system are:
1. Significant reduction in present air and
solid waste pollution (to meet present and
future standards).
2. Decreased land requirements for disposal.
3. Conservation of natural resources (fossil
and nuclear fuels) by recycling wastes to
produce electrical power.
4. Decreased cost of waste disposal.
IV-3-5
-------
While the concept is one in which both
wood and municipal wastes are available,
successful development and plant-scale
demonstration of the proposed concept could
lead to use in areas where either type of fuel
predominates. The concept should find wide
application throughout the Pacific Northwest,
the North, and the Southeast.
A schematic flowsheet of the conceptual
plant is shown in Figure 6. Although the plant
would be designed to process wood waste and
municipal refuse as the principal fuels, the
basic concept is compatible with other waste
(e.g., industrial wastes, sewage sludge, etc.) as
fuel. In the future, the plant could be modified
to accept feed in the form of low sulfur coal as
fuel in the event that advancing technology
results in better use of the waste materials.
After the required preparation (e.g.,
screening, shredding, and storage), the waste
is partially dried and charged to a fluid-bed
burner operating at 150 psia and in the
temperature range of 1400 to 1500°F.
Combustion occurs in an inert bed of sand;
the efficient solids and gas contact in the bed
results in rapid and complete combustion. Ash
and noncombustibles are continually with-
drawn from the burner; some particulate is
carried overhead in the off-gas.
Steam is generated from condensate
passing through the tubes of an in-bed heat
exchanger. The superheated high pressure
steam then flows to a steam turbine-generator
to produce electrical power. Flue gases from
the fluid bed are cleaned of particulate matter
using a combination of cyclones and high
efficiency filters (ceramic or sintered metal).
The clean flue gas then flows to a gas turbine
to generate additional electric power.
The concept of combining gas and steam
turbine cycles in a system for generation of
electric power is not new, and the high thermal
efficiencies possible in such cycles are being
demonstrated in utility power stations. The
San Angelo Station of West Texas Utilities,
for example, has achieved 41 percent thermal
IV-3-6
efficiency while burning natural gas in a gas
turbine exhausting to a conventional boiler-
steam turbine system.
The concept of generating steam in tubes
immersed in fluidized beds of solids also is not
new. Work has been under way for several
years to develop this system, both in the
United States and abroad. The British, in
1969, speculated that the advantages of
pressurized fluid-bed boilers deserved further
study including a mixed cycle incorporating a
gas turbine.4 In the United States, pilot-scale
work has been under way for some four years
to develop a system for pressurized fluid-bed
combustion of solid wastes using hot exit gases
to produce electric power in a gas turbine
generator.5
The concept proposed herein includes
elements from all of the aforementioned work;
however, the cycle proposed is unique and has
distinct advantages over other proposed or
existing processes for processing wood waste
and municipal refuse. The proposed cycle is
shown in Figure 7; wood wastes would be
dried to less than 10 percent moisture. This is
important to the overall thermal efficiency
because low-level turbine exhaust heat, much
of which is wasted in other cycles, would be
used to dry the wood wastes which may
contain up to 50 percent moisture.
Thermodynamically, the cycle is very
attractive, particularly for the combustion of
waste materials with high water content.
Performance data for one set of conditions,
not necessarily optimum, are summarized in
Table 1.
The plant can be divided into four major
sections: feed preparation, fluid-bed burning,
off-gas cleanup, and power generation
facilities. In feed preparation, the removal
efficiency of noncombustibles from the
municipal refuse has a significant impact on
the operation of the fluid bed. Operation of
the fluid-bed burner (including feed introduc-
tion, ash removal, and in-bed heat transfer)
and off-gas cleanup are critical to the
successful operation of the proposed plant.
-------
Table 1. PERFORMANCE DATA FOR
PROPOSED CYCLE
Basic parameters
Waste moisture content
Gross heating value, dry
Air pressure to fluid bed
Gas pressure to gas turbine
Air temperature to compressor
Fluid bed and exit gas temperature
Excess air
Turbine exhaust pressure
Turbine exhaust temperature
Steam pressure
Steam temperature
Feed water temperature
Condensing pressure
Fluid-bed dryer and exit
gas temperature
Calculated performance/1 00 Ib dry
fuel
kWhr, steam turbo-generator
kWhr, gas turbine turbo-generator
Energy, kWhr/100 Ib dry fuel
Thermal efficiency, %
50%
8075 Btu/lb
150psia
145psia
60°F
1440°F
15%
17 psia
720 °F
1500 psia
1000°F
415°F
1.5in. Hg
150°F
77.0
11.5
88.5
37.5
The power generating facilities will be conven-
tional and do not require detailed discussion.
A process flowsheet of the demonstration
plant is shown in Figure 8.
Feed Preparation and Storage
Feed to the fluid-bed burner consists of
wood waste (e.g., sawdust, chips, and
shredded material) and municipal refuse.
Wood wastes are relatively homogeneous and
require only sizing and drying before feeding
to the bed. The heterogeneity of municipal
refuse requires separation of glass and metals
and sizing before being introduced to the
tluidized bed. Based on a minimum amount of
engineering development, the A-E would
select a feed preparation scheme for
facilitating materials handling, reducing
environmental pollution, and providing safe
storage.
Wood Waste Feed Preparation and
Storage—Wood waste would be properly sized
at the mill site for feed to the burner. Raw feed
which is sufficiently dry (<10 percent
moisture) would be transported directly to
feed storage. All other feed would be dried
approximately 10 percent moisture before
storage.
Municipal Refuse Feed Preparation and
Storage — With the exception of moisture
content, a typical municipal refuse content
and composition is shown in Table 2. The
municipal refuse consist of glass, dirt, metal,
and various combustible materials. Size
distribution of the refuse varies from large
pieces of material to dust particles; moisture
content is a nominal 25 percent by weight. For
rapid combustion, good quality fluidization,
and satisfactory materials handling, refuse
must be sized to less than 1-in. pieces. The
water content is usually lowered to less than 10
percent during normal feed preparation (e.g.,
shredding and classification); therefore,
drying of the refuse is not anticipated.
Table 2. TYPICAL MUNICIPAL REFUSE
COMPOSITION
(Yard-free basis)
Material
Paper, > 1/4 inch
Paper, wood, fabric fines
Wood
Fabrics
Plastics
Inerts (glass and metallics)
Dust
Heating value, Btu/lb (dry basis)
Wt%
48.9
11.7
10.2
1.0
2.0
14.0
11.7
6000
A conceptual flowsheet for feed prepara-
tion of the municipal refuse is shown in Figure
9. The raw refuse is dumped into a feed-
conveyor hopper directly from a truck; no raw
feed storage is provided. After charging to the
shredder, the material is transported to a
classifier where the large and/or dense parti-
cles are separated in the underflow. As
dictated by economics, ferrous metals may be
separated from the underflow for recycle.
Feed preparation equipment will be over-
sized to allow for routine maintenance without
limiting the plant capacity and would
normally be operated only one shift per day. If
economics require, refuse feed preparation
IV-3-7
-------
could be done at the source; this is especially
true in the case of a large city such as
Spokane.
Fluid-Bed Combustion
The fluid-bed burner (Figure 10) is the
heart of the proposed system. The maximum
bed temperature is limited by the gas turbine
blade materials and the minimum fusion
temperature of the noncombustibles and ash
present in the bed. Fluid-bed burners have a
high heat release rate per unit volume of bed;
because of the variability in the heat content of
refuse, control of bed temperature could be
difficult when charging municipal refuse
alone. However, temperature control problems
in the proposed concept should be minimal,
since the wood waste (relatively homogeneous
and of constant heat value) will comprise one-
half to two-thirds of the feed to the burner.
Condensate within tubes of an in-bed heat
exchanger is converted to steam by transfer of
heat from the fluid bed. Depending on
economics and technical considerations, bare
or finned tubes will be used. Based on a
practical carbon steel tube bundle configur-
ation and overall heat transfer coefficients in
the range of 300 to 600 Btu/hr-ft2-°F, heat
' transfer rates in the range of 4 x 105 to 8 x 105
Btu/hr-ft3 of bed are possible. For a 125
ton/hr plant, approximately 1 x 106 Ib/hr of
steam (1200 to 1500 psia and 900 to 1000°F)
would be generated. Based on a nominal
steam cycle efficiency of 35 percent, approx-
imately 100 MW would be produced by the
steam turbine.
Flue gases from the fluid-bed burner pass
through a high efficiency off-gas cleanup
system (series of cyclones followed by ceramic
or sintered metal filters) for removal of essen-
tially all particles > 5 Mm in diameter. The
cleaned off-gas is then passed to a gas turbine
where, based on a 125 ton/hr plant and a
turbine cycle efficiency of 20 percent, 14 MW
are produced.
Though most of the ferrous and nonferrous
metals will be removed by magnetic separation
and air classification, some of these
components will remain in the feed to the
burner. In addition, glass, dirt, ceramic
material, and possible agglomerates of
material (e.g., plastics, etc.) must be removed
from the burner. If the material is allowed to
accumulate on the air distributor, fluidization
quality would deteriorate as a result of gas
channeling. A reliable system for removal of
ash agglomerate and noncombustibles is
required.
Off-Gas Cleanup System
Removal of entrained particulates and
corrosive gases is a major problem in the
design of conventional incinerator off-gas
cleanup systems; expansion of the off-gas in a
turbine requires even a higher degree of off-
gas cleanup. Particulates, if not removed, will
erode and foul turbine blades and pollute the
environment. Corrosive gases will also corrode
the off-gas cleanup system and turbine blades.
Fewer corrosive gases are generated during
processing of wood waste; therefore, the wood
waste could result in off-gas concentrations of
corrosive gases of acceptable levels.
Although maximum particulate loading speci-
fications for turbine inlet gases vary, 2 x 10"3
gr/ft3 of off-gas, with no more than 10 percent
of the particles greater than 10 urn in size, is a
typical requirement. Because of the control of
bed temperature by transfer of heat to an in-
bed steam generator, the excess air flow and
hence volume of gas requiring cleanup is
minimized using the proposed concept.
Conventional Particulate Removal —
Particulates are normally removed from off-
gas streams by combinations of cyclone
separators, scrubbers, and electrostatic
precipitators; each type of cleanup device has
distinct disadvantages. Cyclones are normally
effective in removing particulates at high gas
rates as long as the particle size is larger than
20 nm. Scrubbers require low operating
temperatures, and electrostatic precipitators
are often ineffective. Sand or granular filters,
though sometimes employed, are bulky and
regeneration is difficult.
IV-3-8
-------
Combination Multi-Stage Cyclone and
"Candle" Filters for Participate Removal —
The concept of filtering off-gases at high tem-
peratures is not new; sintered metal filters are
commonly employed up to 1500°F. Such
filters are available for removing submicron
particles at pressure differentials less than 1-
in. water. Ceramic "candle" filters, used in
conjunction with multi-stage cyclones, appear
to be most attractive for removing particulates
to the levels recommended by turbine manu-
facturers. Well designed 3-stage cyclones
could almost satisfy the requirements alone,
and candle filters could further reduce
particulate loading to levels below process
requirements. The cyclones would be
constructed of a material resistant to chloride
corrosion at temperatures as high as 1500°F.
In the proposed concept, off-gases at 145
psia and 1400 to 1500°F would leave the
burner and pass through multi-stage cyclones
followed by candle filters for final particulate
removal, as shown in Figure 11. The multi-
stage cyclones would remove essentially all
particulates greater than 20 pirn. Off-gas
passing to the turbine would contain a max-
imum loading of 1 x 10"3 gr/ft3, and
essentially all particulates greater than 5
microns would be removed. The multi-stage
cyclone and candle filter systems are commer-
cially available.
CONCLUSIONS
Disposal of municipal refuse and wood
waste while minimizing harmful effects on the
environment is an existing problem. The per
capita increase of refuse generation rates
coupled with the expected population increase
will almost double the annual refuse generated
by 1980. Wood waste generation rates are also
expected to increase, though not as rapidly as
municipal refuse.
Existing methods for municipal refuse
disposal (landfill and conventional inciner-
ation) frequently result in air and water pollu-
tion and unsightly facilities. Disposal of wood
waste by incineration in teepee burners is
unacceptable from the standpoint of satisfying
air pollution standards. Development of more
advanced systems for disposal of municipal
refuse is based largely on the concept of pro-
duct recycle. The recycled products have ques-
tionable market value when compared with
virgin materials.
Existing and near-term fluid-bed
technology can result in a fluid-bed process for
power generation having a thermal efficiency
greater than 35 percent while reducing the off-
gas mass flow rates per MW-hour of electrical
energy by a factor of eight when compared
with a conventional gas turbine cycle. In such
a process, steam would be generated within an
in-bed heat exchanger and used as the
working fluid in a steam-turbine cycle. The
latest technology in finned-tube heat transfer
within fluid beds would be used. The burner
off-gases would be cleaned of particulate by
passage through a high efficiency cleanup
system consisting of staged cyclones and
sintered metal (stainless steel and Hastelloy C)
or ceramic "candle" filters. Particulate
removal at the temperatures (1400 to 1500°F)
proposed for refuse and wood waste combus-
tion has been tested in development and
operations facilities in the atomic energy field.
Use of existing technology and commercially
available equipment where applicable will
significantly reduce the development costs and
time required to place a full-scale waste
disposal-power generation facility on stream.
Power production costs are estimated to be in
the range of 5 to 6 mills/kWhr.
BIBLIOGRAPHY
1. Bendixsen, C.L. et al. The Third Processing
Campaign in the Waste Calcining Facility.
U.S. Atomic Energy Commission, Oak
ridge, Tenn. Publication Number IN-1474.
May 1971.
2. Kilian, D.C. et al. Description of the Pilot
Plant for the Headend Reprocessing of
Unirradiated Rover Fuels. U.S. Atomic
Energy Commission, Oak Ridge, Tenn.
Publication Number IN-1181. May 1968.
IV-3-9
-------
3. Cox, A.R., et al. Operation of San Angelo Combustion. The Engineer, July 24, 1969.
Power Station Combined Steam and Gas c ,, . . „ T, . Ann /-..*•
„,.„,._ ,. . . 5. Combustion Power Umt-400. Combustion
Turbine Cycles. In: Proceedings American „ _ n , A1, .-, ,.„ „ , ~
„ _ J_ A7 , VVTV AM Power Co., Palo Alto, Calif. Prepared for
Power Conference, Volume XXIX, p. 401- ,. „ ~ c ,., l,T . .» ,
41- ' v the Bureau of Solid Waste Management,
Environmental Protection Agency, Wash-
ington, D.C. under Contract Number Ph
4. Big Economics from Pressurized Fluid-Bed 86-67-259.
IV-3-10
-------
WASTE SOLUTION
_/ RECYCLE
ATOMIZING
AIR
ATOMIZING
OXYGEN
KEROSENE
SCRUBBER SEPARATOR.
VENTURISCRUBBER
ICPP
STACK
BLOWER
• SOLIDS TRANSPORT AIR
SOLIDS STORAGE BINS
BLOWER
Figure 1. Waste calcinating facility.
CO
-------
CIRCULATING
HEAT TRANSFER
MEDIUM
I
HEAT
SINK
BURNING
FUEL
METAL
WALL
LIQUID
METAL
METAL
WALL
FLUIDIZED
PARTICLES
ENDOTHERM1C
REACTION
IN-BED HEAT EXCHANGER
BURNING
FUEL
FLUIDIZED
PARTICLES
IN-BED COMBUSTION
ENDOTHERMIC
REACTION
Figure 2. Comparison of heat flow paths.
BLOWBACK
AIR
TO VESSEL
OFF GAS'
GLOVE BOX
BAG
Z30UT
PORT
CANNER
Figure 3. Product denitration facility.
IV-3-12
-------
TO EXHAUST STACK
CHARGING GLOVE
BOX
ALUMINA
CHARGING
POT
COOLING
AIR
FLUIDIZING
GAS
A TO PLANT OFF-GAS
SYSTEM
TO DISSOLUTION
STEPS
GRAPHITE
BURNER
DRY SOLIDS
COLLECTION
VESSEL
Figure 4. Graphite fuels pilot plant-combustion process flowsheet.
IV-3-13
-------
P TW
COOLING AIR
INLET
GAS INLET TO
INNER BED
OFF GAS
OUTLET
ALUMINA CHARGING PORT
FUEL CHARGING TUBE
AUXILIARY GAS INLET
6-in. SCHEDULE 40 PIPE
TYPE 316 STAINLESS STEEL
BED SAMPLING LINE
COOLING AIR OUTLET
MlO
HEATING ELEMENTS
PNEUMATIC
OPERATOR
BALL VALVE
SAMPLING LINE
Hn. SCHEDULE 40 PIPE
TYPE 304-L STAINLESS STEEL
STAINLESS STEEL SHROUD
/-HEATING ELEMENT
FLUIDIZING GAS
DISTRIBUTOR
PLATE
BED SUPPORT PLATE AND
GAS DISTRIBUTION
GAS INLET TO
ANNULAR BED
QUO
DEFLECTION BAFFLES
FUEL CHARGING TUBE
1-M-in. SCHEDULE 40 PIPE
HASTELLOY ALLOY-C
PERFORATED PLATE BAFFLES
FUEL CHARGING TUBE
3/8-in.x3-in. PERFORATIONS
STAGGERED ON %-ln. CENTER
LEGEND
P-PRESSURE TOP
THNTERIOR TEMPERATURE
THERMOWELL
TW-VESSEL WALL TEMPER-
ATURE THERMOWELL
INNER BED DRAIN
ANNULAR BED DRAIN
Figure 5. Concentric bed burner.
IV-3-14
-------
INERT MATERIALS
(ASH, GLASS, AND
ROCKS) TO A
LANDFILL
RAW
MUNICIPAL
REFUSE
WOOD
WASTES
1
1
1
MUNICIPAL
REFUSE
PREPARATION
RECYCLE OF !
GLASS I
L
r
i
WOOD
STORAGE
PLANT BOUNDARY
FLUIDIZED
BED
COMBUSTOR
ELECTRICAL
POWER
(STEAMS
ELECTRICAL
POWER
GAS TURBINE
SALABLE ELECTRICAL POWER
Figure 6. Schematic of waste disposal-power generation facility.
IV-3-15
-------
CO
h-1
O5
TO ATMOSPHERE
SCRUBBER
SUPERHEATED
STEAM
PREPARED
MUNICIPAL
REFUSE
L
FEED WATER
PUMP
WET WOOD
WASTE
HOT TURBINE
EXHAUST GAS
Figure 7. Proposed cycle for producing electric power from wood waste and municipal refuse.
-------
FLUIDIZED-BED
INCINERATOR
WOOD FEED
STORAGE
MUNICIPAL
WASTE
FLUIDIZED-BED
INCINERATOR
GENERATOR I COMPRESSOR
GAS TURBINE
ASH
<3
CO
Figure 8- Flowsheet of proposed waste recycle-power plant.
-------
RAW MUNICIPAL
REFUSE
PREPARED FEED
STORAGE
TO STACK
GAS
TO FLUID-BED BURNER
NONCOMBUST1BLES
Figure 9. Municipal refuse feed preparation facility.
IV-3-18
-------
SUPERHEATED
STEAM
CONDENSATE-
•-OFF-GAS
PEPARED
FEED
PNEUMATIC
CONVEYOR
li'Vi-'''i£ ''j,j' ^' i''!!z,;.'fjiji''2,
/A v/a Vfa ¥//t ra r/a VIA v//» m
-CYCLONES
SUPER HEATER
SECTION
t* -*STEAM GENERATOR
EVAPORATIVE
SECTION
GRID SUPPORT AND
AIR DISTRIBUTOR
HIGH-PRESSURE COMBUSTION
AND FLUIDIZING AIR
Figure 10. Fluid-bed burner.
IV-3-19
-------
BANK OF
CANDLE FILTERS
MULTI-STAGE
CYCLONES
FEED
Figure 11. Off-gas cleanup system for proposed plant.
IV-3-20
-------
4. POWER GENERATION USING THE SHELL
GASIFICATION PROCESS
A. N. DRAVID, C. J. KUHRE, AND J. A. SYKES, JR.
Shell Development Company
ABSTRACT
Growing concern about sulfur and nitrogen oxide emissions has given rise to a search for means
of converting conventional fuels into clean, non-polluting fuels for electric power generation.
Through its ability for converting liquid fuels into partially oxidized gaseous fuels and recovering
the heat of partial oxidation in the form of high pressure steam, the Shell Gasification Process
(SGP), aided by the Shell Sulfinol or ADIP Process, offers an attractive means of converting sulfur-
laden heavy hydrocarbon feedstocks of high metals content into non-polluting fuel gas and
saleable elemental sulfur.
Conceptual design and economics of an SGP-based power plant utilizing the Combined Gas
and Steam (COGAS) cycle are presented. Besides offering the simplicity, flexibility, and reliability
associated with the SGP, such a power plant can generate electric power at unit costs competitive
with those of future conventional power plants.
INTRODUCTION
Thp Shell Gasification Process (SGP)* is a
process for the partial combustion of hydro-
carbons, and is particularly suitable for the
partial combustion of heavy, sulfur containing
residual fuels and heavy crude oils to produce
a mixture of hydrogen and carbon monoxide.
