U.S. Environmental Protection Agency Industrial Environmental Research
Office of Research and Development Laboratory
Research Triangle Park, N.C. 27711
EPA-600/7-76-011
September 1976
STUDIES OF THE
PRESSURIZED FLUIDIZED-BED
COAL COMBUSTION PROCESS
Interagency
Energy-Environment
Research and Development
Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S.
Environmental Protection Agency, have been grouped into seven series.
These seven broad categories were established to facilitate further
development and application of environmental technology. Elimination
of traditional grouping was consciously planned to foster technology
transfer and a maximum interface in related fields. The seven series
are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from
the effort funded under the 17-agency Federal Energy/Environment
Research and Development Program. These studies relate to EPA's
mission to protect the public health and welfare from adverse effects
of pollutants associated with energy systems. The goal of the Program
is to assure the rapid development of domestic energy-supplies in an
environmentally—compatible manner by providing the necessary
environmental data and control technology. Investigations include
analyses of the transport of energy-related pollutants and their health
and ecological effects; assessments of, and development of, control
technologies for energy systems; and integrated assessments of a wide
range of energy-related environmental issues.
REVIEW NOTICE
This report has been reviewed by the participating Federal
Agencies , and approved for publication. Approval does not
signify that the contents necessarily reflect the views and
policies of the Government, nor does mention of trade names
or commercial products constitute endorsement or recommen-
dation for use.
This document is available to the public through the National Technical
Information Service, Springfield, Virginia 22161.
-------
EPA-600/7-76-011
September 1976
STUDIES OF THE
PRESSURIZED F LUIDIZE D-BED
COAL COMBUSTION PROCESS
by
R.C. Hoke, R.R. Bertrand, M.S. Nutkis,
D. D. Kinzler, L.A. Ruth, andM.W. Gregory
Exxon Research and Engineering Company
P. O. Box 8
Linden, New Jersey 07036
Contracts No. 68-02-1312 and -1451
Program Element No. EHE623A
EPA Project Officer: D. Bruce Henschel
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
ABSTRACT
The pressurized fluidized bed coal combustion process was studied in
two experimental units including the new 2 MW (thermal) "miniplant"
unit. The shakedown phase of the miniplant program was successfully
completed, culminating in a continuous 100 hour demonstration run.
The unit was operated at pressures up to 1020 kPa (10 atm), super-
ficial velocities up to 3.2 m/s (10.5 ft/sec), temperatures up to
980°C (1800°F), coal feed rates up to 155 kg/hr (340 Ibs/hr) and
at combustion intensities of 5 MW/m3 (480,000 BTU/hr ft3) . Improve-
ments in the coal feeding system and in the steam coil design were
required to achieve these performance levels.
Operating results from the miniplant and the older batch combustion
unit indicate that S02 emissions can be controlled to levels meeting
current EPA emission standards with either limestone or dolomite sor-
bents. However, dolomites are more effective and generally will
require less sorbent to achieve the same level of S02 control. NOX
emissions from pressurized fluidized bed coal combustion can be con-
trolled to levels of 0.09 to 0.17 g (as N02)/MJ (0.2 to 0.4 Ibs/MBTU).
This compares to the current standard of 0.3 g (as Nt^/MJ (0.7 Ibs/
MBTU). Control of particulate emissions to current standards cannot
be achieved by two stages of conventional cyclones. A very efficient
third stage particulate removal device will be required.
The miniplant and batch experimental units will be used in additional
studies characterizing the environmental effects of the fluidized bed
coal combustion process. This report was submitted in fulfillment of
Contract Numbers 68-02-1312 and 68-02-1451 by Exxon Research and
Engineering Company under the sponsorship of the Environmental Pro-
tection Agency. Work was completed in August, 1975.
iii
-------
CONTENTS
Page
Abstract iii
List of Figures vi
List of Tables ix
Acknowledgements x
Sections
I Conclusions 1
II Recommendations 5
III Introduction 7
IV Miniplant Shakedown 10
Equipment 10
Materials 33
Procedures 36
Unit Performance 37
Equipment Performance 41
Combustion Results 73
V Batch Combustor Studies 80
Equipment, Materials, Procedures 80
Equipment and Technique Development 96
Combustion Results 105
VI Discussion of Results 132
S02 Emissions 132
NOX Emissions 135
Combustion Efficiency 135
Heat Transfer Coefficients 139
VII References 140
VIII List of Publications and Patent Memoranda
iv
-------
CONTENTS (CONTINUED)
IX Appendix 143
A. Materials of Construction 144
B. Reports of Metallurgical Examinations 152
C. Safety Consequences of a Steam-Coil Break 162
D. Miniplant Alarms Annunciators 164
E. Miniplant Data Logger Points 166
F. Analytical Techniques 169
G. Miniplant Run Summaries 170
H. Batch Unit Run Summaries 198
-------
FIGURES
No.