From this mixture the hydrogen sulfide pro-
duced during partial oxidation can be readily
removed. A non-polluting fuel gas is thus pro-
duced which can be used for power
generation. This type of fuel should be of
particular interest for power generation
because of the following factors:
1. Natural gas, for many years a sulfur-free
fuel, has slid into a declining reserve posi-
tion in the face of an increasing demand at
"Licensed by Shell Development Company, Houston, Texas
77001 and Shell Internationale Research, Maatschappij, N.V.,
The Hague.
present regulated prices.
2. The cost and difficulty of desulfurizing
heavy fuel oils, particularly those with
high metals content, is very high.
3. Eastern and mid western coals have high
sulfur contents which to an increasing
extent make them unsuitable for use for
generation of power in conventional steam
power plants.
4. There is a relatively long lead time for the
development of low sulfur western coals,
and also high transportation cost asso-
ciated with the use of these coals.
5. There is a long lead time required for the
installation of nuclear power plants.
As an alternative method of power gen-
eration, the Shell Gasification Process, with a
IV-4-1
-------
moderate investment and a high thermal
efficiency (>85 percent), converts fuels with
high levels of sulfur, nitrogen and/or metals
into attractive power generating fuels for use
in the COGAS cycle (Combined Gas and
Steam cycle).
An SGP-based power plant consists of a
Shell Gasification Process unit, for converting
residual fuels or low-value crudes into low-Btu
fuel gas and recovered steam, and a gas tur-
bine-steam turbine unit for converting these
products into electrical power. An SGP-based
power plant differs from a conventional
thermal power plant: in a conventional
thermal power plant raw fuel is burned
directly, while in an SGP-based plant the
power plant fuel is the product of partial oxi-
dation of the raw fuel.
The power generation unit recommended
for the SGP-based power plant uses the
COGAS cycle. The COGAS cycleJ-2'3 is ther-
modynamically superior to either the steam
cycle or the gas cycle. It is particularly suited
for an SGP-based power station, since in the
SGP the net exothermic heat of partial oxi-
dation is recovered as high pressure steam
which can be integrated with the steam section
of the COGAS cycle.
The following technical and economic case
study shows that an SGP-based power station
is not only a feasible means of generating
power without contributing to atmospheric
pollution, but it is also economically competi-
tive with conventional power plants of the
future.
IV-4-2
-------
CHEMISTRY OF PARTIAL OXIDATION
Partial oxidation describes the net effect of
a number of component reactions that occur
in a flame, supplied with less than stoichio-
metric oxygen. This net effect can be
approximated:
and is actually a combination of several reac-
tions that occur within the reactor.
Heating-op and cracking phase
In the fuel injection region of the reactor,
hydrocarbons leaving the atomizer at about
preheat temperature are intimately mixed
with air. Prior to combustion they are heated
and vaporized by back radiation from the
flame and the reactor walls. Some cracking of
the hydrocarbons to carbon, methane, and
hydrocarbon radicals may take place during
this brief interval.
Reaction phase
As soon as the ignition temperature is
reached, part of the hydrocarbons will react
with oxygen according to the highly
exothermic reaction:
CnHm+(n + —)O2 -*• n CO +—H O. (2)
As the equilibrium is far to the right,
practically all the available oxygen is con-
sumed in this phase. The remaining hydrocar-
bons which have not been oxidized react with
steam and the combustion products from
reaction (2) according to the endothermic
reactions:
CnHm + nC02
and
2nCO+f-H2
CnHm + n H20- n CO + (21 + n) H2 . (4)
In order to prevent excessive local tempera-
tures, it is essential that all reactants of equa-
tions (2) to (4) are intimately mixed so that the
endothermic reactions tend to balance the
exothermic reactions. In this way the complex
of reactions is brought to a thermal
equilibrium resulting in a measured tempera-
ture of about 2350 to 2550°F.
Soaking phase
Soaking takes place in the rest of the
reactor where the gas is at a high temperature.
The gas composition changes only slightly due
to secondary reactions of methane and carbon
and the water gas shift reaction.
Methane produced by cracking will de-
crease according to:
CH4 + H2O ^ CO+3H2
(5)
and
CH4 + CO2
2CO + 2H2. (6)
As the reaction rate is relatively low, the
methane content is higher than would be
expected from equilibrium.
During the soaking phase a portion of the
carbon also disappears according to the reac-
tions:
C + CO
2CO
and
C + HO - CO + H2.
(7)
(8)
However, some carbon is always present in the
product gas from the reactor in a quantity
equivalent to about 3 weight percent of the oil
feed.
The composition of the fuel gas is deter-
mined by the water-gas shift equilibrium
which appears to freeze after the gas enters the
waste heat boiler at an equilibrium tempera-
ture about 2200 to 2400°F.
CO + HO ^ CO2 + H2
(9)
DESCRIPTION OF SHELL GASIFICA-
TION PROCESS
A simplified SGP flow sheet is given in
Figure 1. The hydrocarbon charge and the
oxidant are preheated and fed to the reactor.
IV-4-3
-------
The hot reactor-effluent gas containing about
3 percent of the feed as soot is passed to a
waste heat boiler, producing high pressure
saturated steam. High heat transfer rates are
achieved, with the result that the temperature
of the gas leaving the waste heat boiler closely
approaches that of the steam produced in the
boiler. The design and construction of the
waste heat boiler are such that the surface
remains clean for an indefinite period (without
using any external cleaning devices); it may be
noted that the waste heat boiler of the Shell
prototype unit has been in operation since
1956 and never has been cleaned on the gas
side. The waste heat boiler can be designed for
steam pressures up to about 1500 psig.
The crude gas leaving the waste heat boiler
at temperatures around 350°F is then passed
to the carbon removal system, consisting of a
bulk removal of the carbon by a special
method of contacting the gas with water, and a
final water wash. The product gas is virtually
free of carbon (< 5 ppm).
The carbon produced in the gasification is
recovered as a soot-in-water slurry (carbon
content 1 to 2 weight percent). In most cases, it
will not be possible to dispose of this carbon
slurry as such. Therefore, a special technique
has been developed for removing the carbon
from the slurry, resulting in carbon-free water
for re-use. Depending upon the metals content
of the feedstock and the economics and
maintenance policy of the process operator, up
to 100 percent of the soot can be recycled to
extinction with the fresh feed.
Sulfur in the feedstock is converted
primarily to H2S and traces of COS. The
carbon-free product gas is treated in a Shell
Sulfinol or ADIP process unit where the sulfur
compounds and most of the CC>2 are
absorbed. The desulfurized gas typically
contains less than 5 ppm of sulfur. The acid
gas effluent from the Sulfinol unit is fed to a
Claus process unit which recovers elemental,
salable sulfur.
Depending on the desired LHV (Lower
Heating Value) in the product gas, either
IV-4-4
oxygen or air (enriched or unenriched) may be
used as the oxidant. Nitrogen present in the
air acts as a moderator for temperature
control in the reactor. When either oxygen or
air enriched with oxygen is used as the
oxidant, a certain quantity of steam must be
injected into the reactor for temperature
moderation. Air oxidation produces a low
heating value (120 Btu/scf) fuel gas due to the
presence of nitrogen, while oxygen feed
produces a medium heating value gas (300
Btu/scf). Typical product gas compositions for
air and oxygen gasification are shown in Table
1.
Tabte 1. TYPICAL PRODUCT GAS COMPOSITION
Hydrogen
Carbon monoxide
Methane
Nitrogen
Argon
Sulfur
Total
% vol, dry basis
Oxygen
oxidation
48.0
51.0
0.6
0.2
0.2
5 ppm
100.0
Air
oxidation
12.0
21.0
0.6
66.0
0.4
5 ppm
100.0
COGAS CYCLE THERMODYNAMICS
Although the idea of combining a gas
turbine and steam turbine is old, its applica-
tion to power generation has been studied only
lately. Wood3 has presented an excellent
summary of the development of the COGAS
cycle. In a COGAS cycle (Figure 2), air is
compressed and heated by burning fuel in it.
The hot gases are then expanded in a gas
turbine coupled to the air compressor and a
generator. The gas turbine exhaust, still at a
high temperature, is used to raise and
superheat high pressure steam; it is also used
as a heat source for deaeration and boiler feed
water preheating. The steam generated by the
gas turbine exhaust is expanded in a steam
turbine to produce additional electric power.
Heat rejection occurs in the stack exhaust and
in the condenser of the steam cycle. It is well
known that the greater the difference between
-------
the heat source and heat sink temperatures of
any heat engine, the higher its thermodynamic
efficiency. In the COGAS cycle, the heat sink
of the gas cycle becomes the heat source of the
steam cycle, increasing the overall spread
between source and sink temperatures for the
combined cycle. As a result, the COGAS cycle
has a higher thermodynamic efficiency than
either the simple gas cycle or the steam cycle.
In the case presented in this paper, a COGAS
cycle efficiency (based on the net useful energy
input to the power plant) of 44 percent was ob-
tained. Simple gas cycle and steam cycle
efficiencies are of the order of 25 percent and
37 percent, respectively.
In application of the Shell Gasification
Process to the COGAS cycle, the addition of
the steam generated by the heat of partial
oxidation to the steam generated in the
COGAS cycle largely compensates for the loss
of heating value of the oil caused by gasifica-
tion.
POWER PLANT FLOW SCHEME
A flow diagram of the power plant using
desulfurized fuel gas as fuel is shown in Figure
3. Air is compressed to 14 atm and split into
two parts. One part is cooled by heat exchange
with 100° F product gas from the SGP section
and compressed again in a booster compressor
to 18 atm before entering the SGP unit as the
oxidant. The other part of the compressed air
is combusted to a temperature of about
2200°F in the gas turbine combustor by
burning the sulfur-free fuel gas supplied by
the SGP/Sulfinol units, and expanded to 1.5
atm (absolute) pressure in the gas turbine
which is coupled to an electric power
generator. The gas turbine exhaust is cooled to
about 350 °F in a waste heat
boiler/superheater and boiler feed water
deaerator before being vented to atmosphere.
Steam (1250 psig) generated in the SGP waste
heat boiler is combined with the steam
generated in the power plant waste heat boiler,
and the combined steam is superheated to
1000°F in the superheater section of the latter.
The superheated steam is expanded typically
to about 4-in. Hg vacuum in a steam turbine
coupled to a second electric power generator.
From the point of view of startup and control,
it is advisable to use separate generators for
the gas and steam turbines. Approximately 56
percent of the total power generation is
contributed by the gas turbine.
SGP-BASED POWER PLANT — CASE
STUDY
Using the foregoing technology, an
economic study has been made of an SGP-
based power plant of 200-MW nominal
generating capacity. Air was used as the
oxidant in the partial oxidation step. The
composition of the typical heavy residue used
as feedstock is stated in Table 2. The material
balance of various gas streams in the power
plant appears in the inset of Figure 3.
Table 2. TYPICAL RESIDUAL FEED PROPERTIES
Gravity, °API, 60°F 12.0
Specific gravity, 60/60 0.986
Composition, wt %
Carbon
Hydrogen
Sulfur
Nitrogen
Oxygen
Ash
Total
Viscosity, centistokes
470°F
212°F
100°F
Ash analysis:
Nickel
Vanadium
Sodium
Iron
Others
Total
86.00
10.73
2.65
0.30
0.30
0.02
100.00
4
55
800
ppm in feed
30
100
1
4
65
200
% of ash
15.0
50.0
0.5
2.0
32.5
100.0
Turbine manufacturers have indicated to
us that industrial gas turbines are currently
designed for a compression ratio of 12 and
IV-4-5
-------
turbine inlet temperature of 1800°F. However,
it is predicted that by 1975, these conditions
are likely to be upgraded to 14 and 2200 °F,
respectively. In this evaluation we have
assumed 1975 technology. Other simplifying
assumptions made were:
1. The gas turbine can accommodate an
increase of approximately 25 percent
(mole) in the gas flow across the
combustor as low-Btu gas is injected into
the combustor. (This has been shown to be
feasible.)
2. Polytropic efficiencies of the various
components are : compressor 90 percent,
gas turbine 90 percent, steam turbine 75
percent2.
3. Approximately 2 percent of boiler feed
water is evaporated in the deaerator, and 5
percent is rejected in the boiler feed water
blowdown. Thus, fresh boiler feed water
requirement is 7 percent of the total steam
generation.
4. Friction and generator losses, which are
normally small, have been neglected.
RESULTS AND DISCUSSION
A power flow diagram for this case (Figures
4A and 4B) shows the factors contributing to
an overall station efficiency of 38 percent.
Figure 4A refers to the SGP unit and 4B shows
the power flow through the power generation
unit. The downstream end of Figure 4A thus
matches with the upstream end of Figure 4B.
About 87 percent of the energy input to the
SGP unit (LHV basis) is available to the power
plant as input energy.
Table 3 is a summary of the capital and
manufacturing costs. The major uncertainties
in the unit power cost in this estimate, the
power plant capital cost and the oil price, are
shown parametrically (Figure 5) based on 1972
U.S. East Coast costs.
Estimated unit power costs are about 1 to 2
mills/kWhr higher than current rates,
depending on the location. However, the latter
costs do not include the cost of stack gas
scrubbing or other alternatives of treating
sulfur and NOX emissions. With the growing
shortage of natural gas and restrictions on
sulfur, NOX, and particulate emissions, the
power cost is certain to rise rapidly in the next
few years. At such time, SGP-based power
plants will offer an attractive alternative to
conventional power plants.
Some of the advantages of the Shell
Gasification Process relative to coal
gasification processes are:
1. The SGP unit flexibly accepts a wide
variety of fuels ranging from heavy
residues (e.g., flasher pitch) to natural gas.
Thus, a consistent and continuous quality
of fuel gas can be generated despite
variations in the quality of fuel supply.
2. The SGP-based power station handles
fluids, avoiding the complex solids
handling andash disposal steps involved in
coal gasification units.
3. Both SGP and coal gasification units
require comparable installation times.
The lag in coal availability may lead to a
longer project realization time for coal
gasification-power units.
Relative to other means of clean power
generation, such as the conventional power
plants with stack gas clean-up units, the SGP-
based power plant offers the following
advantages:
1. The SGP unit, followed by a Sulfinol and
Claus unit, produces the most marketable
quality of recovered sulfur.
2. The NOX concentration in SGP-based fuel
gas is low because the C-N bond in the fuel
is broken mostly into CO and N2. With
careful gas turbine combustor design
there is little breaking of the N-N bond, to
produce NOX.
The unit evaluated in this paper is intended
for intermediate to base load applications.
IV-4-6
-------
TableS. POWER GENERATION COST
Power produced (gross), MW'
Power consumed, MW
Net power output, MW
Overall efficiency, %
Capital costs, $ x 106
Fuel processing unit
Power generation unit
Total capital cost, $ x 106 b
Operating cost
Variable costs
Oil @ $x/bbl
Sulfur credit @ $10/ton
Catalysts and chemicals
Cooling and boiler feed water
$x106/yr
2.61x
(0.11)
0.10
0.69
Total 0.68 + 2.61 x
Fixed costs
Operating labor @ $83,500/job
(4 operators)
Maintenance @ 3% of capital
Local overhead @ 100% labor
plus 25% maintenance
Taxes and insurance @ 1 %
of capital
Total
Net operating cost 3
Capital charges @ 14%
0.34
1.45
0.70
0.50
2.99
67 + 2.61 x
6.94
Total power cost 10.61 + 2.61x
18.2
31.4
49.6
Mills/kWhr
1.52x
(0.06)
0.06
0.40
0.40+ 1.52x
0.20
0.85
0.41
0.29
1.75
2.15+ 1.52x
4.06
6.21 + 1.52x
200.0
4.7
195.3
38.0
a Yearly average value. Actual capacity is 11% higher to compensate for
b 90% stream factor.
C1972, U.S. East Coast dollars.
C.W.at1.5
...
Cond. at 7% of circulation @ $0.50/10 3 gal.
Basis: 200 MW Nominal Generating Capacity
90% Stream Factor
IV-4-7
-------
CONCLUSIONS
The Shell Gasification Process in
conjunction with a COGAS power generation
unit offers an attractive combination for
generating electric power from high-sulfur,
heavy residual fuels. At the same time, sulfur
and NOX emissions are almost completely
eliminated, and the sulfur is recovered as a
salable byproduct. The process flexibly
accepts wide and frequent variations in the
feed quality and composition ranging from
natural gas to flasher pitch. Effluents are
minimized. Efficient heat recovery in the
gasification unit combined with the
advantages of the COGAS cycle in the power
generation unit ensures very little loss of
energy despite the additional fuel processing.
High stream factors render this process
suitable for base load applications. In view of
these merits, the SGP-based power station has
promise for use in the electric utility industry.
REFERENCES
1. Business Week, March 11, 1972, p. 44C.
2. Robson, R. L., et al. Technological and
Economic Feasibility of Advanced Power
Cycles and Methods of Producing Non-
polluting Fuels for Utility Power Stations.
United Aircraft Research Laboratories,
East Hartford, Connecticut. Report
Number J-970855-13, p. 258.
3. Wood, B. Combined Cycles: A General
Review of Achievements. Combustion.
April 1972, p. 12.
IV-4-8
-------
STEAM
PREHEATERS
HIGH PRESSURE STEAM
OXYGEN
OR AIR
REACTOR
WASTE
HEAT
BOILER
CARBON
SLURRY
SEPARATOR
FUEL GAS TO
SULFINOL UNIT
FRESH
WATER
CARBON-FREE
CIRCULATION
BOILER
FEED
WATER
BOILER FEED STOCK
WASTE
WATER
Figure 1. SGP for fuel gas manufacture.
IV-4-9
-------
•* FUEL
COOLING WATER
Figure 2. COGAS cycle for power generation.
IV-4-10
-------
414 °F
1250 psig
STEAM
FROM WHB
OF SGP UNIT
BAROMETRIC
CONDENSER
WASTE
HEAT
BOILER
(UNFIRED)
CONDENSATE
PUMP
1250 psig, 212 °F
350 °F
STEAM
TURBINE
DRIVE
-*- AIR TO REACTOR
470 °F, 18 aim
PREHEATED
BFW TO WHB
OF SGP UNIT
COMPONENT
HYDROGEN
CARBON MONOXIDE
CARBON DIOXIDE
METHANE
NITROGEN
OXYGEN
WATER
MOL
WT
2.0
28.0
44.0
16.0
28.0
32.0
18.0
B.P.
SP.
GR.
TOTAL
PRESSURE AND TEMPERATURE,0 F
scfm_x 103
ALL NUMBERS IN Ib-mole hr
CASE
4,158
6,490
397
85
17,230
28,540
116.
<3>
17,485
4,647
22,150
64,412
17,112
81,534
14atm,780°F
179
140
516
j Btu/scf
<4>
6,972
81,642
11,668
4,308
104,560
2,200°F
661
Figure 3 . Power plant flow diagram.
-------
AIR FROM
POWER UNIT
OIL
ENTHALPY
PLUS
HEATING
VALUE
CONDENSATE
OIL
PRE-
HEAT
-^:
M
I
GAS ENTHALPY
10 MW
H
-s;
525 MW
1786x106
Btu/hr
•yr
SGP REACTOR
GAS
ENTHALPY
575 MW
\^^-^
•y*
3 WASTE HEAT
g BOILER
o
fe
UJ
3=
fe
s -^
(
•^
GAS
ENTHALPY
446 MW
^^ ^*
SOOT AND SULFl
RECOVERY
GAS
ENTHALPY
366 MW
^-^ ^^
;=»
IR
?*•
ft 15
\JWAT
ENT
WATE
MW
ER
HALPY
IR ENTH
OIL
PREHEATER
f.
LT>
ftLPY
CD
r-
•"xT
cc
UJ "=*"
g =
i§
o
o
0
l_, ^.J
*
Q.
-------
GAS ENTHALPY
25 m
AIR TO SGP
AIR
mm
FUEL FROM SGP
WATER ENTHALPY
114 m
STEAM FROM SGP
COWBUSTOR
GAS
ENTHALPY
IRQ MW
CONDENSER
GAS ENTHALPY
50 MW
STACK GAS
WASTE HEAT
BOILER
WATER
ENTHALPY
292 MW
T|
I* WATER ENTH
J1 15MW
GAS ENTHALPYj
65 MW n
DEAREATOR
CONDENSATE
COOLING
WATER
LOAD
205 MW
COOLING WATER
Figure 4B, Power flow through power plant.
IV-4-13
-------
20
PARAMETER: TOTAL CAPITAL INVESTMENT, $
5
2 3 4
OIL PRICE, $/bbl
Figure 5. Economics of power generation using SGP.
IV-4-14
-------
5. FLUIDIZED-BED OIL GASIFICATION FOR CLEAN
POWER GENERATION-ATMOSPHERIC AND
PRESSURIZED OPERATION
R. A. NEWBY, D. L. KEAIRNS, E. J. VIDT,
D. H. ARCHER, AND N. E. WEEKS
Westinghouse Research Laboratories
ABSTRACT
This paper evaluates high sulfur residual oil gasification for the purpose of clean power
generation. It also considers both atmospheric pressure operation with conventional boilers and
pressurized operation utilizing combines operation with conventional boilers and pressurized
operation utilizing combined steam and gas turbine cycles.
The outlook and status of atmospheric pressure fluidized-bed oil gasification is reviewed. The
process has been studied by Esso (England) on a 1-MW pilot plant scale and has been shown to be
an effective pollution control device. Preliminary cost estimates for retrofit systems on
conventional plants indicate a potential energy cost reduction of 30 to 50 percent over wet
scrubbing or low sulfur fuel alternatives. A 30 to 100-MW demonstration plant is scheduled.
A pressurized fluid-bed oil gasification process has been proposed, and preliminary assessment
is being carried out in which plant performance, capital costs, and energy costs are examined. The
projected ability of the fluid-bed process to meet both pollution regulations and gas turbine
requirements is based on Esso (England) atmospheric pressure data. Capital and energy costs for
the pressurized fluid-bed combined-cycle (PACE) plant process are compared with a conventional
oil-fired plant with wet scrubbing, pressurized fluid-bed combustion of oil, PACE plant operation
with No. 2 distillate fuel oil, and conventional pressurized oil gasification processes developed by
Shell and Texaco. The fluid-bed process has the potential to reduce energy costs > 20 percent
below a conventional plant with wet scrubbing and > 15 percent below a PACE plant using
distillate fuel oil.
INTRODUCTION
Electric utility demand for residual oil is electric power industry as the availability and
projected to increase by around 250 percent by cost of natural gas and clean fuel oils de-
1980. Low sulfur residual to meet Federal and creases and because the technology for
local regulations will not meet the demand removing sulfur dioxide from stack gases is
without an intense effort on desulfurization. proving to be expensive. A gasification system
must provide a clean gas, reliability, util-
Gas producers for making low Btu gas ization of a wide range of fuel, competitive
from oil are becoming more attractive to the energy cost, and high efficiency.
IV-5-1
-------
Fluidized-bed oil gasification can be
applied to power generation to produce a
clean, low-Btu fuel gas—200 to 500 Btu/scf.
In a fluidized-bed gasifier, oil is added to a
fluidized-bed of limestone or dolomite with
sufficient air—15 to 25 percent of stoich-
iometric—to maintain the bed at ~ 1600°F
and react with the oil to produce a fuel gas.
The limestone or dolomite removes sulfur
from the fuel gas during the gasification
process.
Gasification of oil can be carried out at
either atmospheric or elevated presure. Oper-
ating at atmospheric pressure, a fluidized-bed
gasifier provides clean fuel to a combined
cycle gas and steam turbine power plant. The
potential for high efficiency and low capital
cost of such a plant makes this system
attractive.
It has been demonstrated that oil can be
gasified and sulfur removed from the resulting
fuel gases in a fluidized bed. The design and
evaluation of a fluidized-bed oil gasifier oper-
ating at atmospheric pressure have been com-
pleted. A 30 to 100-MW demonstration plant
is scheduled. A pressurized fluidized-bed oil
gasification system has been designed. The
economics and performance of a pressurized
oil gasification combined cycle power plant
have been evaluated.
IV-5-2
-------
ATMOSPHERIC PRESSURE FLUID-BED
OIL GASIFICATION
Westinghouse is evaluating the
gasification/desulfurization of residual oil at
atmospheric pressure under contract to the
Office of Research and Monitoring* (ORM) of
the Environmental Protection Agency. The
concept is being studied for the purpose of
producing, on-site, a low-sulfur fuel gas
suitable for power plant utilization in a con-
ventional boiler. Esso Petroleum Company has
provided experimental information on the oil
gasification/desulfurization process based on
small scale batch fluidized units and a 750-
kW continuous unit. Westinghouse has
carried out preliminary conceptual design
studies to evaluate the commercial process,
and is presently attempting to locate a utility
partner with which to carry out a demon-
stration plant operation.
Gasification/Desulfurization Concepts
Two possible modes for gasification/
desulfurization operation are the regenerative
mode and the once-through mode. Figure 1
illustrates the major process streams and
identifies the basic elements of the two
operational modes.
The elements of the regenerative operation
are the gasifier vessel and the regenerator
vessel. The gasifier is an air-fluidized bed of
lime operated at 1600°F with sub-stoichio-
metric air, ~ 20 percent of stoichiometric. The
oil is injected into the gasifier vessel where it
cracks and is partially combusted to form a
hot, low-sulfur fuel gas. Hydrogen sulfide is
produced during the gasification which reacts
with the lime to produce a sulfided lime.
H2S + CaO -» CaS + H20
(1)
The fuel gas is transported to the boiler
burners where combustion is completed and
the sulfided lime is sent to the regenerator.
The regenerator is an air-fluidized vessel
operated with a slight excess of air at 1900°F.
*Previously under the Office of Air Programs.
Regeneration takes place by reaction of
oxygen with the utilized lime to give an SO 2
rich stream (of about 10 mole percent SO2)
and a regenerated lime having a slightly
decreased activity compared to that of fresh
lime.
3
The SO 2 stream is transported to a sulfur
recovery system and the regenerated lime is
returned to the gasifier along with a stoichio-
metric amount of fresh make-up limestone.
The elements of the once-through
operation shown in Figure 1 consist of a
gasifier vessel and a sulfate generator, or pre-
disposal vessel. The operation of the gasifier
for once-through limestone operation is the
same as for the regenerative operation. The
sulfate generator operates similarly to the
regenerator, but at a lower temperature
( ~ 1500°F) so that the sulfided lime from the
gasifier is converted to calcium sulfate rather
than calcium oxide. The dry calcium sulfate
may be disposed of while the gas stream from
the sulfate generator is sent to the gasifier. A
limestone addition rate of 3 to 4 times that
used in the regenerative operation is necessary
in the once-through operation to achieve
similar sulfur removal of 90 to 95 percent.
Experimental Work
Under contract to the Office of Research
and Monitoring, Esso (England) is carrying
out laboratory tests on atmospheric-pressure
batch fluidized-bed equipment to investigate
lime sulfur absorption and lime regeneration
operating variables.1'2 The results of the Esso
Petroleum Company experimental program
have identified the critical phenomena
associated with atmospheric-pressure gasi-
fication/desulfurization, and have established
the probable operating conditions and
behavior of a commercial system. This
evaluation has been based on the results of the
small scale batch work. The batch work has
been carried out in conjunction with the con-
struction and operation of a 750-kW con-
tinuous pilot unit.
IV-5-3
-------
Design
Energy and material balances provide the
basic information with which the feasibility of
applying the gasification/desulfurization con-
cept as an add-on to an existing boiler, or as a
new plant design feature, has been examined.
The feasibility of the retrofit concept has been
examined in terms of the availability of space
in an existing power plant, the modifications
necessary to retrofit an existing boiler, and the
performance of a modified boiler. The specifi-
cations assumed for the conceptual design are
presented in Table 1.
The feasibility of converting an existing
boiler to one which utilizes the fuel from a
gasification/desulfurization system depends
on a number of factors, many of which will
differ from one boiler to the next. The space
available for a gasification/desulfurization
system in an existing power plant, the modifi-
cation necessary to retrofit an existing boiler,
and the performance of a retrofit boiler will
depend on the specific gasification/desulfur-
ization system design and choice of operating
conditions, the location of the gasifi-
cation/desulfurization system in the plant, the
type of boiler (coal-, oil-, or gas-fired), size of
boiler, the turn-down needed, the boiler load
factor, and the specific design features of the
boiler.
The gasification system may be placed
internal to the boiler (directly beneath) or
external to the boiler. Preliminary considera-
tions seem to favor the external retrofit design
over the internal retrofit design, and the
feasibility of retrofitting a coal-fired boiler
over the feasibility of retrofitting an oil- or gas-
fired boiler. The internal design concept
appears to be limited by the available space
beneath the boiler and the cost for modifying
the boiler. The external design concept should
reduce modification costs and should allow
greater uniformity and flexibility in system
design. Coal-fired boilers have more potential
space near the boiler than oil- or gas-fired
boilers and may provide some of the needed
solids handling and particulate clean-up
equipment. Boiler derate should be of little
concern with coal-fired boilers because they-
are designed to operate with slag on the boiler
heat transfer surface. Derate may be a concern
with oil- and gas-fired boilers due to a
reduction in the flame temperature and a
Tabte 1. SPECIFICATIONS FOR CONCEPTUAL DESIGN
Process specifications
Design variables
Sulfur removal: 90-95%
Fuel oil: 3 wt % sulfur; LHV = 17,700 Btu/lb
Limestone: 98.6% CaO yield
Boiler size: 600 MW
Load factor: 40%, 80%
Turn-down: 4/1
Operating variables
Gasifier temperature
Regenerator (sulfate
generator) temperature
Stone make-up rate
Air/fuel ratio
Limestone utilization
Fluidization velocity
Minimum fluidization
velocity
Particle sizes (avg)
Gasifier bed depth
(static)
Regenerative system
1600°F
1900°F
1 mole CaO/mole sulfur
20% of stoichiometric
~ 5 wt % sulfur in bed
8 ft/sec
3 ft/sec
~2000 Mm
2.5-3.5 ft
Once-through system
1600°F
1500°F
3 moles CaO/mole sulfur
20% of stoichiometric
~ 19 wt% sulfur in bed
8 ft/sec
1 ft/sec
~ 1000 pirn
3.5-4.0 ft
IV-5-4
-------
redistribution of the heat release within the
boiler which may occur with the fuel.
In contrast to boiler retrofit considerations,
the feasibility of incorporating a gasifica-
tion/desulfurization system into a new boiler
design will be limited only by the overall
economics of the system and the market
potential for new oil-fired boilers. The total
space occupied by the gasification/desulfuri-
zation system will be a small percentage of the
total plant volume. Also, because of the
flexibility in boiler design, boiler performance
will not be affected by the presence of the gasi-
fication/desulfurization system.
Figure 2 shows a plant layout for a 600-
MW once-through gasification/desulfur-
ization system. The plant consists of two gasi-
fication modules, each utilizing air from one
of the two power plant forced draft fans
present in the existing boiler.
Evaluation
The design study, coupled with the
experimental work of Esso Petroleum
Company, points out the technical and
economic feasibility of oil gasification/desul-
furization as a retrofit SO 2 control system for
utility boilers, or as an SC>2 control system to
be incorporated into a new boiler design. A
market for retrofit and new oil-fired boilers
exists.3
Preliminary investigations indicate that,
overall, the once-through operation may be
somewhat more attractive to a utility customer
than the regenerative operation. Capital
investment is reduced with the once-through
operation and, although the limestone feed
rate is expected to be three times the rate with
regenerative operation, the operating costs for
once-through operation may be less than those
for the regenerative operation. A complete
cost breakdown for once-through and regen-
erative operation has been presented.4 Once-
through operation has fewer technical prob-
lems at this time, and is an overall simpler
process than regenerative operation.
A comparison is made in Table 2 between
atmospheric pressure oil gasification/desul-
furization and the alternative schemes of low-
sulfur oil and stack gas cleaning. Capital costs
and operating costs are compared for new and
retrofit systems. Oil gasification/desulfuriza-
tion compares favorably with low-sulfur oil
and stack gas cleaning, based on prelimimary
cost estimates. A reduction of about 40 per-
cent in the capital costs involved in stack gas
cleaning is estimated for new and retrofit
gasification/desulfurization systems. Oper-
ating costs appear to be about the same for
once-through stack gas cleaning and regen-
erative gasification/desulfurization with sulfur
recovery. Once-through gasification/desulfur-
ization may reduce the operating costs 30 to 50
percent as compared to stack gas cleaning.
The cost estimate indicates that the operating
cost with low-sulfur fuel oil will be about 30 to
50 percent greater than the operating cost for
gasification/desulfurization. These conclu-
sions are based on the desulfurization of high-
sulfur residual oil (3 weight percent sulfur)
and may be altered when a lower sulfur oil is
considered (1 to 1.5 weight percent sulfur).
Environmental factors are also compared in
Table 2. The low-sulfur oil is advantageous in
that capital costs are limited to possible boiler
modifications necessary when changing from
gas or coal to low-sulfur oil. On the other
hand, operating costs are higher than those for
stack gas cleaning or oil gasification/desulfur-
ization. Capital costs are extremely high with
stack gas cleaning, especially on the retrofit
case, while operating costs are very near those
estimated for oil gasification/desulfurization.
Advantages of atmospheric-pressure oil
gasification over stack gas wet scrubbers
include:
Corrosion and fouling problems mini-
mized in SO 2 removal process and
boiler (minimum SOx and V),
No flue gas reheat required,
Uses crushed limestone — no lime-
stone pulverizing system needed,
IV-5-5
-------
Table 2. ASSESSMENT OF FLUIDIZED BED OIL GASIFICATION/DESULFURIZATIOIM
Cost
Capital, $/kW
New
Retrofit
Fuel adder, <£/106 Btu
New
Retrofit
Efficiency (thermal)
Environmental factors
S02, lb/106 Btu
NOX, Ib N02/106 Btu
Particulates, lb/106 Btu
Solid waste, ft3/MW-de
Sulfur removal
Stone
Ca/S
Make-up Ca/S
Low-sulfur
oil
a
25-35
25-35 a
—
0.35
0.40
0.06
V —
NA
NA
NA
Stack gas
cleaning
25-40
40-75
11-14
14-20
0.95-0.98
0.45
0.8
0.05
25
Oil gasification
Regenerative
operation
12-15
22-27
9.5-16.0
11-18
0.89-0.96 b
0.35
0.16
0.02-0.2 c
15
Limestone
rv 150
1.0
Once-through
ooeration
8-10
18-22
9-10.5
10.5-12.5
0.96-0.97 b
0.35
0.16
0.02-0.2 c
45
Limestone
3.0
NA
Equipment modifications are required when converting from gas or coal to low
sulfur oil.
bOverall efficiency is dependent on mode of temperature control.
C0.02 figure based on installing electrostatic precipitator (ESP).
0.2 figure based on installing high efficiency cyclone before burners and no ESP.
dCa/S rate dependent on regenerator temperature control scheme.
Basis: 3% Sulfur, 90% Sulfur Removal, 600 MW Capacity, 8% Load Factor
Simplified disposal — dry solids and
no disposal pond,
More compact system,
Reduced structural costs,
Lower auxiliary power requirement,
Reduced energy cost,
Improved NOX control, and
Potential market for spent CaO.
DEMONSTRATION PLANT PROGRAM
A three phase demonstration plant pro-
gram has been conceived.
Phase I. Preliminary design and cost esti-
mate of demonstration plant
installed on an existing boiler.
IV-5-6
-------
Phase II. Detail design and construction of
demonstration oil gasification pro-
cess.
Phase III. Developmental operation of the gas-
ification process and integrated
power plant.
A utility (or utilities) is required as a
cooperating party to carry out this program.
In Phase I the utility would supply technical
information on its existing power plant,
provide price and availability information on
oil fuels which might be utilized in the plant
currently and/or in the future, supply
information concerning the load requirements
placed on the plant, cooperate in the selection
of an engineering firm to prepare the
preliminary design, assist in selecting various
options in the design of the system and
evaluate, in cooperation with Westinghouse,
Esso (England) and EPA, the effectiveness and
economy of the oil gasifier/desulfurizer in
power generation and pollution abatement.
An engineering firm will carry out designs
in sufficient detail that fixed price bids can be
solicited for detailed design and construction
of the system. The design effort will be based
on experimental data from Esso (England) on
their 1-MW continuous pilot plant and the
conceptual design and assessment by
Westinghouse. EPA will provide general
guidance and funding for Phase I.
If the preliminary design confirms the
effectiveness and economics of the system, a
proposal would be prepared for Phases II and
III. In Phase II the utility would work closely
with the engineering firm in the design and in-
stallation of the gasifier/desulfurizer system
and share in the costs of the design and
installation.
In Phase HI the utility would operate the
plant, collect basic data on the operation of
the gasifier/desulfurizer system, aid in its
interpretation and analysis, and cooperate in
the evaluation of the effectiveness, technical
and economic, of the process in power
generation and pollution abatement.
Presentations have been made to 13
utilities who currently operate or plan to
operate oil-fired power plants. Presentations
began in March 1972 and have included
utilities from the East Coast, West Coast,
Florida and the South Central U.S. A West
Coast utility is actively interested in the pro-
gram. Seven utilities are currently evaluating
the proposal, and five utilities have indicated
they are not interested in participating in the
program at this time.
Problem areas for which utilities have
expressed concern are:
1. Hot fuel gas piping and valves.
2. Control and emergency conditions.
3. Space requirements.
4. Solid waste disposal.
5. Modifications and time required for
modifications to boiler.
6. Availability.
These and similar technical concerns will be
evaluated in detail during the preliminary
design phase. In addition, the prospect of
obtaining funds from EEI (Edison Electric
Institute) has been pointed out.
Conclusions
Performance
The concept has been technically demon-
strated with a 750-kW development gasifi-
cation plant.
Sulfur removal up to 95 percent can be
achieved.
Nitrogen oxide emissions of 150 ppm appear
possible.
Particulate emissions will be higher than
conventional oil- or gas-fired systems but can
easily be removed to achieve proposed
standards.
Further development effort is required in key
areas—e.g., calcium sulfate generation for
IV-5-7
-------
once-through operation, temperature control,
sulfur recovery.
Economics for comparable pollution abate-
ment
Capital cost of a retrofit, once-through gasifi-
cation system may be 50 to 70 percent less
than a retrofit wet scrubbing system.
Fuel adder cost for a retrofit, once-through
gasification system may be 30 to 50 percent
less than wet scrubbing, low-sulfur oil, or
desulfurized oil.
Market
Initial market is expected to be small boilers
( < 600 MW) on the East Coast, in the
Southwest where gas may be limited, or on the
West Coast. Once-through operation may be
favored over operation with sulfur recovery for
these plants.
PRESSURIZED
GASIFICATION
FLUID-BED
OIL
Two different processes are being con-
sidered for the pressurized gasification of
residual oil: (1) a pressurized version of the
fluidized bed oil gasification/desulfurization
process, being developed by Esso (England),
which has been explored only at atmospheric
pressure, and (2) the pressurized oil gasifica-
tion process of the type which has been
operated for gas manufacture by Shell5 and
Texaco.6 The Shell and Texaco processes are
not identical, but they are very similar in con-
cept, performance and cost. The process con-
cept is the important factor for this study so
the Shell and Texaco processes are not dis-
cussed individually. The Shell and Texaco
processes generate a low temperature (100-
250°F), clean fuel gas having a low heating
value of about 120 Btu/scf. Steam generated
by waste heat boilers in the gasification pro-
cess is also provided for the combined cycle
plant. The fluidized-bed process generates a
hot (~1600°F), clean fuel gas having a low
heating value of 200 to 500 Btu/scf (hot). Pre-
liminary cost and performance estimates for
these two process concepts have been
developed and are compared with alternative
oil-fueled power generation techniques having
pollution control.
Process Concepts and Options
Flow diagrams for the Shell and Texaco
processes and the pressurized fluid-bed
process are shown in Figures 3 and 4,
respectively. The Shell and Texaco processes
consist of an air-blown oil gasification vessel
(partial oxidation reactor) operated at a
temperature of about 2500°F with an air/fuel
ratio of about 45 percent of stoichiometric.
The hot gas is cooled in waste heat boilers,
producing saturated steam, prior to purifica-
tion of the gas. The gas purification process
consists of a carbon (soot) recovery step and a
sulfur removal section. Recovery soot ( ~ 3
weight percent of the fuel oil feed rate) is
recycled to the gasifier vessel, and H2S
produced in the sulfur removal section is sent
to a sulfur recovery plant to recover elemental
sulfur.
The pressurized fluid-bed oil gasification
process shown in Figure 4 consists of a
fluidized bed gasifier/desulfurizer vessel, a
limestone/dolomite regeneration section, and
a sulfur recovery section. The phenomena
taking place in the pressurized fluid bed
gasifier are essentially the same as have been
described for the atmospheric pressure fluid
bed case. The gasifier is operated at about
1600°F with an air/fuel ratio of about 14-25
percent of stoichiometric. Both limestone and
dolomite are considered as sulfur absorbents
in the pressurized fluid-bed case. The major
process options which have been examined
with respect to cost and performance are:
1. The gasifier air/fuel ratio. Air/fuel
ratios of 14 percent and 25 percent of
stoichiometric giving fuel gas low heating
values (hot) of about 500 and 280 Btu/scf,
respectively, have been considered. An air/fuel
ratio of 14 percent of stoichiometric is
assumed to be the minimum air/fuel ratio at
which a gasifier temperature of 1600°F can be
maintained, while 25 percent of stoichiometric
IV-5-8
-------
is assumed to be a conservatively high air/fuel
ratio according to the experience gained from
the Esso (England) atmospheric pressure
operations. The physical feasibility of
operating at an air/fuel ratio as low as 14
percent of stoichiometric without excessive
carbon deposition in the gasifier must be
demonstrated.
2. The limestone/dolomite regeneration
method. The regeneration of calcium suflide
and the production of a high sulfur gas for
sulfur recovery can be achieved by either of
two processes:
H2O
to
Regeneration with CO2 and
produce H2S. With this process the reaction
- CaCO+ H2S (3)
is utilized to produce calcium carbonate and
H2S-rich gas. The reaction is favored by high
pressure and would be carried out at a
temperature of about 1100°F, with H2S
concentrations of about 9 percent by volume
being projected. Two sources of CO2 for the
regeneration scheme have been considered--
CO2 provided by scrubbing flue gas from the
combined cycle plant, and CO2 provided by
scrubbing a gas stream produced by
combustion of a portion of the fuel gas.
Regeneration with air to produce SO2 . The
reaction of oxygen with calcium sulfide,
CaS + 3/2O2 - CaO + SO;
(4)
is favored by low pressure and high
temperature (~2100°F). Although this regen-
eration scheme would yield an SO 2
concentration of only about 2 percent by
volume at pressures of 10-15 atm, it is
considered because of its apparent simplicity.
If a high stone make-up rate is required for
the regenerative processes and if high stone
utilization can be achieved in the gasifier, a
once-through system may be attractive. Once-
through operation would require conversion of
calcium sulfide to calcium sulfate before
disposal of the stone. The reaction
CaS + 2O2 ^ CaSO4 (5)
could be applied for this purpose, and would
be carried out at 1400-1700°F. Limestone
utilization in a once-through process is
expected to be 35 percent or higher.
Other process options examined are the
pressure drop required across the gas turbine
combustor and the option of cooling the gas
produced by fluidized-bed oil gasification
before it is combusted in the combined cycle
plant.
Process Specification and Design Basis
Table 3 lists the factors specified for the
conceptual design of the pressurized fluid-bed
oil gasification process. The specifications for
the plant capacity, capacity factors, and the
turndown ratio are also assumed for the Shell
and Texaco processes. The combined cycle
power generating plant is a Westinghouse
PACE Plant consisting of two Westinghouse-
501B gas turbines and a single 28.5-in. steam
turbine. The pressure of the fuel gas to the gas
turbine is 215 psia. Designs and energy costs
are based on operation using limestone for de-
sulfurization. Sulfur removal of 90 to 95 per-
cent from a 3 weight percent sulfur residual oil
Table 3. SPECIFICATIONS FOR FLUID-BED
OPERATION
Plant electrical capacity
Capacity factor
Turndown ratio
Number of gasifier modules
Modes of operation
Gasifier pressure
Pressure of gas turbine
Sulfur removal
Residual oil
Gasifier temperature
Regenerator temperature
Suifate generator terriperature
Air/fuel ratios
Lime particle diameters
Regenerative lime utilization
Once-through time utilization
Limestone make-up rate
Gasifier temperature control
Regenerator temperature control
Sulfate generator temperature
control
Plant heat rate
250 MW (PACE Plant)
70%
4/1
2
Regenerative or once-through
~15 atm
11 atm 1165 psia to turbine)
90-95%
3 wt % sulfur
1600°F
1100°Fa
1700°F with once-through option
14 (minimum} Er 25% of stoichiometric
500-2000 (im average diameter
10%
35%
1 mole CaO/mole sulfur fed
Stack gas recycle, steam or water
injection
Lime circulation rate; water injection
Excess air circulation {-v 200%
excess)
9000 Btu/kWhr {assumed for purposes of
material balances)
With Co2/H2O regeneration; 2100°F .with air regeneration.
IV-5-9
-------
is specified for the fluid-bed process, with
specifications for vessel temperatures, lime-
stone utilizations, and limestone make-up rate
based both on thermodynamic information
and the atmospheric pressure data of Esso
(England). Control of the vessel temperatures
is assumed to be easily carried out by any of
the methods suggested in the table, based on
atmospheric information. A plant heat rate of
9000 Btu/kWhr was assumed for the purpose
of finding the approximate fuel consumption
of a PACE Plant.
Shell and Texaco both supplied energy and
material balance information for their
processes along with capital investment
estimates. The Shell and Texaco processes had
not necessarily been modified to provide the
optimum power generation performance, but
were based mainly on gas manufacturing
experience.
Material and energy balance information
for the pressurized fluid-bed oil gasification
process was based on Esso (England)
atmospheric pressure data. Performance and
vessel sizes were modified for effects of
pressure. Regeneration system designs were
based largely on thermodynamic behavior.
The general behavior of the fluidized-bed gas-
ifier for sulfur removal, vanadium removal,
and carbon deposition is assumed to be
independent of pressure once the gasifier bed
diameter and bed depth has been scaled for
the affect of pressure.
Material and Energy Balances
Figures 5 and 6 show simple block flow
diagrams for the Shell and Texaco processes
and the pressurized fluid-bed oil gasification
process, respectively. Block diagrams of this
type have been utilized to analyze the per-
formance of the gasification processes and the
performance of the complete power plant.
Inputs to the Shell and Texaco processes,
Figure 5, are shown to be residual oil, air,
water, booster compressor power, and
auxiliary power for pumping oil, boiler water,
cooling water, scrubber recycle, etc. Output is
IV-5-10
cold fuel gas, steam, sulfur, and energy losses.
Energy losses for the Shell and Texaco
processes arise from heat losses, cooling water
losses, and sensible heat of flue gas from gas
purification and sulfur recovery sections. The
thermal efficiency of the Shell and Texaco
processes as presently conceived is less than 80
percent; though improvements in this per-
formance factor may be made by simple
process alterations. Shell has recently
indicated 87 percent thermal efficiency for
their process.
Inputs to the fluidized-bed process (Figure
6) are residual oil, air, water and steam, lime-
stone (or dolomite), booster compressor power,
and auxiliary power for pumping oil, water,
solids circulation, and gas compression in the
regeneration section of the gasification system.
Output is hot fuel gas for the combined cycle
plant, sulfur, and energy losses in the form of
sensible heat of the spent limestone, heat
losses, carbon deposition losses, and sensible
heat of flue gas from the regeneration and
sulfur recovery sections. The thermal
efficiency of the fluidized-bed gasification
system will depend slightly on the air/fuel
ratio and the regeneration method used, but
will be about 90 to 95 percent for all the
options considered.
Fuel compositions and heat values are
shown in Table 4 for the two oil gasification
processes. Shell provided expected product
Table 4. GASIFICATION PRODUCT
COMPOSITIONS
s
N2
H2
CO
C02
H20
CH4
C2H4
H2S
Low heating
value (hot),
Btu/scf
hell process,
Vol%
60.79
14.57
22.92
1.37
0.00
0.35
0.00
0.00
117
Fluidized bed process,
Vol%
14A/F
50.74
0.82
13.40
6.70
0.00
9.43
18.86
0.05
~500
25A/F
47.74
2.64
7.97
7.97
20.56
6.55
6.55
0.02
~280
-------
compositions for a dry gas, while the product
gas compositions for the fluidized-bed process
at air/fuel ratios of 14 and 25 percent of
stoichiometric have been estimated from
atmospheric pressure data. Projected plant
performance has been based on these product
compositions.
Capital Investment Evaluation for the
Pressurized Fluid-Bed Process
Figure 7 is a flow diagram for the
pressurized fluid-bed oil gasification process
with limestone regeneration be reaction with
CO2 and H2O to form CaCO3 and H2S^ The
source of CO2 is flue gas from the combined
cycle plant. The alternative CO2 source,
combustion of a portion of fuel gas, has been
eliminated based on the comparative
economics of the two options. The equipment
shown in Figure 7 consists of fuel oil and lime-
stone handling equipment, air booster
compressor, gasifier vessels (2 modules) with
multi-stage particle collection, and a
CO2/H2O regeneration system.
The process has been separated into four
component systems—the gasification system,
the CO2/H2O regeneration system, the
booster compressor, and the Claus plant.
These four component costs are shown in
Figure 8, in units of S/106 Btu-hr of oil feed
(HHV), as a function of the pressure at the
gasifier product gas outlet, with the air/fuel
ratio as a parameter. Two cases of pressure
drops required in the gas turbine combustor
and the fuel distribution equipment are
considered in the figure—8 psi AP and 53 psi
A P. The pressure drop across the gasifier and
particle control equipment is assumed to be 9
percent of the gasifier pressure.
Figure 8 leads to the following conclusions:
1. Reducing the air/fuel ratio from 25 to 14
percent of stoichiometric reduces the
capital investment by about 10 percent in
units of $/106 Btu-hr.
2. For the case of a 10 atm gas turbine,
increasing the combustor pressure drop
from 8 psi to 53 psi reduces the process
cost by about 10 percent because of the
highly pressure sensitive nature of the
CO2/H2O regeneration scheme.
3. A once-through scheme would reduce the
cost of the process by more than 40
percent, neglecting the slight cost increase
due to a sulfate generating vessel.
4. The cost of the fluidized-bed process with
an air-blown regenerator would be some-
where between the cost for the CO2/H2O
regeneration system and the once-through
gasification system.
The true capital investment for pressurized
fluid-bed oil gasification will depend on the
plant performance as a function of the design
options-air/fuel ratio, combustor pressure
drop, and regeneration method.
Performance
Cycle studies based on estimated material
and energy balances and approximate gas
compositions (Table 4) have led to estimates of
the plant heat rate, the plant power, capital
investment, and energy costs, using
pressurized oil gasification with combined
cycle power generation (PACE Plant). The
following conclusions concerning the
combustor pressure drop have been deduced:
1. With the pressurized fluid-bed oil
gasification process an increase in the
required combustor pressure drop of 10
psi increases the plant heat rate by only 10
Btu/kWhr with operation at an air/fuel
ratio of 14 percent of stoichiometric, and
30 Btu/kWhr percent.-
2. With the pressurized fluid-bed oil gasifi-
cation process an increase in the required
combustor pressure drop of 10 psi reduces
the plant power by about 0.5 MW with
operation at 14 percent of stoichiometric,
and about 1.0 MW at an air/fuel ratio of
25 percent of stoichiometric.
3. With the Shell and Texaco processes,
increasing the required combustor
pressure drop 10 psi increases the plant
IV-5-11
-------
heat rate by about 70 Btu/kWhr, and
decreases the plant power by about 3
MW.
From these points it is concluded that the
combustor pressure drop is a fairly insensitive
parameter with respect to plant performance
and should be determined by the requirements
for fuel distribution and control.
Table 5 summarizes the performance,
capital cost, and energy cost of the Shell and
Texaco processes and the fluidized-bed
process with a combustor pressure drop of 50
psi for the two air/fuel ratios, and for regener-
ative operation by CO2/H2O, and for
once-through operation. The figures shown in
the table refer to the case in which the 1600°F
fuel gas from the fluid-bed process is not
cooled before combustion. Table 5 indicates
that the plant heat rate with the fluid-bed
process increases with increasing air/fuel
ratio, while the heat rate is comparable for the
regenerative and once-through operations.
The plant heat rate is less than 10,000
Btu/kWhr even for the high air/fuel ratio case
of 25 percent of stoichiometric. The Shell and
Texaco processes yield a plant heat rate of
about 13,000 because of the low thermal
efficiency, the high booster compressor power
requirements, and the relatively high ratio of
power produced by the steam turbine to the
power generated by the gas turbine. A plant
heat rate of 11,000 Btu/kWhr is estimated for
87 percent thermal efficiency. Plant power is
comparable for all of the cases shown, with the
Shell and Texaco processes yielding the
highest power. Capital investment is based on
PACE Plant cost information and the cost
estimates shown in Figure 8. The cost basis is
listed in Table 5. Capital cost is reduced for
the fluid-bed process with once-through
operation due to the great expense involved in
limestone regeneration by the CO2/H2O
method. The change in plant capital cost in
going from 14 percent air/fuel ratio to an
Tables. PROCESS PERFORMANCE
Plant heat rate
(HHV), Btu/kWhr
Plant power, MW
Plant capital cost,
$/kW
Total energy cost,
mills/kWhr
Break-even distillate
cost, tf/106 Btu
Fluid bed process
CO2^H2O limestone regeneration
14% A/F 25% A/F
9,007
266.0
179.5
9.43
65.2
9,828
276.0
182.1
9.93
71.1
Fluid- bed process
once-through operation
14% A/F 25% A/F
8,906
269.0
158.8
8.96
59.7
9,716
279.0
159.5
9.42
65.1
Shell and Texaco
thermal efficiency
75% 87%
13,000
282.4
246.7
13.34
111.1
11,000
282.4
246.7
12.44
98.7
Conditions Assumed:
1. 50 psi combustor pressure drop.
2. 1600°F fuel gas temperature from fluid-bed process with no cooling.
3. Fixed charges at 15%.
4. Residual oil at 45i|;/106 Btu.
5. Limestone at S6/ton.
6. 5% contingency; 4% escalation; 8% interest during construction; 2 years construction time; 2%
A&E; 70% capacity factor.
7. No credit for sulfur recovered.
IV-5-12
-------
air/fuel ratio of 25 percent of stoichiometric is
slight. The plant capital investment for the
Shell and Texaco processes are based on the
estimated gasification system capital cost of
about $100/kW provided by both Shell and
Texaco.
Energy cost assumptions are listed in Table
5. The 14 percent air/fuel ratio operation of
the fluid-bed process is more economical than
the 25 percent air/fuel ratio, while the energy
cost for once-through operation is also less
than regenerative operation. The relationship
between energy cost of regenerative and once-
through operation is dependent on the cost of
limestone disposal, but even if the limestone
cost should double to $12/ton the once-
through and regenerative system energy costs
will be about equal.
Break-even costs for No. 2 distillate fuel oil
is given in $/10& Btu and represent the price
at which No. 2 distillate must be available for
the PACE Plant to operate at the same energy
cost as with residual oil gasification. With No.
2 distillate prices ranging from 85-95
-------
capital about $10/kW, not including the
cost of a waste heat boiler system and soot
removal equipment. The second option
would result in inefficient operation with
high energy losses, as with cooling of the
gas by water injection. Both options seem
unattractive for power generation, though
further analysis is required to develop
quantitative conclusions.
Comparisons With Alternative Power
Generation Systems
Tables 6 and 7 compare capital invest-
ments and energy costs of pressurized fluid-
bed oil gasification with alternative oil-fueled,
pollution controlled, power generation
systems. Capital and energy costs for a con-
ventional oil-fuel power plant utilizing lime-
stone-wet scrubbing for pollution control, a
pressurized fluid-bed combustion plant fueled
by oil, and a PACE Plant fueled with No. 2
distillate, are compared with capital and
energy costs for a PACE Plant with pres-
surized fluid-bed oil gasification. The fluid-
bed process is carried out at a 25 percent
air/fuel ratio with limestone regeneration by
the CO2/H2O method, representing the
highest capital cost and lowest efficiency of
the fluid-bed cases considered. Due to the low
cost of the PACE Plant package, the PACE
Plant with fluid-bed oil gasification is about
140$/kW cheaper than a conventional power
plant and 60$/kW cheaper than a regenerative
pressurized fluid bed combustion plant fueled
with oil. Energy cost of the PACE Plant with
fluid-bed oil gasification, assuming 4520
percent are projected. This conclusion holds
over the range of factors explored — air/fuel
ratios, combiistor pressure drops, and lime-
stone regeneration methods.
Table 6. CAPITAL INVESTMENT COMPARISONS
Total capital cost,
$/kWc
Assumptions
Construction time,
years
Sulfur removal equip-
ment, $/kW
Plant capacity, MW
Conventional oil-
fired power plant
with scrubber
323.28
4.5
50.0
635
Pressurized
fluid bed oil
composition3
243.67
3.5
12.68
635
PACE plant
with no. 2
distillate fuel
137.80
2.0
—
269
PACE plant
with residual
oil gasification6
182.1
2.0
44.3
269
aOperated with limestone regeneration.
bC02/H20 regeneration of limestone and air/fuel ratio of 25% of stoichiometric-no
cooling.
C5% contingency; 4% escalation per year; 8% interest during construction; 2% AErE;
70% capacity factor.
IV-5-14
-------
Table?. ENERGY COST COMPARISONS
(mills/kWhr)
Fixed charges
Fuel
Limestone
Operating and main-
tenance
Total
Conventional oil-
fired power plant
with scrubber
7.91
4.11
0.12-
0.91
13.05
Pressurized
fluid bed oil
combustion
5.96
4.35
0.13
0.82
11.26
PACE plant
with no. 2
distillate fuel
3.37
7.80
—
0.52
11.69
PACE plant
with residual
oil gasification
4.45
4.42
0.13
0.93
9.93
Assumptions
1. Fixed charges at 15%; 70% capacity factor.
2. 3 wt% sulfur residual oil at 45^/106 Btu.
3. Limestone at $6/ton.
4. No.2 distillate at gO^/IO6 Btu.
5. No credit for sulfur recovered.
2. Cooling of the 1600° F fuel gas produced
by pressurized fluid-bed oil gasification
with no recovery of the energy or by water
injection is unattractive from the
standpoint of capital and energy cost.
Cooling to 400° F with 25 percent recovery
of the energy may be uneconomical and
further analysis is required.
3. An experimental study of pressurized
fluid-bed oil gasification should be carried
out to determine the process behavior in
critical areas and to obtain a better
understanding of the economic
advantages to be gained from the process.
4. The process concept utilized by Shell and
Texaco, though not as efficient or
economical as the fluid-bed process
concept, should be considered for
combined cycle power generation because
it depends largely on existing technology.
Improvement of the process performance
may be possible.
ACKNOWLEDGMENTS
This work was performed under contract to
the Office of Research and Monitoring,
Environmental Protection Agency. P. P.
Turner served as Project Officer. W. L.
Wright, Westinghouse Power Generation
Systems Division, arranged and was a con-
tributor to utility presentations on the atmos-
pheric pressure demonstration plant. Esso
(England) also contributed to the utility pre-
sentations.
REFERENCES
1. Moss, G. The Desulfurizing Combustion of
Fuel Oil in Fluidized Beds of Lime
Particles. (Presented at First International
Conference on Fluid-Bed Combustion.
Hueston Woods. November 1968.)
2. Craig, J. W. T., G. L. Johnes, G. Moss, and
J. H. Taylor. Study of Chemically Active
Fluid Bed Gasifier for Reduction of Sulfur
Oxide Emissions. Interim Report. Esso
Research Centre, Abingdon, Berkshire,
England. Prepared for the Air Pollution
IV-5-15
-------
Control Office, Environmental Protection
Agency, Research Triangle Park, N. C.
under Contract Number CPA 70-46.
August 1970.
3. Hauser, J. G. Optimum Uses of Energy
Sources. (Presented at Spring Conference,
Southeastern Electric Exchange. New
Orleans. April 1971.)
4. Newby, R. A., D. L. Keairns, and D. H.
Archer. Assessment of Fluidized Bed Oil
Gasification for Power Generation.
(Presented at the 65th Annual Meeting of
the Air Pollution Control Association.
Miami Beach. June 1972.)
5. Information provided during 'discussions
with Shell, May 3, 1972.
6. Information provided during discussion
with Texaco, May 24, 1972.
IV-5-16
-------
T0
STACK
COMBUSTION AIR
LIMESTONE
MAKE-UP^
•HH^
~l
I
GASIFIER
CLEAN
BOILER
FUEL GAS k
r^*»^ TEMPERATURE
•*^ CONTROL
AIR
REGENERATED
LIME
S02RICH
STREAM
STREAM
SULFUR
RECOVERY
SULFATED
LIME
DISPOSAL
REGENERATIVE MODE
COMBUSTION AIR
CLEAN FUEL GAS
, ^\
A T0
I STACK
I
BOILER
LIMESTONE FEED
GASIFIER
FUEL
| OIL
I A
I. »»,
I
SULFATE
GENERATOR
TEMPERATURE
CONTROL
STREAM
GAS
LIQUID
SOLID FEED
SOLID CIRCULATION
SULFATED
LIME
DISPOSAL
ONCE-THROUGH MODE
Figure 1. Modes of operation.
IV-5-17
-------
40ft
1. GASIFIER-DESULFURIZER
2. SULFATE GENERATOR
3. SULFATED-LIME BUNKER
4. LIMESTONE BUNKER, FEEDER
5. HIGH-EFFICIENCY CYCLONE
6. FORCED DRAFT FAN
7. GASIFIERFAN
8. SULFATE GENERATOR FAN
9. OIL FEED LINE
10. FUEL GAS LINE
11. LIME CIRCULATION LINE
12. SULFATED-LIME DISPOSAL LINE
13. BURNERS
SOLIDS TRANSPORT
IHHIHIHH GAS TRANSPORT
Figure 2. Retrofit of 600-MW boiler, external once-through design.
IV-5-18
-------
OIL
ELECTRIC
POWER
COOLING
WATER
BOILER FEED
WATER
RESIDUAL
FUEL
OIL
TO
MOTORS, etc.
TO C. W.
USERS
-*• TOW. H.
BOILERS
FUEL OIL
STORAGE, PUMPING,
AND PREHEATING
OIL TO/FROM —
C RECOVERY -*-
AIR
FUEL
GAS
AIR
AIR
1
STEAM
STEAM
WASTE HEAT
STEAM
GASIFICATION
AND
WASTE HEAT
STEAM
GENERATION
AIR
COMPRESSOR
AND MOTOR
ASH
SLURRY
TO DISPOSAL
POND
TO
RAW
FUEL
GAS
UR
)SAL
RBON
SULFUR
RECOVERY
i
H2S
REMOV
SYSTE
H£S
AL
M
STACK
GAS
I
FUEL
CLEAN
FUEL
GAS
TO
WESTINGHOUSE
SYSTEM
FREE
FUEL
GAS
CARBON
RECOVERY
CARBON/OIL
SLURRY TO
FUEL OIL
STORAGE
FUEL OIL
TO
OIL
PREHEAT
MAKEUP
NAPHTHA
FROM STORAGE
Figure 3. Pressurized oil gasification for gas turbine fuel.
en
-------
FUEL GAS
COIBBUSTOR
•f
H2$ OR SOj
STREAM TO
SULFUR
RECOVERY
LIME/DOLOMITE
DISPOSAL
REGENERATOR
SULFIDED
UME/
DOLOMITE
GENERATOR
BOOSTER
COMPRESSOR
H20, C02
OR
AIR
STEAM
OR
WATER
INJECTION
BOILER
^ J—I
SUPERHEATED
STEAM
•CONDENSER
Figure 4. Regenerative pressurized oil gasification process.
IV-5-20
-------
SATURATED
STEAM TO
COMBINED CYCL
PLANT
WATER
RESIDUAL
FUEL
OIL
I. i COLD (100 TO 250 °F)
FUEL GAS TO
E COMBINED CYCLE PLANT
SHELL/TEXACO
OIL GASIFICATION
SYSTEM
ENERGY LOSSES
. ^. (HEAT LOSSES SENSIBLE HEAT
FLUE GAS, COOLING WATER)
i i *
T A | AUXILIARY POWER (BOILER FEED WATER PUMPING
1 / \ | PUMPING, GAS CLEANING SYSTEM PUMPING)
OF
,OIL
BOOSTER I
COMPRESSOR
POWER
BOOSTER COMPRESSOR
AIR FROM GT COMPRESSOR
(AIR/FUEL RATIO*45% OF STOICHIOMETRIC)
Figure 5. Energy balance for the Shell/Texaco oil gasification process.
IV-5-21
-------
HOT (1600 °F)
FUEL GAS TO
LIMESTONE/
DOLOMITE
80°F
STEAM _
WATER _
UUINDINC.U UTL,
FLUIDIZED BED
Oil G/KIFlPATinN
SYSTEM
LC ru\n I
** ENERGY LOSSFS (HFAT 1 n«F<:,
CARBON DEPOSITION, SENSIBLE
HEAT OF FLUE GAS, COOLING WATER)
»^e|i| Clip
fl A AUXILIARY POWER (COOLING WATER PUMP-
X 1 ING, OIL PUMPING, SOLID CIRCULATION,
•-— /\ 1 COMPRESSION FOR REGENERATION)
BOOSTER
COMPRESSOR I
POWER
BOOSTER COMPRESSOR
AIR FROM GT COMPRESSOR
(AIR/FUEL RATIO=14 TO 25% OFSTOICHIOMETRIC)
Figure 6. Energy balance for the fluidized-bed oil gasification process.
IV-5-22
-------
BASIS: 250-MW COMBINED CYCLE PLANT
TWO 125-MW GASIFIER MODULES
15-atm GAS TURBINE PRESSURE
50 ton
FUEL GAS
TO COMBINED
CYCLE PLANT
COVERED
HOPPER
CARS
No. 2 FUEL
OIL FOR
STARTUP TO
I FLUE GAS FROM
y COMBINED CYCLE PLANT
__ __
TO STACK
CONDENSER
LIMESTONE
RECEIVING
HOPPER
ton
C02
ABSORBER,
STRIPPER
WATER -*-
(ACIDIC)
MULTI-
STAGE
CYCLONES
STARTUP
OIL
STORAGE
COMPRESSOR
V
REGENERATED
_ LIMESTONE
INJECTION .
TRANS-
PORT
AIR
OIL PUMPS
150 gal/min -
SOOpsiAP
PETROCARB
INJECTION
TO 2nd MODULE
TO 2nd
MODULE
8,800 bbl/day
T__L_'k
BOOSTER I
AIR
COMPRESSOR TEMPERATURE
CONTROL
WATER
STONE
DISPOSAL
WATER FOR
TEMPERATURE
CONTROL INJECTION
PARTICLE
COLLECTOR
REGENERATOR GAS
TO CLAUS PLANT
SOLID
LIQUID
GAS
Figure 7- Pressurized oil gasification plant flow diagram.
IV-5-23
-------
3200
2800
2400
± 2000
§ 1600
o 2200
2 800
BASIS:
MID-1972 COSTS
AP GASIFIER/P GASIFIER OUTLET=0.09
3 wt % SULFUR IN FUEL OIL
95% SULFUR REMOVAL
AIR/FUEL RATIO=14% OF STOICHIOMETRIC
AIR/FUEL RATIO =25% OF STOICHIOMETRIC
INDEPENDENT OF MR/FUEL RATIO
53 psi AP IN COMBUSTOR AND FUEL CONTROL
I Vps\ A P IN COMB. AND FUEL CONTROL I ICLAUS PLANT
. I BOOSTER
?COMPRESSOR-
•J
120 140
160
180 200 220 240 260
PRESSURE AT GASIFIER PRODUCT GAS OUTLET, psia
280
300
32
Figure 8. Fluidlzed-bed oil gasification capital investment.
IV-5-24
-------
6. FUEL GASIFICATION AND ADVANCED POWER
CYCLES-A ROUTE TO CLEAN POWER
F. L. ROBSON
United Aircraft Research Laboratories
ABSTRACT
The United States is currently faced with a growing gap between th,e demand for electrical
energy and the supply of economic fuels for generating this energy with minimum environmental
impact. The use of advanced power cycles utilizing technological spinoffs from the aerospace
industry in conjunction with fuel gasification/desulfurization offers a solution which could prove
to be not only technically feasible but economically attractive. A review of one such system, the
Combined Gas And Steam (COGAS).is presented and the technical and economic advantages are
enumerated. There are, however, several problem areas, particularly in the interface between the
power system and the fuel system which must be resolved before the overall concept becomes a
commercially viable one. These problem areas are presented with the intent of provoking
thoughtful discussion and perhaps of opening new areas of research among the conference
attendees.
INTRODUCTION
The majority of the people attending this
conference were aware of the current energy
crisis facing this country well before there was
in fact a crisis. This is not the time to talk of
the reasons for the current situation but rather
to discuss methods of alleviating it by using
the Nation's vast supply of coal, a fuel source
now held in low esteem by a large segment of
the air pollution regulatory agencies. It is
apparent that unless extensive effort is directly
applied towards this goal, or unless equally
extensive institutional changes are brought
about in this country, we will be faced with a
utility system based upon foreign sources for
one portion of our fuel and what can only be
termed an adolescent nuclear industry for the
remainder.
The use of fossil fuels in the utility industry
will hopefully not parallel that currently being
followed in the transportation industry where
each new reduction in emissions is
accompanied by a corresponding increase in
fuel consumption. To assure this, any power
system utilizing the advantages promised by
fluidized-bed combustion/gasification must
have the potential of achieving operating
efficiencies significantly higher than currently
attainable in conventional boilers.
However, the various economic forces,
natural and imposed, which now and in the
future affect the energy scene are such that a
delay in the introduction of these advanced
power systems could result in electrical power
becoming more of a luxury than a necessity.
Thus, these advanced power systems must be
based upon technology which is now well in
hand, but which will continue to grow, thus
affording better performance and economics
as each new generation of power system is
achieved.
rv-6-i
-------
The system described briefly in the using evolutionary changes of technology cur-
following paragraphs does indeed use the cur- rently being demonstrated will offer power
rent technology available in the aircraft gas systems which could make use of the pollution
turbine industry applied to the industrial seg- reductions offered by fluidized-bed com-
ment to realize efficiency and economic gains. bustors and still attain efficiencies well beyond
Second and third generation power systems those of conventional steam-electric plants.
IV-6-2
-------
THE POWER SYSTEM
Current steam power plants have
efficiencies approaching 40 percent at 1000 °F,
a value which is limited not by
thermodynamics but by economics. There
have been steam systems designed and
operated at temperatures of 1200 °F, but the
initial boiler and turbine costs were high and
the maintenance associated with operation at
this temperature eventually caused derating to
1000° F.1 What, then, are the alternatives
available to increase the efficiency of power
systems? The answer comes of course from
Carnot's law which defines the efficiency limit
of any heat engine operating between two
temperature limits. Referring to Figure 1, it
can be seen that while the theoretical Carnot
efficiency is well above that obtained by a real
power cycle, several of the advanced power
cycles demonstrate efficiencies which are
nearly 70 percent of the theoretical limit. The
three fossil fuel-fired sytems having the
highest efficiency are based on the potassium
topping cycle, the COGAS cycle, and an MHD
topping cycle. Each of these cycles uses a high-
temperature cycle to top a more or less
conventional steam cycle. Of the three
mentioned, however, only the COGAS system
has been demonstrated in commercial size
and, in fact, nearly 2500-MW of these systems
are currently on order by the utilities.2
The advantage the COGAS system has over
the potassium and MHD topping systems is
that of development. The COGAS system is
evolutionary in nature, adapting the advances
demonstrated in military and commercial air-
craft to the large industrial turbomachines.
Thus, the technology of the JT9D engines,
used on the 747, which allows operation at
2000 °F and above will be adapted to the next
generation of industrial machines.
In order to estimate the performance of the
COGAS system, the relation of system
elements must be established. While there are
many ways of combining the gas turbine and
steam system equipment, essentially two
generic types emerge: (1) the waste-heat
recovery type in which the turbine exhaust
raises steam with or without additional firing,
and (2) the pressurized-boiler type in which
steam is raised and superheated using heat
from a gas turbine combustor. These
variations are shown in Figures 2 and 3,
respectively. The effect of configuration of
performance can be obtained by interpreting
Figure 4, which shows station efficiency versus
gas turbine participation for the waste-heat
recovery type configuration.
The parameter gas turbine participation is
really a measure of oxygen utilization. For a
given turbine inlet temperature, a fixed
amount of oxygen is required for combustion;
the remainder, plus a large amount in the
dilution air, is exhausted through the turbine.
In the waste-heat recovery system, if there is
no additional firing, the efficiency is the
highest value, i.e., the right end of the lines of
efficiency in Figure 4. If there is firing using
up the additional oxygen and generating more
steam, the gas turbine participation declines,
and the efficiency becomes lower until the left
end of the line, the point of which all the
oxygen is consumed, is reached. The foregoing
applies to combined-cycle systems with
turbines operating at 2000°F and above.
Below this turbine inlet temperature, the
combined-cycle efficiency could, in fact,
increase with high steam fractions since it is
possible to have steam-cycle efficiencies signi-
ficantly better than the efficiency of the lower
temperature gas turbine.
In the supercharged cycle, fuel and air are
burned essentially at stoichiometric conditions
and the combustor exhaust is cooled to tur-
bine inlet temperature by raising steam rather
than be dilution air. This raising of steam by
combustion has the same effect as supple-
mentary firing, i.e. the system efficiency is
reduced. Actually these systems are, obeying
Carnot's law, which in essence says that the
system which uses the entire heat input at the
combined-cycle efficiency will be more effi-
cient than the system which utilizes only part
of the heat input at combined-cycle efficiency
and part at the lower steam-cycle efficiency.
IV-6-3
-------
There are also considerations outside of the
power cycle which influence the ratio of gas
turbine power to steam power which will
appear again when the integrated gasifica-
tion/power system is discussed.
THE GASIFICATION SYSTEM
There is a wide variety of methods for
converting solid or liquid fuels to gaseous
form. Rather than discuss the operational
characteristics of any one of these processes, it
would prove more fruitful to discuss those
characteristics which are important from the
viewpoint of utilizing the fuel in an advanced-
cycle power system.
Perhaps the most important aspect of fuel
supply is cleanliness. While the turbine can
handle a wide variety of fuels, each of these
fuels must meet rather rigid contamination
criteria. For gaseous fuels, the current specifi-
cations3 will allow no more than 0.08
lb/106ft3 of total solids. No particulate size is
specified, but filters are located in the fuel
lines capable of removing particles of 30
jum and above. To minimize blade erosion,
however, particulates should be small enough
to follow the air stream through the blading
without impingment. The exact particulate
size has not been firmly defined but measure-
ments of smoke particles seem to indicate few
larger than 20 to 25 jum.
Current specifications limit total sulfur
content in the fuel to 162 lb/106ft3 of which
HzS can be no more than 0.18 percent by
volume. This amount, assuming it was all
converted by combustion to SC>2, would be the
equivalent of about 3 lb/106 Btu of methane-
type fuel gas or about 2 lb/106 Btu of low-
heating-value gas from coal. Both of these
values are above EPA regulations (no
allowable SO2 for gaseous fuel, 1.2 lb/106 Btu
for coal) so that any sulfur removal methods
which will meet EPA regulations could be
suitable from the turbine viewpoint.
A gas turbine can handle a wide latitude of
fuel heating values, ranging from blast fur-
nace gas ( ~ 100 Btu/ft3) to propane ( ~ 2000
IV-6-4
Btu/ft3). Thus, the chemical heating value,
per se, is not a problem. There is, however, a
heating value dependent problem which must
be considered. This is the problem of fuel
delivery pressure. The first aspect of this
problem is essentially one of hardware. The
sizing of fuel manifolding, injection nozzles,
etc., is a function of fuel heating value, i.e., a
given Btu/min must be supplied to the engine
and a fuel with a heating value of 150 Btu/ft3
requires a higher volume throughput than one
with 1000 Btu/ft3. Thus, to reduce the fuel
handling equipment to sizes compatible with
high release gas turbine combustors, the low-
heating-value fuels should be supplied at a
pressure higher than the pressure ratio of the
engine. The relationship between the fuel
delivery pressure and the Btu/ft3 requirement
differs among engine types but is of the form:
P = (f) LHV/(Specific gravity x delivery
temperature)1/2. Unfortunately the func-
tional form is inverse; thus, as the heating
value decreases or temperature increases, the
fuel delivery pressure increases. This has a
significant effect on overall system per-
formance and will be discussed later in the
paragraphs dealing with the integrated station
performance.
The most important characteristic of a
gasifier designated for use with a power system
is its efficiency in converting coal Btu's to fuel
gas Btu's. There are two gasifier efficiencies
that need to be considered. The first is the cold
gas efficiency (chemical heating value of fuel
gas/chemical heating value of coal) and the
second is the hot gas efficiency (chemical &
sensible heat in fuel gas/chemical heating
value of coal). The manner in which these two
efficiencies affect the integrated system
performance is complex, but if one thinks
again in terms of Carnot's law, all of the
chemical heating value (represented by the
cold gas efficiency) is used at combined-cycle
efficiency, while the sensible heat may or may
not be so utilized. If the hot gas can be used in
the engine, there would be no degradation in
the performance. However, if the gas must be
cooled by raising steam, then the sensible heat
-------
could be used only at steam cycle efficiency.
There is of course, a "however", attached to
the foregoing. If the sensible heat were to be
used in regenerating the gasification system, it
would be used at essentially combined cycle
efficiency. This use of the fuel gas sensible
heat requires relatively expensive heat ex-
change equipment.
Another gasifier attribute necessary for use
with power plants is operational flexibility.
While power systems designed for base-load
operation do not require fast startup
capability, they do operate over a range of
power settings ranging from perhaps 70
percent to full power. Thus, even base-load
applications require some turndown
capability. One of the more attractive features
of the COGAS system is its capability of rela-
tively fast startup, 5 minutes to full power for
the gas turbine power and less than 1 hour for
the total plant. This capability lends itself to
applications in mid-range load factor (3000 to
6000 hr/yr) in which daily startup would be
common place, with shutdown over weekends
or holidays. A gasifier for use in this system
would have to have a fast-start capability.
Since this type of power system typically
operates over a range of power settings from
40 percent to full power, flexibility in gasifier
operation would be necessary.
INTEGRATED FUEL
POWER SYSTEM
PROCESSING/
One of the basic tenets of mathematics is
that the whole is equal to the sum of its parts.
The power system has, in a sense,
circumvented this by putting together two
parts of comparable efficiency into a
combined system of significantly better per-
formance. It would be indeed fortunate if this
symbiotic relationship could extend to the
joining together of the fuel processing and
power cycle portions into an integrated
system. However, the serendipity does not
carry over; in fact, the requirements at the
interface of the two parts can cause a noticable
reduction in combined performance. There is
also an environmental consideration involving
the production of NOX which could influence
the selection of overall systems configurations.
One of the simplest of all integrated
systems is shown in Figure 5. Air for the
gasifier is bled from the compressor, raised to
the required gasifier pressure in a booster
compressor, mixed with the fuel in the gasifier
with the resultant hot fuel gas being supplied
directly to the turbine. Some heat exchange
between the air streams would be possible.
This system utilizes the fuel sensible heat in
the engine.
A much more complex configuration
(Figure 6) results if the fuel sensible heat must
be recovered for use elsewhere in the cycle. In
Figure 6, the gasifier is run at essentially
atmospheric pressure with the fuel gas exiting
to a boiler/superheater. From there the fuel
gas passes through a heat exchanger where
bleed air from the compressor enroute to the
gasifier is heated, then through a feed water
heater and finally into the booster compressor
were it is raised to the required pressure for
injection into the gas turbine burner. The
booster compressor could be driven, in part,
by an expansion turbine in which the heated
bleed air is let down to gasifier pressure. The
feedwater and superheated steam would be
utilized in the waste heat boiler.
What are the performance differences of
these two systems? The simple systems utilize
all the fuel heating value at combined-cycle
efficiency, while the second utilizes only the
chemical heating value at combined
efficiencies with the sensible heat being used
at steam cycle efficiencies. The efficiency
differences can be found by inspecting Figure
7, based upon data from Reference 4, in which
the percentage change in efficiency with fuel
temperature is shown for three different
turbine inlet temperatures. From all indica-
tions, the system utilizing the sensible heat in
the gas turbine is the most efficient. This
reinforces the results shown in Figure 2, which
indicated that large gas turbine participation
was more efficient.
IV-6-5
-------
There are, however, several considerations
that must be made before a final choice can be
made. Besides the very difficult hardware pro-
blems associated with fuel control systems
handling gases at 35 to 50 atm and tempera-
tures above 2000°F, there are the environ-
mental constraints to be considered. The
systems pictured in Figures 5 and 6 are based
upon both sulfur and particulate removal
within the gasifier; i.e., a method of desulfuri-
zation and particulate removal that operates
at 1500°F and above. Currently, the majority
of sulfur removal systems operate at 600° F or
below, some even requiring below zero gas
temperatures. With the exception of fluidized-
bed gasifiers and perhaps one or two other
types, the majority of gasifiers require external
desulfurization and thus require some method
of heat recovery from the fuel gas stream.
A second environmental constraint is that
of NOX formation. It is well established that
the formation of NOx is strongly dependent
upon the combustion temperature. Combus-
tion temperature, in turn, is a function of both
fuel heating value and of fuel and air preheat.
This dependency is shown in Figure 8 where it
can be seen that the combustion temperature
is a stronger function of sensible heat than of
HHV. Using Figure 9, which shows NOX con-
centrations as a function of temperature, the
rise in combustion temperature due to fuel gas
sensible heat will increase; in the extreme, the
emission of NOx by a factor of about 25 (i.e.,
from less than 5 to nearly 100 ppm with a fuel
gas having an HHV of 120 Btu/ft3). As the
fuel HHV increases, the base NOX emission
factor increases and the multiplying factor due
to sensible heat decreases — slowly at first and
then rapidly as NO equilibrium is approached.
The allowable concentration, using EPA regu-
lations for NOX from coal-fired power-plants
would be in the order of 120 to 160 ppm,
depending on engine efficiency. (If the system
were to be considered as a gas-fired station,
the allowable concentrations could be 35 to 45
ppm. A brief discussion of emission regu-
lations and their form is given in Appendix A.)
The situation;'therefore, is that as the HHV
of the product gas increases, the sensible heat
must be decreased (or vice-versa) in order to
meet the NOX standards. Since the NOX pro-
duction increases more rapidly with sensible
heat, it would seem to be more advantageous
to have a gasifier with a high cold gas
efficiency (more HHV in the gas) as was men-
tioned before. This is doubly beneficial since
the HHV of the fuel also has a noticeable
effect on the integrated system efficiency
(Figure 10). (This effect is somewhat
exaggerated in Figure 10 since the steam
system efficiency used in preparing this figure
was relatively low, i.e., ~30 percent.) This is
because of two effects, a reduction in steam
generation due to the decrease in mass flow of
fuel, and a concurrent decrease in booster
compressor work.
There are several ways to increase the HHV
of the product gas. Two of the more promising
are: (1) regeneration of the gasifier air, and (2)
oxygen enrichment of gasifier air. In the first,
the air to fuel ratio needed to attain a given
gasifier temperature is reduced as the air is
preheated. This means less nitrogen dilution
and more Btu/ft3. Oxygen enrichment
accomplishes the same thing; i.e., a reduction
in nitrogen dilution thereby increasing HHV.
In fact, the use of oxygen alone to blow the
gasifier would result in the production of
synthesis gas having an HHV of about 315
Btu/ft3. Unfortunately, the combustion of
synthesis gas could result in greater NOX
production than the combustion of methane.
There are, or course, modifications which
can be made to the combustion process which
could effect a reduction in NOX emission. One
of these, off-stoichiometric combustion
resembles, in theory, the method being used
with some success in gas-fired steam boilers.
At this time, there is no assurance that the
combustion efficiency during this staged-
combustion will be comparable to that
currently obtained in gas turbines, e.g.,
greater than 99 percent. This is especially true
with the low-Btu fuels whose combustion
characteristics in gas turbine burners have not
been extensively studied.
IV-6-6
-------
A SUMMING UP
The foregoing discussions have briefly
described the advanced power system which
offers the potential of attaining efficiencies
nearly 50 percent greater than those currently
obtainable yet utilizes equipment which would
be evolutionary developments of the com-
bined-cycle systems that are currently being
placed on-line by utilities. When used in con-
junction with coal/residual oil gasification,
such systems could generate electrical power
with a minimum of pollution while using rela-
tively abundant high-sulfur fuels. Although
economics have not been treated thus far,
prior studies have indicated that the combined
fuel gasification/advanced power system could
generate this pollution-free power at costs less
than conventional steam power plants having
equivalent emission characteristics and at
costs which could easily be less than those
associated with nuclear power (Figure 11 from
Reference 5).
That this answer to a maiden's prayer is not
without problems is apparent even to the
casual reader of this paper. Some of these
problems have been touched on — config-
uration of the power cycle, pressurization level
of the gasifier, trade-offs between sensible and
chemical heating value, etc. A myriad of
practical problems exist in interfacing the two
complex systems.
However, many of these problem areas can
be resolved only by construction and operation
of a prototype plant of large enough scale to
demonstrate the concept that clean power can
be generated from high-sulfur fuels with
acceptable economics. Until a successful
demonstration(s) takes place, utilities will view
the concept as just that—a concept—and will
continue to displace fossil-fired systems with a
greater dependence on nuclear power. The
recent announcements by the Office of Coal
Research describing its plans for one or more
demonstration plants gives hope that a
reasonable program will soon be underway —
one which will lead to the introduction of com-
mercial systems by the latter part of this
decade.
REFERENCES
1. Giramonti, A. J. Discussions of Steam and
COGAS Systems with the Babcock and
W^lco'x Co., Barberton, Ohio. UARL
Report UAR-H246, 1969.
2. Packaging Sells the Combined Cycle.
Electrical World. September 1,1972.
3. P&WA Specification 526 - Gaseous Fuel.
Industrial for Turbine Engine. June 1968.
4. Giramonti, A. J. Advanced Power Cycles
for Connecticut Utility Systems, UARL
Report L-971090-2, January 1972.
5. Robson, F. L. Clean Power from Gas
Turbine-Based Utility Systems.
Combustion. July 1972.
APPENDIX A
Emission Standards
Current emission standards are based upon
the amount and type of fuel burned; x
pounds/106 Btu for coal, y lb/106 Btu for oil,
and z lb/106 Btu for gas. If the full advantage
is to be taken of the advanced power systems,
a new basis for standards must be used.
First, as was alluded to in the main text,
fuel type should be defined in a different
manner. The power system burns a gaseous
fuel even though coal, residual oil, coke, or
garbage is used in the gasifier. In fact, in an
actual system it would be hoped that the input
to the gasifier could be switched more or less
as the fuel market dictates.
Secondly, the current standards are based
upon input rather than the output, which is
the real purpose of the power system. It is now
possible for a power station to meet the
regulations but emit, in absolute numbers,
significantly more pollutants than would an
advanced power system of equal power. For
example, a conventional steam station of
1,000-MW output running at off-design
conditions to meet NOX regulations could
have a heat rate of perhaps 12,000 Btu/kWhr
and, meeting EPA standards, would put out
IV-6-7
-------
2,400 Ib/hr or 2.4 Ib/MWhr of NOX. An
advanced system having a heat rate of 8,000
Btu/kWhr and meeting the same standards as
currently written would emit, for the same
power demand, l,6001b/hr or 1.6 Ib/MWhr of
NOx.
As turbine inlet temperatures increase,
there is a second-order increase in NOX
emissions. While minor, this increase could
result in an emission/106 Btu above the
prescribed level. However, the power system
has become more efficient, and there could
well be a decrease in emissions/output
compared to the lower temperature system.
Referring to the above example, suppose a
500°F increase in turbine inlet temperature
caused a 10 percent increase in NOX emissions
but also a 20 percent decrease in heat rate.
The system would no longer meet the current
EPA regulations but would, in fact, emit only
1.4 Ib/MWhr.
For these reasons it appears that standards
based upon emissions/MWhr would be a more
reasonable choice.
IV-6-8
-------
5
o
5
I
CARNOT EFFICIENCY
POTASSIUM-STEAM
BINARY SYSTEM
(NUCLEAR FUELED)
TYPICAL PWR
SYSTEM
CLOSED-CYCLE
HELIUM
TURBINES
TYPICAL
SYSTEM
rnr.. COMBINED GAS TURBINE
cuwta AND STEAM TURBINE
NUCLEAR
-TEAM
CURRENT DESIGN
30
20
2000 3000
MAXIMUM CYCLE TEMPERATURE, °F
Figure 1, Comparison of estimated thermal efficiencies for advanced-cycle power stations.
4000
IV-6-9
-------
COAL OR RESIDUAL OIL
POWER
TURBINE
ELECTRIC
GENERATOR
ELECTRIC
GENERATOR
PUMP
Figure 2. Combined gas-steam turbine system.
IV-6-10
-------
COMPRESSOR
TURBINE
POWER
TURBINE
ELECTRIC
GENERATOR
ELECTRIC
GENERATOR
PUMP
Figure 3. Supercharged combined gas and steam turbine system.
IV-6-11
-------
70
60
0>
i*
as
£5 50
Oil
Ul
40
30
STEAM CYCLE EFFICIENCY 38.8%
TURBINE INLET
TEMPERATURE, °F
3100
2800
2200
20
80
40 60
GAS TURBINE OUTPUT, % of station output
Figure 4. Performance of exhaust-fired combined system.
100
IV-6-12
-------
FUEL GAS
BOOSTER COMPRESSOR
STACK
Figure 5. Schematic of high-pressure system.
IV-6-13
-------
FUEL GAS
FROM
CONDENSER
FROM TURBINE
WASTE • HEAT
BOILER
TO FUEL GAS
SUPER HEATER
FROM FUEL GAS
SUPER HEATER
TO FUEL GAS
WASTE - HEAT BOILER
Figure 6. Schematic of low-pressure system.
IV-6-14
-------
o
TURBINE INLET
TEMPERATURE,0 F
COMPRESSOR PRES
SURE RATIO, atm
500
1000 1500
FUEL TEMPERATURE TO TURBINE,°F
2000
2500
Figure 7. Effect of fuel temperature on integrated station efficiency.
IV-6-15
-------
4400
4200
M»
& 4000
3800
3600
3400
INCREASE FUEL
TEMPERATURE
INCREASE FUEL HHV
REFERENCE FUEL HHV-120 Btu/scf
REFERENCE FUEL TEMPERATURE=80°F
STOICHIOMETRIC FUEL-AIR RATIO
INITIAL AIR TEMPERATURE=825"F
100
120 140 160
FUEL CHEMICAL PLUS SENSIBLE HEAT, Btu/scf
180
200
Figure 8. Effect of fuel gas chemical and sensible heat on combustion temperature.
IV-6-16
-------
1000
i
100
.0
§
0.1
i i i
o LOW-HEATING-VALUE FUELS
0 METHANE
_* JP-5
BASED ON P & WA THREE-ZONE BURNER MODEL
TURBINE INLET TEMPERATURE=1760°F
A I I I I I
3200 3400 3600 3800 4000 4200 4400 4600
MAXIMUM COMBUSTION TEMPERATURE,°R
Figure 9. Effect of combustion temperature
on nitric oxide emissions.
100 110 120 130 140 ISO 160 170
FUEL HIGHER HEATING VALUE, Btu/scf
Figure 10. Effect of fuel HHV on integrated
station efficiency.
IV-6-17
-------
1.50
125
P
&
o
^ i on
2» ItUU
o±
2
3
S °-75
>
I—
0.50
0.25
0
1. CONVENTIONAL STEAM-ELECTRIC STATION BURNING UNTREATED COAL
2. CONVENTIONAL STEAM-ELECTRIC STATION BURNING UNTREATED COAL
WITH 85% EFFECTIVE SULFUR OXIDE STACK GAS CLEANING
3. COGAS STATION BURNING DESULFURIZED FUEL
- 4. BASE-LOAD GAS TURBINE STATION BURNING DESULFURIZED FUEL
5. NUCLEAR
—
—
1
2
•MM
WMMI
4
•BMMDB
M1MMB
1
1
^^^m
2
•MMB
3
4
MMBMi
5
4
'
MMM
2
.
^BHMI
3
4
5
-
-
-
1970 DECADE
1980 DECADE
LEVEL OF TECHNOLOGY
1990 DECADE
Figure 11. Busbar power costs for coal-based power plants.
IV-6-18
-------
7. A DESIGN BASIS FOR UTILITY GAS FROM COAL
C. W. MATTHEWS
Institute of Gas Technology
INTRODUCTION
The preferred solution for control of
atmospheric pollutants from the stacks of
electric utility boilers is to substitute clean fuel
for polluting fuel. In many cases a clean fossil
fuel such as natural gas is not a practical sub-
stitute for coal in the generation of electricity
because of scarcity and cost. Furthermore, the
available supply of clean fuel may combat pol-
lution more effectively when used to fulfill
residential and small commercial needs.
The combustion products of coal con-
tribute one-eighth of the total atmospheric
pollutants emitted in the United States,
including one-half of the sulfur oxides and
one-quarter of both the nitrogen oxides and of
the particulates. Sulfur emissions from coal
combustion may be reduced by: (1) using low-
sulfur coal, (2) cleaning high-sulfur coal by
physical methods, (3) removing sulfur oxides
from coal combustion gases, (4) removing sul-
fur during the combustion step, (5) producing
de-ashed low-sulfur fuel by solvent processing
of coal, and (6) gasifying coal and removing
sulfur from the gas before combustion.
The last method, coal gasification with gas
cleaning before combustion, promises the
greatest reduction in sulfur emission. Most of
the sulfur gasified appears as hydrogen sul-
fide. Several different commercial gas
cleaning processes are available today which
can reduce the hydrogen sulfide content of gas
streams to less than 10 ppm; some processes
can remove hydrogen sulfide to 1 ppm or less.
This paper discusses the selection of design
criteria for the gasification of coal and
cleaning of the generated gas before com-
bustion in an electric utility boiler. The pre-
liminary plant design description is for a large
pilot plant installation that will demonstrate
the feasibility of this concept.
The gas produced from coal for the boiler is
called utility gas, although producer gas and
low-Btu gas are equivalent names. It is made
by gasifying coal with air and steam at
elevated pressure. Dust and sulfur compounds
are removed from the utility gas before it is
burned in the power generating system.
Heating value of the utility gas will be between
140 and 250 Btu/ft3, depending on the gasifier
and the plant design.
The low heating value of utility gas limits
the distance it can be transported
economically. When used to fuel electric
power stations, it probably will be generated
onsite. Retrofitting coal gasifiers to existing
boilers is one of the most important appli-
cations now .for coal gasification plants. In the
future installations, combined gas turbine-
steam turbine systems will be served by gasifi-
cation plants not only for reduced atmospheric
pollution, but also for greater efficiency in the
generation of electricity. From these plants we
expect savings in investment and decreased
electricity costs as well as less heat rejection to
the environment and better conservation of
our coal resources.
IV-7-1
-------
We believe that this country has an urgent
interest in the demonstration of the practica-
bility of coal conversion to clean utility gas for
the electric power industry as soon as possible.
Successful achievement of this goal in the
shortest possible time through government
and industry support will provide substantial
benefits to the country and to the electric
utility industry.
The program proposed by the Institute of
Gas Technology 0GT) for proving this concept
includes construction in the near future of a
large pilot plant which will be located near an
existing boiler. The pilot plant will be capable
of feeding from 10 to 50 ton/hr of coal and will
fuel a power plant with a generation capacity
of 20 MW or more. We believe that with
favorable results from this pilot plant com-
mercial plant designs can be undertaken by
the end of 1975.
IV-7-2
-------
GENERAL DESIGN PRINCIPLES
The entire concept hinges on the coal
gasifier performance. The gasifier operation
must be reliable; it must gasify a high per-
centage of feed carbon; and it must be load-
following, that is, capable of operating hi
response to the power system requirements.
The focal point of our effort is to demonstrate
gasifier operation.
We anticipate that coal conversion to utility
gas will be practiced to a great extent in the
coal belt from Illinois to New York. Because
the coals from this region have caking
properties, the plant must be capable of
accepting these coals as feed.
Initially, utility gas will be produced for
existing boiler systems. Operation of the large
pilot plant facility will show the practicality of
retrofitting coal gasifiers to existing boilers.
This is one of the most important applications
now for coal gasification plants. In the future,
for savings in investment and for decreased
electricity costs, combined-cycle systems will
be served by gasification plants. Flexibility will
be provided in the pilot plant design to test
advanced power generating system
components.
The pilot plant gasifier will be large enough
so that scale-up to commercial-size gasifiers in
the future can be done with confidence. For
example, a single gasifier fed at a rate of 10
ton/hr and operating at 300 psia will fuel a
power station generating about 20 MW of
power. Operation at 1000 psia will increase its
capacity by more than 3 times. The diameter
of this gasifier permits shop fabrication and
rail shipment for reduced construction time
and cost.
Our guiding principles in design of the first
large pilot plant for conversion of coal to
utility gas are:
1. Prove the gasifier design and operation.
2. Accept caking coal as feed.
3. Demonstrate application of coal gasifica-
tion to existing boilers.
4. Build gasifier large enough so that it can be
directly scaled to commercial size.
UTILITY GAS FROM COAL PLANT
Figure 1 is a block flow diagram for our
proposed utility gas from coal plant. This
design is suitable when the gasification plant
is installed to fuel a small to medium size
boiler. I will briefly describe the flow scheme
in Figure 1 and then discuss in more detail the
important parts of the plant.
The coal feed is crushed to the desired size.
Lock hoppers are used to transfer coal from
atmospheric pressure to the elevated pressure
of the gasifier. Heat is recovered from the hot
raw gases, and a small part of the cooled gas is
used to pressurize the lock hoppers. The main
gas stream expands from high to low pressure
through a gas expander, thereby generating
power needed to drive the large gasifier air
compressor. Gas vented from the lock hopper
system rejoins the low-pressure main gas
stream. The combined gas streams are cleaned
of sulfur at low pressure by the Stretford
process or one that is similar. The Stretford
process produces elemental sulfur directly
from hydrogen sulfide. After sulfur removal,
the gas flows to the boiler. Associated with this
process train are a large air compressor, ash
handling and disposal equipment, waste-water
treatment, and possibly oil stabilization and
storage equipment. When an efficient, high-
temperature sulfur removal system has been
developed, we believe that it will replace the
equipment shown within the dashed lines.
DESIGN PRESSURE
In plants manufacturing pipeline gas from
coal, maximum methane formation within the
gasifier is desired for improved thermal
efficiency. To obtain this, gasifier pressures of
1000 psia or more are preferred. The thermal
efficiency of utility gas plants does not depend
on methane formation within the gasifier, and,
therefore, we have more flexibility in selection
of plant pressure. If necessary — and this is
what we have done — the plant pressure is
established on other considerations than the
chemistry of gasification.
Lock hoppers were selected to transfer coal
into the plant. A dry solid feed permits a less
complicated gasifier design, and the plant fol-
lowing the gasifier is also less complicated. For
high plant reliability when using dry feed
IV-7-3
-------
systems, the performance of lock hopper
valves will probably set the upper pressure
limit. Today, the best commercial lock hopper
operation is that demonstrated with Lurgi
gasifiers. In these, the pressure difference
between lock hopper and gasifier is about 300
to 350 psi. Even though we want to gasify at
pressures up to 1000 psi, we will design for a
pressure of about 300 psi because lock hopper
valves are available which work at this pres-
sure. We intend to search for improved
methods to feed dry coal into higher pressure
systems.
COAL FEEDING SYSTEM
Figure 2 presents a simplified illustration
of the solids handling system, which includes
the coal feed system, pretreatment, gasifica-
tion, ash removal, and dust removal from the
gas.
As described above, the single-stage lock
hopper was selected to transfer coal from
atmospheric pressure to the elevated pressure
of the gasifier. This feed system was chosen so
that the gasification plant will be simple,
reliable, and less costly.
The disadvantages of the lock hopper are
evident. The most important is the difficulty in
obtaining reliable operation of the lock hopper
valves. During the hopper cycle, these valves
alternately seal against the gasifier pressure,
and then open to pass a fine, dry, abrasive
solid. Coal dust tends to pack in the valve, cut
the packing, and jam in the guides and seat.
The valve must seal fairly well. Leakage
overheats the valve and lock hopper;
introduces dirty, raw gases into the hopper
and its vent gas system; and wets the cool coal
with moisture, oil, and tar causing the coal to
bridge and no longer flow easily.
A second disadvantage is the need to vent
gas from the hopper during its operating cycle.
Loss of this vent gas decreases process
efficiency. Recompressing the gas into the
system is expensive. In this utility gas plant
design, the vented gas is collected and mixed
with the low-pressure main gas stream before
sulfur removal. In case the valves leak, the
vent gas system includes a cooler and vapor-
liquid separator to reduce contamination of
product gas.
The IGT HYGAS process is designed to
manufacture pipeline-quality gas from coal.
In this process coal is fed to the gasifier in the
form of a slurry which is pumped to system
pressure. This is a more reliable feeding
system than high-pressure lock hoppers, and
the slurry pumps are capable of good
operation to 1000 psi and higher discharge.
You might ask: Why not use a slurry feed
system for utility gas generation?
Slurry feeding introduces additional
complexity in the plant. Equipment for
making the slurry must be provided. When the
slurry enters the gasifier, the liquid must be
vaporized. Therefore a drying bed is added to
the gasifier, and the hot, raw gasifier gases are
used to supply the heat needed for drying and
for stripping. The lowered gas temperature
makes efficient recovery of heat from the gas
difficult. Stripped slurry oil must be recovered
efficiently for recycling. In its recovery, more
than one stage of quenching may be needed;
the gas stream is cooled to 100°F, thus adding
to the cooling water demand; and activated
carbon towers or sponge oil scrubbing
completes final oil removal from the gas.
Then, the recovered oil has to be dewatered
and stripped of dissolved gas before returning
to the slurry tanks.
We concluded that utility gas plants for the
electric power industry will be less complex,
less costly, easier to operate, and more
efficient when using lock hoppers to feed coal
into the gasifier.
COAL PRETREATMENT
Most bituminous coals have the property of
caking or agglomerating when heated.
Agglomeration of coal within the gasifier
cannot be tolerated because of the possibility
of pugging. So that the utility gas process can
accept the widest variety of coal as feed,
facilities for modifying or destroying this
property must be a part of the process.
Pretreatment takes place in the presence of
air at 750 to 800 °F. The particle surfaces of
coal are mildly oxidized, destroying the caking
properties. Heat is evolved and must be
removed to control the temperature.
Pretreated char yield is about 90 percent of the
coal feed weight. Off-gas from pretreating
FV-7-4
-------
contains tars, tar acids, carbon oxides, sulfur
dioxide, water vapor, and sometimes small
amounts of oxygen.
In the utility gas process, we propose to
pretreat coal at gasifier pressure and to
comingle pretreater off-gas with gases from
the gasifier. Hot pretreater char is fed directly
into the gasifier, thus avoiding thermal losses.
This design eliminates the waste-water and
gas-treatment problems associated with low-
pressure pretreatment. Furthermore, a smal-
ler pretreater vessel is needed for pressure
operation. The overall plant complexity is
reduced and so is its capital cost.
The heat of reaction in the pretreater is
removed by generating steam in heat
exchanger tubes contained in the fluidized
bed. The amount of steam generated is more
than the gasifier steam requirement.
THE GASIFIER
We want to obtain rapid gasification rates
which will permit a higher throughput for a
given reactor size and, therefore, will result in
a less costly plant. Rapid, precise control of
the gasifier operation is needed to follow
changes in the power demand. Production of a
clean, low-carbon ash is a primary economic
consideration.
The gasifier is designed to gasify coal with
air and steam in a fluidized bed.
Simultaneously, the coal ash is agglomerated
into larger and heavier particles for selective
separation from the bed. The principle of ash
agglomeration and separation was discovered
by A. Godel1 and developed into the Ignifluid
boiler. The concept was described by Jequier
et al.2-3 following laboratory and pilot plant
gasifier development at Centre d'Etudes et
Recherches des Charbonnages de France. We
have adapted and modified the Jequier design
in development of this reactor concept.
We call the gasifier the ABR
(Agglomerating Bed Reactor). The ABR
concept resolves the main disadvantage of coal
gasification in a fluidized bed rich in carbon:
How can low-carbon-content ash be selectively
removed from the bed? Advantages of fluid-
bed gasification are retained. These are:
1. Bed temperature can be uniformly and
;.,., readily controlled.
2. High reaction rates can be attained
because of excellent gas-solids contact and
large surface area of the solids.
3. Coal fines from mining and crushing can
be used in the feed.
4. The mass of carbon in the fluid bed
ensures reducing conditions at all times.
The ABR is fluidized by a mixture of air
and steam. Gasification takes place at about
1900°F in the fluidized bed. Part of the
fluid izing gas enters through a grid which is
sloped toward one or more cones contained
in the grid. Heavier particles migrate along the
sloped grid toward the cones. The rest of the
fluid izing gas flows upward at high velocity
through the throat at the cone apex, creating a
submerged jet within the cone. The tempera-
tures generated within the jet are somewhat
greater than in the rest of the bed. As carbon
is gasified in and near the jet, ash is heated to
its softening point. The sticky ash surfaces
cling to one another, and ash agglomerates
grow in the violently agitated jet. When heavy
enough, the agglomerates fall counter to the
high-velocity gas in the throat and are thus
separated from the fluid bed.
To protect the ash lock hoppers from the
hot agglomerates, they are filled with Water
which is boiling from the heat contained in the
ash. The steam generated reduces the amount
of external steam needed for the ABR. When
filled, the hopper is flushed into filters to
recover a wet cake of ash for final disposal.
Filtered water returns to the lock hoppers.
TAR AND DUST REMOVAL
Above the ABR fluid bed we have designed
for a gas residence time of 10 to 15 seconds;
the gas temperature will be between 1500 and
1900°F. By "soaking" the gas at high temper-
ature, tars and oils which may be evolved are
thermally cracked to gas and carbon.
Elimination or reduction of tarssand oils in the
raw gas will reduce heat exchanger fouling
and will simplify by-product and
waste-stream cleanup and treatment.
IV-7-5
-------
Most of the dust contained in the gasifier
gases is removed by cyclone separators and
returned directly to the ABR bed. Very fine
dust is separated in the second stage of dust
removal and is returned to the gasifier by
injection beneath the gasifier cones. Within
the cones the carbon contained in the fine dust
is gasified. The fine ash sticks to the heavy
agglomerates and is removed from the system.
Although cyclone separators are shown in
Figure 2, we plan to investigate other high-
temperature solids separators. We will provide
space and plant flexibility in the utility gas
pilot plant for large-scale testing of alternative
separators. Efficient removal of hot dust is
important in retrofitting utility gasification
plants to existing boilers to prevent erosion,
contamination, and plugging in the raw gas
heat exchangers. When utility gasification is
applied to combined-cycle power generation,
even greater gas cleanliness is needed to
protect the gas turbines.
SULFUR REMOVAL SYSTEM
Most of the sulfur produced by coal gasifi-
cation appears in the form of hydrogen
sulfide. Because the pilot plant produces low-
pressure, low-temperature fuel for a boiler, we
can use the Stretford or a similar process for
product gas sulfur cleanup. The Stretford
process is commercial; it is effective when
scrubbing low-pressure gas; it can produce a
cleaned gas containing as little as 1 ppm
hydrogen sulfide; and the process converts
hydrogen sulfide directly to elemental sulfur,
avoiding the need for a Claus plant. For this
application the Stretford process is easy to
operate and is inexpensive.
Figure 3 is a simplified flow diagram for
the Stretford process. The scrubbing liquor is
an aqueous solution of sodium carbonate,
sodium vanadate, and ADA (the sodium salt
of anthraquinone 2:7 disulfonic acid). Gas
enters the scrubbing tower at less than 140°F
and usually at a pressure of less than 75 psia.
We have selected 120°F and 25 psia as pilot
plant conditions. Absorbed hydrogen sulfide is
oxidized by the solution to fine, suspended
sulfur particles. After completion of the
oxidation reaction, the reduced solution is re-
oxidized by blowing with air at atmospheric
pressure. The fine sulfur concentrates in the
froth during air blowing and is collected from
the solution as an elemental sulfur product.
The preferred system for sulfur removal
may change depending on the gasification
plant capacity and the final use for the gas.
For large gasification plants sulfur removal at
high pressure using processes such as Selexol,
Purisol, or Alkazid may be more economical
than low-pressure sulfur removal. In
conjunction with combined-cycle plants, yet-
to-be-developed, high-temperature sulfur
removal is desirable for improved plant
efficiency and for decreased cost. For small-to-
medium-sized power plants backfitted with
coal gasification systems, we believe that the
Stretford process will be widely applied for
sulfur removal from the gas.
ENERGY RECOVERY
Raw gas leaves the gasifier at a
temperature between 1500 and 1700°F and at
a pressure of 300 psi or higher. The main gas
flow from this section of plant enters the sulfur
removal system at 25 psia and 120°F. Sensible
heat contained in the gas represents about 20
percent of the heat available from combustion
of the coal feed. The energy recovery section
(Figure 4) is designed to recover as much of
this energy as possible. Since most of the heat
recovered must be used in the power cycle, the
design of this section of the plant will be
strongly influenced by the heat levels that can
be used in the power cycle.
After withdrawing from 1 to 5 percent of
the main gas stream for use as lock hopper
pressurizing gas, the pressure of the main gas
stream is broken by expansion through a gas
expander. We want the expander exhaust gas
condition to be suitable for feeding directly to
the Stretford scrubbing tower. In our design,
the condition is at 25 psia and 120°F with the
gas water saturated. Condensation should not
IV-7-6
-------
occur in the expander. Having defined the gas
condition at the expander exhaust, the desired
moisture content of the gas is obtained by ad-
justment of the main gas separator
temperature. With a 300-psia gasifier, this
temperature is 224°F; with a 1000-psia
gasifier, it is 300°F. The gas expander inlet
temperature is adjusted to give the desired
exhaust temperature. Power recovered by gas
expansion is used to drive the gasifier air
compressor. The air compressor power
requirement is about 19 bhp/ton-day of coal
feed in a 300-psi plant and 26 bhp/ton-day in
a 1000-psi plant.
Condensed water from the main gas
separator and from the lock hopper gas
separator is fed into a low-pressure stripping
tower to remove dissolved gas. The stripped
gas rejoins the main gas flow entering the
Stretford scrubbing tower. Stripped waste
water is cooled and sent to either biological
treatment or active carbon treatment for
phenol removal.
The advantages of the proposed design are:
(1) the gasifier air compressor is driven by
process energy, (2) the main gas stream is not
water-cooled to obtain the desired 120°F, (3) a
minimum of waste water is produced, and (4)
the conditions of heat recovery to the power
cycle are well defined so that an efficient
recovery system can be designed.
COOLING WATER
TREATMENT
AND WASTE
Cooling water requirements are minimal in
this utility gas plant design as a result of some
of the process choices that were made. These
are: (1) use of a dry feed system, (2) pretreat-
ment at pressure, (3) gas expander exhausting
at 120°F, (4) gasifiying at higher pressure, and
(5) use of the Stretford process for sulfur
removal. In addition, air cooling is used where
feasible.
Waste-stream treatment problems
associated with coal gasification plants can be
serious. We reduced the severity of these
problems by choosing the dry feed system,
pressure pretreatment, single-stage high-tem-
perature gasification with ash agglomeration,
raw gas "soaking" at high temperature, and
the Stretford process.
SYSTEM PERFORMANCE
Table 1 shows the calculated plant
performance for 300-psi and 1000-psi utility
gasification plants. Note that although more
methane is formed in the 1000-psi gasifier, the
product gas heating value is not greatly
different for the two cases. Higher-heating-
value gas can be produced if the gasifier is
designed as a two- or three-stage unit with
gasification temperatures increasing progres-
sively in each stage. We do not believe that this
increased complication is warranted. Also,
higher-heating-value gas can be made if the
pretreating unit off-gas is diverted from the
gasifier. This will involve additional plant
equipment for condensing, separating, and
waste-stream treating of the off-gas and its
components. This, too, is considered to be an
undesirable plant addition.
Table 1. CALCULATED PLANT PERFORMANCE
Gasifier pressure, psia
Product gas
Heating value, Btu/scf
Composition, vol %
CO
C02
H2
H?0
CH4
N2
Total
Thermal efficiency, %
To all products
To gas only
To steam only
300
Wet
140
17.8
9.2
12.1
8.5
4.3
48.1
100.0
Dry
153
19.4
10.0
13.3
__
4.7
52.6
100.0
86.5
73.2
12.7
1000
Wet
150
12.5
13.6
11.6
8.5
7.1
46.7
100.0
Dry
164
13.7
14.8
12.6
__
7.8
51.1
100.0
88.2
73.9
12.6
The thermal efficiency of the plant is very
good in both cases, with the 1000-psi plant
being slightly higher. The thermal efficiency is
determined by comparing product heating
values, heat contained in net steam, etc., to the
heating value of the coal feed. Compare the 86
percent efficiency of the utility gas process
with the 66 percent efficiency of coal gasifica-
tion plants producing synthetic pipeline gas.
IV-7-7
-------
The gasification plant output should
respond at nearly the same rate as the power
plant responds to electrical demand changes.
The ABR gasifier design should be operated
nearly all of the time; its temperature should
not fluctuate drastically or serious internal
refractory damage may result. However, the
throughput can be cut back significantly, and
in this way the plant may serve as a load-
following plant in addition to its use in base-
load operation. If the plant is initially
operating at design capacity, the first move in
decreasing output is to decrease steam and air
flow to the gasifier. This flow change can
reduce plant output by a factor of 3 or 4 very
rapidly. If a further reduction is needed, the
ratio of air to steam is reduced, causing a slow
decline in gasifier temperature. By lowering
this temperature from 1900 to 1300 or 1400°F,
an additional reduction in output of 10 to 20 is
obtained. The gasifier temperature should be
held within a few hundred degrees of its
normal operating temperature most of the
time to avoid thermal shock and consequent
cracking and spalling of the gasifier internal
refractory. Therefore, if the boiler plant is shut
down temporarily, we prefer that the gasifier
operation continue at minimum rates and that
the produced gas is burned in a flare or
burning pit.
CONCLUSIONS
We have described the preliminary design
of a coal gasification plant for manufacturing
clean utility fuel gas for the electric power
industry. Its operation will demonstrate the
economics and reliability of such a plant when
used to fuel an existing boiler. Modifications
in this design will adapt the gasifier concept to
combined-cycle power generation and to the
manufacture of clean fuel for other industrial
uses.
We expect that the plant cost and the
product energy cost will be less by a significant
amount than those costs for an equivalent-size
synthetic pipeline gas pla,nt. Proof of this
design will provide electric utilities with a
realistic method for conversion of coal to a
clean fuel.
REFERENCES
1. God el, A. A New Combustion Technique.
Eng. Boiler House Rev. 71:145-153, May
1956.
2. Jequier, L., L. Longchambon, and G. Van
de Putte. The Gasification of Coal Fines. J.
Inst. Fuel. 33:584-591. 1960.
3. Jequier, L. et al. Apparatus for Dense-
Phase Fluidisation, U.S. Patent 2,906,608.
September 29, 1959.
IV-7-8
-------
LP
QUENCH
VENT GAS
PRESSURIZING GAS
COAL
CRUSHING
1
LOCK
HOPPERS
GASIFIER
PRODUCT GAS
SULFUR
GAS
CLEANING
(STRETFORD)
EXPANDER
~1
HEAT
RECOVERY
WASTE-
WATER
TREATMENT
OIL
FABILIZA
AND
STORAGE
TION
E
AIR
COMPRESSOR
I
\ (FUTURE)
REPLACE WITH HIGH-TEMPERATURE
PARTICULATE AND SULFUR
REMOVAL
ASH
HANDLING
AND
DISPOSAL
Figure 1, Coal gasification pilot plant block flow diagram for clean utility gas.
IV-7-9
-------
RAW GAS TO
TREATING
COAL REED
LOCK HOPPER
PRETREATMENT
FOR BITUMINOUS
COALS
STEAM
GENERATION
DUST \ /
REMOVAL W
2ND STAGE
DUST
REMOVAL
AIR AND STEAM
SOUDS FEEDER
AIR AND
STEAM
ASH LOCK HOPPER
(WATER - FILLED)
Figure 2. Agglomerating bed reactor.
IV-7-10
-------
CLEANED GAS.
GAS FEED.
SCRUBBING
TOWER
AIR
SULFUR
Figure 3. Stretford process.
IV-7-11
-------
RAW GAS FROM GASIFIER ^
1500 • 1700 °F, 1
HIGH PRESSURE /^
HEAT /
EXCHANGER V
t
HEAT
RECOVERY
TO
POWER
CYCLE ^.
\.
GAS
^XPANDER
DM ^^^^^^"
JL
S^\ GASIFIER
K A AIR
\^J/ COMPRESSOR
J _^
^~~y* s~
•M^ i^nni co
^\^ IrUULtK
MAIN
GAS
SEPARATOR
GAS-TO-SULFUR
25 psia,
WATER SATURATED
, GAS TO LOCK
l»*~ HOPPER, 120 °F
1 HIGH PRESSURE
TOR J
l^ WATER TO
^^ STRIPPING, 120 °F
WATER TO STRIPPING
200-300 °F '
Figure 4. Energy recovery section.
IV-7-12
-------
SESSION V:
Pilot Plant Design, Construction, and Operation
SESSION CHAIRMAN:
Mr. H.B. Locke, National Research Development
V-0-1
-------
1. THE DESIGN, CONSTRUCTION, AND OPERATION
OF THE ABINGDON FLUIDISED BED GASIFER
G. MOSS AND D. E. TISDALL
Esso Research Centre, England
ABSTRACT
A detailed description is given of the design and construction of the desulphurising fluidised-
bed gasifier which was built and operated at the Esso Research Centre at Abingdon under the
terms of GAP Contract CPA 70-46.
The unit was operated under gasifying conditions for a total of 450 hours during the commis-
sioning period. Information is presented concerning the operational problems. which were
encountered and the remedial steps which were taken.
INTRODUCTION
The information in this paper supplements combustion of fuel oil. The Abingdon gasifier
that given in reference 1, which provides pro- is the first of its kind and incorporates a num-
cess data relating to the retention of sulphur in ber of unique features which were designed
fluidised beds of lime during the in situ partial specifically for pilot scale operation.
V-l-1
-------
THE DESIGN AND CONSTRUCTION OF
THE GASIFIER
Three primary decisions determined the
size, configuration, and mode of operation of
the gasifier. These were as follows:
1. The internal configuration of the regener-
ator of the gasifier was to be identical
with that of the batch units which had
been used in the exploratory phase.
2. The gas produced was to be burned within
a standard packaged boiler fitted with a
suitably modified combustion system.
3. A monolithic form of construction was to
be used in which all vessels and transfer
lines were to be formed as cored holes in a
solid block of refractory concrete.
The first decision eliminated one area of
uncertainty because it was known from
experience that the batch units functioned
satisfactorily under regenerating conditions. It
also set a limit to the capacity of the unit based
on what was then known concerning the
capacity of the batch units. At the time that
the decision was made these had only been
operated at a superficial gas velocity of 4
ft/sec. At an assumed SO2 concentration of 10
percent by volume in the regenerator off-gas,
this gave a sulphur handling capacity of about
8.6 Ib/hr. In the case of a 2.2 percent by
weight sulphur fuel, oil this limited the fuel
throughput to 391 Ib/hr, giving an energy
throughput of 7.1 x 106 Btu gross/hr or 6.7 x
10 6 Btu net/hr. The gasification of 391 Ib fuel
oil/hr at 900 °C and a gas velocity of 4 ft/sec
with 20 percent of stoichiometric air indicated
a cross-sectional area for the gasifier of 4.8 ft2.
The decision to use a packaged boiler
rather than a flare for the second stage of
combustion was influenced by a number of
considerations. A flare might have had an
unfortunate impact on local public relations;
it was in any case considered desirable to use
the gasifier to fire a standard piece of equip-
ment in order to demonstrate it as a practical
proposition. The capacity of the boiler which
was available was 10 x 106 Btu/hr delivered as
pressurised hot water. This appeared to be
quite suitable for use as an energy sink for a 7
x 106 Btu/hr gasifier; it was decided to
dissipate the output to the atmosphere via a
pressurised heat exchanger and an atmos-
pheric evaporative cooler. What was not
realised at the time, was that it was possible to
operate the batch units at up to 8 ft/sec super-
ficial gas velocity. Consequently, although on
the original design basis the heat disposal
equipment provided a reasonable margin of
spare capacity, it subsequently turned out that
heat disposal was a factor restricting the range
of operating conditions which could be
explored.
The use of the packaged boiler enabled flue
gas recycle, to be used to control the gasifier
temperature to levels lower than those dictated
by adiabatic operating conditions. This was
advantageous since steam would most likely
have been used for this purpose in other
circumstances, involving the use of an addi-
tional utility.
So far as the gasifier was concerned, the
choice of monolithic construction imposed its
own logic upon the geometrical configuration.
The reasons for choosing monolithic
construction were: (1) the absence of joints
between the various components enabled them
to be grouped in a very compact arrangement;
(2) the position of the bed transfer ducts in the
heart of the block enabled the sensible heat of
the transferred bed material to be conserved to
the same degree as it would be in a large scale
unit; and (3) this was a very simple and cheap
form of construction.
The major disadvantage with monolithic
construction is that it does not readily lend
itself to modification. It was necessary to be
sure that transfer lines would work
satisfactorily before they were cast in refrac-
tory concrete. An additional disadvantage was
the reliance on the durability of the concrete
when subjected to temperature stresses. In
fact some minor cracks did develop but these
were found to be self-sealing under gasifying
conditions.
V-l-2
-------
When the choice of monolithic construc-
tion was made it was found expedient to adopt
a rectangular cross section, because this
enabled the casing to be constructed from flat,
edge-stiffened panels. It was then found that
the overall height of the unit was dependent on
the size of the gasifier cyclone; the decision
was finally made to use two smaller cyclones
instead of one large one. This gave the
configuration shown in Figure 1. It can be
seen that the gasifier bed is rectangular in
plan and that the regenerator is between the
two gasifier cyclones to one side of the
gasifying reactor.
The bed transfer system which was selected
was an adaptation of a system proven in com-
mercial practice, where it is employed to
control the rate at which bed material
descends through a series of stacked beds. The
adaptation was necessary in order to enable
the system to be utilised to transfer bed
material between two beds in parallel and at
the same height. In this system the bed trans-
fer duct is almost vertical, but incorporates a
horizontal section at its lower end. In the
absence of any external agency the duct simply
fills with static bed material. When a pulse of
gas is introduced into the horizontal section,
however, bed material flows down the duct
under the influence of gravity. The adaption
involved utilising the two fluidised beds as lift
pumps for the bed material, each bed dis-
charging the material into a cavity at the top
of a transfer duct. The two transfer ducts can
be seen in dotted outline in Figure 1 and they
are situated on either side of the regenerator in
plan. The existence of these transfer ducts
adjacent to each of the cyclones suggested that
they might also be used as cyclone drains. This
was in fact done and in operation half of the
gasifier cyclone fines are returned to the
gasifier itself and half are passed on to the
regenerator. The fines leaving the regenerator
bed are trapped by an external cyclone and
drained from the system.
It was necessary to test a novel bed transfer
system of this complexity before casting it in
concrete. This was done by building the full
scale cold rig illustrated in Figure 2. The bed
material used in this rig was a crushed brick of
roughly the same density and size distribution
as the lime. Although the gas velocities were
matched in the fluidised beds, there was a
strong element of conservatism in the opera-
tion of the transfer system; no allowance was
made for the very considerable volumetric
expansion of injected gas under hot condi-
tions. It was also anticipated that the higher
gas viscosity at the normal operating tempera-
ture would improve the flow properties of the
bed material. Tests run with this rig soon
showed that the introduction of the cyclone
fines into the transfer ducts via simple
branches led to severe bridging problems. It
was deduced that this bridging was caused by
fines being blown back up the transfer lines to
block the interstices between the particles of
descending bed material.
A way out of this difficulty was found by
devising the mixing cavity shown in Figure 3.
The shape of this cavity is such that a pocket
of gas is trapped within it, leaving a free
surface of bed material above the horizontal
section of the transfer duct. The theory behind
this design is that when the transfer duct is
activated, bubbles of gas rise into the pocket
and displace gas already there; this gas re-
enters the free surface of bed material and
proceeds up the transfer duct. In this way the
free surface can act as a filter for fines. In
practice the device which was built of plexi-
glass worked very well. Lenses of fines were
trapped by the coarser solids, and these
inclusions moved down towards the horizontal
transfer duct where they disappeared. This
modification solved the bridging problem.
Other design problems related to the
thermal expansion of the refractory block
when it was brought up to operating
temperature. There was also the question of
thermal insulation to be dealt with. The
construction of the unit allowed the transverse
thermal expansion in a very simple fashion.
The plates forming the casing of the gasifier
were lined with 3 in. thick slabs of 50 lb/ft3
castable insulating material. When the casing
V-l-3
-------
was assembled the inner surface was lined
with 1/4 in. thick expanded polystyrene sheet;
the refractory concrete block was cast within
this lining. The expanded polystyrene was
subsequently melted out leaving a gap
between the refractory block and the insula-
tion. The vertical expansion of the block posed
more difficult problems because it was
necessary to connect the gasifier to the burner
in a gas-tight manner, but which allowed 3/8-
in. relative vertical movement. Since there
were two gasifier outlets it was necessary to
employ a Y-shaped bifurcated duct consisting
of a mild steel casing enclosing a refractory
concrete lining in which the gas passages are
cast (Figure 4). A layer of calcium silicate slab
insulation is sandwiched between the
refractory concrete and the steel casing to
minimise heat losses.
The duct is suspended immediately above
the cyclone outlets with an expansion
allowance between the corresponding faces.
The expansion joint is sealed by a stainless
steel bellows to provide a gas-tight assembly.
The duct is supported at the other end by a
roller which accommodates the horizontal
expansion of the duct and burner; it is sealed
to the boiler face with a compressible
insulating seal.
The burner design presented some
problems because there was no information
available concerning the combustion charac-
teristics of this hot gas. Some simple burners
had been tested on the early batch units which
showed that the gas would burn easily. By
introducing some premix air to the burner it
was possible to burn the gas with a steady
smoke free flame and low excess air.
The 7 x 106 Btu/hr burner used for the
continuous gasifier is illustrated in Figure 5. It
consists of two sections--a premix zone in
which about 10 percent of the combustion air
may be introduced and a main section in
which the balance of the air is added. In both
of these sections an inner stainless steel
assembly is used which is insulated from the
cold combustion air introduced around the
assembly. The insulation is necessary to main-
tain the temperature of the hot gas duct which
otherwise might become obstructed with con-
densing material from the hot gas.
These inner insulated assemblies are fixed
to the outer casing at one end and are free to
expand along the axis of the burner at
operating temperature. The gas issues from
the burner to mix with the main combustion
air through a stainless steel orifice sized to give
a pressure drop between 3 and 4-in. water
gauge.
For safety reasons it was decided to use a
continuous pilot flame on the experimental
plant; here problems arose because this flame
must be stable at normal running conditions
which means a high gas and air rate, unlike a
conventional burner which lights off a pilot at
a low flame setting. The problem was
overcome by placing a small stainless steel
deflector plate to shield the flame of the pilot
from the main air.
The only other problem with the burner
arose from gas turbulence at the entry to the
burner orifice nozzle which threw out deposits
in the burner inner duct. This was overcome
by smoothing the gas flow into the orifice by a
suitable entry duct; no further deposits were
observed.
Because the overall height of the gasifier is
considerably greater than the centre height of
the boiler furnace, it was necessary to place
the gasifier in a pit in order to line up the two
components. A general plant layout is shown
in Figure 6.
The air distributor of the gasifier is
provided with horizontal nozzles made by
drilling six radial holes of 0.177-in. diameter
through each of 32 stainless steel capped
tubes. The distributor and its plenum form a
removable box structure built of mild steel;
the top of the box, from which the nozzle
assemblies project, is covered with refractory
concrete.
V-l-4
-------
There are a number of penetrations
through the walls of the gasifier to provide
access for thermocouples, manometer probes,
fuel injection tubes, bed drains and the gas
injectors used to activate the bed transfer
system. These penetrations were all cored
before the block was cast. In addition, a pre-
cast quarl for the startup burner was also
placed in position before the refractory
concrete was cast.
Figure 7 shows the lower core assembly of
the gasifier during an early stage of construc-
tion. The cores for the transfer system were
made of plexiglass and were filled with wax in
order to avoid the possibility of damage and
filling with concrete during the casting opera-
tion. Figure 8 shows the unit after the first two
lifts had been cast.
G.R. Stein Refractories Ltd. advised on the
method of construction and built the unit
using their refractory concrete Durax C.1600.
This material contains about 50 percent
alumina, 42 percent SiO2, 5 percent CaO,
with traces of Fe2O3 and MgO. The
maximum operating temperature is 1600°C
with a melting point of 1710°C.
In the first instance the only internal metal
components of the gasifier were the cyclone
outlet tubes. These proved to be unsatisfactory
and were subsequently replaced by tubes of
self bonded silicon carbide.
THE ASSOCIATED SYSTEMS
The pilot plant flow plan is shown in Figure
9. It will be seen that the startup burner is
fired by propane. This burner is used to heat
the gasifier to its working temperature and is a
standard commercial burner with a variable
output, delivering 700,000 Btu/hr at the
maximum firing rate. Propane is also used to
fuel the pilot burner fitted to the boiler. The
main fuel system utilises three small metering
pumps which draw fuel oil from a circulating
stream in a ring main and deliver it to the
three fuel injectors of the gasifier. A small
amount of air is injected with the oil in order
to prevent coking in the injector tubes. A
switch from fuel oil to kerosine is also
provided. This enables the consumption of
propane during the warm-up period to be
reduced while avoiding the introduction of
sulphur into the gasifier. The flue gas recycle
system which is used to control the tempera-
ture of the gasifier is also shown in the flow
plan. The flue gas recycle stream is first
cleaned in a cyclone, then passed through an
orifice plate flowmeter and a control valve to
the inlet side of the first gasifier blower. A
second control valve throttles the air supplied
to this blower; by making suitable adjust-
ments to these two valves it is possible to vary
both the total supply of gas to the plenum of
the gasifier and the composition of the gas in
terms of the proportions of flue gas and air
which it contains.
The operating temperature of the regener-
ator tends to be self regulating when no
oxygen is present in the tail gas, because CaS
can yield two oxidation products — CaSC>4
and CaO + SO2. The first reaction releases
much more heat than the second reaction; but
as the temperature rises the second reaction
tends to predominate so that in effect the
calorific value of the sulphur fed to the regen-
erator tends to fall. In order to hold the
temperature at a specified level however the
bed transfer system is arranged to auto-
matically increase the transfer rate when the
temperature falls below the set point. Because
of the temperature difference between the two
beds, the consequent adjustment to the rate of
heat transfer from the regenerator to the
gasifier brings the regenerator temperature
into line. The rate of SO2 release at any set
regenerator temperature depends on the rate
at which air is fed to the regenerator. For
experimental purposes this is a manual
adjustment; when completely automatic
control is used, the air rate to the regenerator
may be controlled to hold the O2 concen-
tration in the regenerator tail gas at a constant
level. The return of solids from the regenerator
to the gasifier is controlled by the pressure
drop across the regenerator bed.
V-l-5
-------
Equipment was not installed to deal with
the 10 percent SC>2 stream from the regenera-
tor because this did not form part of the
development programme at this stage. The
SO2 stream from the pilot plant was therefore
connected into the boiler stack, thus creating a
flue gas identical to that resulting from direct
combustion of the test fuel.
The instrument flow plan for the
installation is shown in Figure 10. This,
however, does not show the packaged pres-
surisation system which maintains constant
boiler water pressure and temperature.
Considerable attention has been given to
safety measures; the plant is protected by a
number of sensor systems which detect both
hazards and conditions. Signals from the
sensors will, according to a predetermined
selection, either shut down the whole plant
and operate an alarm, or give an alarm and a
visual indication of the trouble, or merely give
a visual indication.
During test runs the unit is operated on a
24-hour shift basis, with one professional and
two non-professional personnel in each shift
team.
OPERATIONAL PROBLEMS
Three operational problems became
apparent during the first attempts to run the
unit for a prolonged period:
1. The formation of lime/coke deposits in the
ducting between the gasifier and the
burner.
2. A tendency to form too large a proportion
of CaSO4 within the regenerator.
3. Poor containment of bed fines.
The deposits in the gas ducting were
formed locally in areas of high turbulence. In
the first instance deposits tended to choke the
inlet ports of the cyclones which were origin-
ally square edged. This problem was greatly
alleviated by chamfering these edges to give a
smoother flow transition. Two other critical
areas were found where the vertical cyclone
exit channels intersected the converging hori-
zontal channels within the Y-shaped duct. The
design of the ducting has since been modified
to provide a smoother gas passage at these
elbows.
The deposits laid down in the gas ducting
could be removed without shutting down the
unit by a controlled burn-out procedure. A
more serious form of deposit was, however,
found within the regenerator and within the
transfer line from the regenerator to the
gasifier. The deposits in these areas built up
relatively slowly, but could only be removed
when the unit was shut down. It was deduced
that the cause of these deposits, which con-
tained no carbon,was the excessive formation
of CaSO4 within the regenerator.
The oxidation of CaS in the regenerator is
analagous to the oxidation of carbon in the
gasifier in that two reactions occur and they
appear to occur sequentially. In the gasifier
the carbon deposited on the lime is first
oxidised to CO 2 which is subsequently
reduced to a large extent to CO on passing
through the bed. Within the regenerator there
is a tendency to form CaSO4 near the distri-
butor which subsequently reacts with CaS to
form CaO + SO 2. It has been observed by
Curren, Fink and Gorin2 that during the
course of this reaction a transient liquid is
formed which can cement particles together.
This is thought to be the mechanism by which
deposits were laid down within the regenerator
and in the regenerator to gasifier transfer
duct. These deposits were largely composed of
calcium sulphate; a photograph of the deposit
which was removed from the regenerator at
the completion of Run 3 is shown in Figure 11.
This deposit was wedge shaped and grew from
the wall of the regenerator opposite the outlet
of the bed transfer duct from the gasifier. As
can be seen in Figure 12 the distributor
nozzles under the deposit were themselves
blocked by coarse bed material; it seems likely
that the initial presence of a dead zone
induced deposit formation in this area. The
design of the regenerator distributor has since
been modified and it is now more similar to
that of the gasifier.
V-l-6
-------
The formation of deposits in the regen-
erator is best avoided by reducing the
tendency to form CaSC>4. This may be done by
introducing the bed material from the gasifier
into the regenerator at a level well above that
of its distributor.
The effect of this change would be to
ensure that both the fresh bed material from
the gasifier and the incoming air are brought
up to the working temperature before they
meet. Under these circumstances there should
be a greater .tendency to form unstable CaSO3
in a single step; less CaSC>4 should be avail-
able for the liquid phase decomposition. In the
case of the experimental unit this has been
attempted by lowering the level of the regen-
erator distributor.
The gasifier cyclones were not found to be
very effective in retaining the fines which were
produced by attrition and decrepitation. In
normal operation they passed solids at a rate
amounting to about 2 percent by weight of the
fuel used. It is likely that this poor per-
formance was largely due to the rough surface
finish of the cyclones, but it was considered
that the best way to prevent solids from
entering the boiler would be to reduce the
amount of fines produced.
Decrepitation occurs during calcination
and may be reduced by lowering the bed
replacement rate. The batch results indicate
that increasing the depth of the bed will not
impair desulphurization efficiency. An indi-
cation of the importance of bed losses incurred
shortly after the entry of stone into the gasifier
was the fact that at the end of Run 3 the
vanadium content of the bed was three times
higher than could be accounted for on the
assumption of a uniform bed life.
There were two mechansims which might
have contributed to this effect-decrepitation
and elutriation during addition. During these
runs the bed material was fed into the gasifier
through its lid and fell in a stream past the
entry of one of the cyclones. It was thought
likely that some of the finer material was
swept out of the unit before it entered the bed.
In order to reduce this possibility the stone
feed system was modified; stone now enters
the unit through the wall of the gasifier in the
vicinity of the bed surface and at a point
remote from the cyclone inlets. Attrition in
fluidised beds increases rapidly as the superfi-
cial gas velocity rises. The highest gas
velocities within the bed occur at the distribu-
tor nozzles and are required in order to ensure
bed stability. The problem here was to
combine bed stability with a low nozzle efflux
velocity. The adopted solution was to use two
stage nozzles in which the kinetic energy
imparted to the gas by the pressure drop
through the first nozzle was dissipated prior to
the low velocity entry of the gas into the bed
via the second nozzle.
• During the first phase of operation the unit
was run for 450 hours under gasifying condi-
tions; the main structure does not appear to
have suffered any significant deterioration. A
major objective of future work is to reduce the
quantity of fines leaVing the bed to a level
which will enable us to envisage the construc-
tion of a full scale unit which would, not
require hot cyclones. If this could be done
there would be a considerable saving on
investment and also an improved operability
due to the decrease in deposit formation
arising from the simplification of the gas
ducting. Finally, it seems reasonable to
conclude that the operational problems which
have so far been encountered do not appear to
be unduly severe or intractable.
REFERENCES
1. Craig, J.W.T., G. Moss, J.H. Taylor, and
D.E. Tisdall. Sulphur Retention in
Fluidised Beds of Lime Under Reducing
Conditions. (Presented at 3rd International
Conference on Fluidized-Bed Combustion.
Hueston Woods. October 29 - November 1,
1972.) (See Session HI, Paper 4 this
volume.)
2. Curren, G.P., C.E. Fink, and E. Gorin.
Phase II Bench-Scale Research on C.S.G.
Process, Research and Development No.
16. Consolidation Coal Company, Library,
Pa. Prepared for Office of Coal Research,
U.S. Department of the Interior,
Washington, D.C. under Contract Number
14-01-0001-415. July 1969.
V-l-7
-------
CONNECTION BETWEEN CYCLONESv
EXPANSION BELLOWS TO
ABSORB VERTICAL EXPANSION
ON GAS OUTLET
REMOVABLE LID
BURNER
FUEL SUPPLY
SECTION A-A
BED
DRAIN
OUTER METAL CASING
INSULATING REFRACTORY^
CASTABLE REFRACTORY
SECTION B-B
/ / /// //// / ///////
3.25 ft
FUEL
SUPPLY — —
CYCLONE FOR
GASIFIER
REGENERATOR
CYCLONE FINES
FED INTO BED
TRANSFER PIPES
V-l-8
GAS PULSE
Figure 1. Layout of continuous gasifier unit.
7.75ft
r? "I RETURN TUBES FOR
BED RETURN
FROM
REGENERATOR
m
JISTRIBUTOF
4.75 It
-------
Figure 2. Full-scale cold rig.
V-l-9
-------
FINES FROM CYCLONE
SOLIDS FROM GASIFIER
DESCENDING SOLIDS
GOOD LEG SEAL
FLUIDISED REGION
INLET TO REGENERATOR
Figure 3. Successful cyclones fines return zone.
V-l-10
-------
SLAB INSULATION
REFRACTORY LINING
EXPANSION ALLOWANCE
BELLOWS SEAL
REFRACTORY LINING
CYCLONE OUTLET CONNECTION
GAS OUTLETS PIPES -
BURNER
\
1 1
1 1
SECONDARY
AIR
PREMIX
AIR
SECTION A - A
SUPPORT YOKES
SUPPORT FRAME
PLAN
Figure 4. Cyclones to burner manifold.
V-l-11
-------
r ."n \ . . j^'T'i "Tnrr v1i \ iTlrTl
m\\\\\\\\\\\\\ft\\
/Cv\\\\\\\\\\\\\\
*•-••' t*f^ * * ' f * t j * ' -*->'~
\\\\\\\\\\\\\\\\
SECONDARY
AIR
Figure 5.. Main gas burner.
PRIMARY
AIR
V-l-12
-------
REGENERATOR CYCLONE
ELECTRICAL
CONTROL
CABINET
16ft
FUEL INJECTOR (3 off)
AIR SUPPLY TO
GAS I HER
GASIFIER
\
\
TO
CHIMNEY
COOLER
HEAT EXCH.
BOILER
AIR SUPPLY TO
REGENERATOR
CIRCULATING OIL
SUPPLY FROM
30- BY 9-ft TANK
SIDE SECTION
ELECTRICAL
CONTROL
CABINET
SWING JIE
PILLAR C
1
• OJ 11
MECHANICAL r
CONTROL
CONSOLE
"" GASIFIER
IRANE , 1 Tj
/XB*
I/ \
1 x_
iS5
BLOWERS
FOR
GASIFIER
OFFICE
| HEAT
BOILER
•rn c
EXCH. 1
Q
32
r
«
T
» ft uiru
PLAN
STACK
Figures. General plant layout.
V-l-13
-------
Figure 7. Gasifier lower core assembly during early construction stage.
V-l-14
-------
Figure 8. Gasifier lower core assembly following casting of first two lifts.
-------
COMBUSTION AIR BLOWER
LIME DRAIN
REGENERATOR AIR BLOWERS
N2 FOR SOLIDS
TRANSFER
FUEL INJECTION
AIR
(3)
PROPANE FOR
START-UP
Figure 9. CAFB pilot plant flow plan.
-------
FLUE GAS
BOILER GAS ANALYSIS
f/c
M
d/p
P/s
- S02 TO STACK
REGENERATOR GAS ANALYSIS
BED DENSITY
BED LEVEL!
P/s d/p M d/p
M
.nr.gr-
DISTRIBUTOR
A P
L..
GASIFIER
TR
FREQUENCY CONTROLLER
MANOMETER
DIFFERENTIAL PRESSURE CELL
PRESSURE SWITCH
TEMPERATURE RECORDER
TEMPERATURE CONTROLLER
GASIFIER-REGENERATOR
A?
REGENERATOR INLET
SOLIDS TRANSFER
N2
GASIFIER PLENUM
GAS ANALYSIS
STATIC PRESSURE LINE
—\— TRANSMITTED SIGNAL
MANUAL CONTROL
^™"" "*t« ?\^
A^J
^""" *
___l M
T
; !r
L
\IR
REGENERATOR OUTLET
SOLIDS TRANSFER
—>
AIR
FLUE GAS RECYCLE
Figure 10. CAFB pilot plant instrument flow plan.
-------
Figure 11. Deposit removed from regenerator after Run 3.
-------
Figure 12. Distributor nozzles blocked by coarse bed material.
-------
2. DESIGN OF FLUIDIZED-BED MINEPLANT
M. S. NUTKIS AND A. SKOPP
Esso Research and Engineering
ABSTRACT
Fluidized-bed combustion of coal offers potential both as an efficient compact combustion-
boiler system and an air pollution emissions control system. Esso Research and Engineering
Company, under contract to the Office of Air Programs of the Environmental Protection Agency,
has designed a system capable of fluidized-bed coal combustion and desulfurization with
continuous limestone regeneration.
The fluidized-bed miniplant will operate at pressures up to 10 arm and with an input of
approximately 6.3 x 106 Btu/hr. This energy input is equivalent to a power plant rated at 635kW
(0.63 MW).
In the miniplant, combustion and heat transfer to tubes take place in a reactor containing a
fluidized-bed of limestone at 1500-1700°F, providing good heat transfer and an efficient desulfuri-
zation reaction between the sulfur dioxide and limestone. The calcium sulfate produced during
desulfurization is transferred to an adjacent fluidized-bed reactor and contacted with a reducing
gas at 1900-2050°F. This regenerates the lime for reuse in the combustor and produces a by-
product off-gas stream of concentrated sulfur dioxide. Thus, regeneration minimizes the limestone
feed requirements and the calcium sulfate disposal problems.
The fluidized-bed miniplant design incorporates a 12.5-in. ID combustor and a 5-in. ID regen-
erator vessel with continuous transfer of solids between these two refractory lined reactors. In the
combustor, the fluidizing air enters a plenum, passes through the distributing grid, up through the
fluidized bed of solids and the combustion products discharge through two refractory cyclones in
series.
Superficial bed velocities and pressures in the combustor and regenerator are automatically
controlled. The pressure differential between the two vessels can also be automatically controlled.
Heat extraction and temperature control in the fluidized-bed combustor are accomplished by
vaporizing demineralized water in 10 independent loops located in discrete vertical zones of the
reactor. The water flows to these loops are controlled by valves whose positions change to maintain
bed temperature in each of the zones.
Coal and makeup limestone to the combustor are fed continuously from a system designed for
controlled solids feeding under pressure. Solids transfer between reactors and discharge of solids
from the system (i.e., from the regenerator reactor) are accomplished using a pulsed gas transport
technique controlled by pressure differentials across and between these fluidized beds.
V-2-1
-------
INTRODUCTION
The fluidized-bed coal combustor provides
a new boiler technique where coal is com-
busted in a bed of particles maintained in a
state of fluidization by the air required for
combustion. The use of limestone or other
suitable sorbent as the bed material in such a
system permits the capture and removal of
sulfur dioxide simultaneously with the com-
bustion process.
Fluidized-bed boilers (FBB) offer the
potential of an efficient and compact boiler
combustion technique also capable of pro-
viding pollution control. Some of the
advantages and economic factors are:
1. Capability to combust lower quality fossil
fuels in a fluidized bed.
2. Immersion of the boiler tubes directly in
the fluidized bed achieves improved heat
transfer rates compared to conventional
boilers.
3. The higher volumetric heat release rates in
a fluidized-bed combustor will permit
reduced boiler unit sizes.
4. Since efficient combustion can be achieved
at comparatively low bed temperatures
(i.e., 1500-1700°F), boiler tube corrosion
and fouling should be reduced.
Pressurized fluid-bed combustion offers even
greater benefits in size reduction, efficiency,
and load control.
Within the fluidized-bed boiler, limestone
is calcined to lime which reacts with sulfur
dioxide and oxygen in the flue gas to form
calcium sulfate. When used on a once-through
basis, relatively high limestone feed rates are
required to the fluidized-bed boiler if sulfur
dioxide removal in excess of 90 percent is to be
maintained.
In order to reduce these high limestone
feed rates, a system was proposed by Esso
Research and Engineering whereby the cal-
cium sulfate formed would be regenerated
back to calcium oxide in a separate fluidized-
bed reactor (i.e., regenerator) by reaction with
a reducing gas at a temperature of approx-
imately 2000°F. The regenerated lime would
be returned to the fluidized-bed combustor,
where it would again react with the sulfur
dioxide.
In a study completed by Esso for the
National Air Pollution Control Administra-
tion,1 the following essential features of the
proposed regenerative-limestone FBB system
were demonstrated:
1. Removal of over 90 percent of the SO2
formed by combusting coal in fluidized
beds of lime.
2. Reductive regeneration of the sulfated
lime to yield an off-gas containing 7 to 12
mole percent SO2. This is a sufficiently
high concentration to permit its con-
version to H2SO4 or elemental sulfur with
conventional technology.
3. Good activity maintenance of the lime
cycled back and forth between combustion
and regeneration. The make-up require-
ment for fresh limestone in a commercial
plant was estimated to be about 15 percent
of that required for once-through use of
this material.
These experimental results were obtained
at atmospheric pressure conditions. Since
completing this study, engineering and cost
analyses carried out by Westinghouse2 for the
Environmental Protection Agency (EPA) have
indicated a much greater commercial
potential for a pressurized FBB system when
used in conjunction with a combined gas-
s |