III-l Pressurized Fluidized Bed Coal Combustion System
IV-1 Exxon Fluidized Bed Combustion Miniplant 11
IV-2 Exxon Fluidized Bed Combustion Miniplant 12
IV-3 Coal Feed System 13
IV-4 Combustor Vessel 15
IV-5 Combustor Vessel Lining 16
IV-6 Combustor Preheat Burner 18
IV-7 Combustor Fluidized Grid 19
IV-8 Combustor Cooling Coils 21
IV-9 Solids Pulse Pot 22
IV-10 Solids Discharge Lockhopper 23
IV-11 Combustor First Stage Cyclone 25
IV-12 Combustor Second Stage Cyclone 26
IV-13 Flue Gas Sampling System 28
IV-14 Regenerator Vessel 30
IV-15 Regenerator Cyclone 32
IV-16 Coal Particle Size Distribution 35
IV-17 Coal Feeder Orifice Assembly 42
IV-18 Coal Flow Through a Tee 44
IV^-19 Combustor Temperature Control Schematic 46
IV-20 Horizontal Cooling Coils 47
IV-21 Horizontal Cooling Coils 48
IV-22 Vertical Cooling Coils 50
IV-23 Vertical Cooling Coils 51
IV-24 Bed Temperature Profile - Horizontal Coils 54
IV-25 Bed Temperature Profile - Vertical Coils 55
IV-26 Damaged Horizontal Cooling Coils 59
IV-27 Damaged Vertical Cooling Coils 61
IV-28 Changes in Secondary Cyclone Dimensions 65
VI
-------
FIGURES (CONTINUED)
No. Page
IV-29 Particle Size Distribution - Primary
Cyclone Collection 67
IV-30 Comparison of Particle Size Distribution
for Solids Collected by Secondary Cyclone
Before and After Modification 68
IV-31 Ignitor - Pilot 71
IV-32 NOX Emissions 74
IV-33 Combustion Efficiency 76
V-l Batch Fluidized Bed Coal Combustion Unit 81
V-2 Batch Fluidized Bed Coal Combustion Unit 82
V-3 Petrocarb Coal Injector 83
V-4 Combustor Vessel 84
V-5 Fluidizing Grid 86
V-6 Preheater Burner 87
V-7 Location of Cooling Coils 88
V-8 Coal Injector 89
V-9 Flue Gas Sampling System 91
V-10 Particle Size Distribution of Arkwright Coal 92
V-ll Particle Size Distribution of Sorbents 95
V-12 Coal Feeder Orifice Assembly 99
V-13 Vertical Cooling Coil 100
V-14 Comparison of Bed Temperature Profiles
for Horizontal and Vertical Cooling Coils 101
V-15 S02 Emissions - Limestone Sorbent 106
V-16 S02 Emissions - Dolomite Sorbent 107
V-17 NOX Emissions 109
V-18 Particle Size Distribution - Overhead Samples 120
V-19 Combustion Efficiency 122
V-20 Comparison of Temperature Profiles - Wyoming
and West Virginia Coal 127
vii
-------
FIGURES (CONTINUED)
No. Page
VI-1 Comparison of S(>2 Removal Results -
Limestone Sorbent 133
VI-2 Comparison of S02 Emissions from Batch Unit
and Argonne NL Study - Dolomite Sorbent 134
VI-3 Comparison of S(>2 Emissions from Limestone
and Dolomite Sorbents - Batch Unit Data 136
VI-4 Comparison of NOX Emissions 137
VI-5 Comparison of Combustion Efficiencies 138
viii
-------
TABLES
No. Page
IV-1 Miniplant Coal Analyses 34
IV-2 Miniplant Fluidized Bed Combustion
Test Conditions 38
IV-3 Control of Operating Variables 39
IV-4 Miniplant Cooling Coil Modifications 52
IV-5 Miniplant Cooling Coil Damage 58
IV-6 Miniplant Overall Heat Transfer Coefficient
Measurements - Run 19.2 78
IV-7 Miniplant Overall Heat Transfer Coefficient
Measurements - Runs 14.1, 14.2, 15.1 79
V-l Coal Particle Size Distribution Penn-Rillton
Co. Grind B-2 Specification 93
V-2 Composition of Coals Used in Batch Fluidized
Bed Coal Combustion Program 93
V-3 Properties of Sorbents Used in Batch Fluidized
Bed Coal Combustion Program 94
V-4 Data for Run No. 3675-2C from Which Samples
Were Analyzed for Trace Elements 111
V-5 Elements Detected by Neutron Activation Analysis 112
V-6 Neutron Activation Analysis - Upper Limits 113
V-7 Material Balances for Typical Components 115
V-8 Retention by Stone of Elements Present in Coal 117
V-9 Comparison of Exxon and Argonne N.L. Data
on Trace Element Recoveries 118
V-10 Heat Transfer Coefficients (Bed to Tube) Measured
During Coal Combustion 124
V-ll Heat Transfer Coefficients Measured During
Bed Pre-Heating 125
V-12 Particle Size Distribution of Sulfated Sorbents 128
V-13 Sulfur Balances for Batch Combustor 129
V-14 Calcium Balances for Batch Combustor 130
ix
-------
ACKNOWLEDGEMENTS
The authors wish to express their appreciation to the many individuals
who played major roles in the shakedown and operation of the FBC units
at Exxon Research and Engineering Company. Specifically, we wish to
acknowledge the efforts of V. J. Siminski, H. R. Silakowski, H. C.
Bunje, T. C. Gaydos, R. E. Long, G. E. Walsh, D. T. Ferrughelli,
E. Hellwege and J. E. Bond of the Government Research Laboratories.
We also wish to acknowledge the efforts of personnel of the Mechanical
Division, in particular R. A. Van Sweringen, E. C. Vath, S. Pampinto,
T. Morrison, F. Huber, T. Artz, T. Morgan and E. Sullivan. A special
acknowledgment goes to Mrs. N. Malinowsky who typed this report.
The personnel of the Industrial Environmental Research Laboratory of
the E.P.A. have also been most helpful. We wish to express our grati-
tude for the help of D. B. Henschel the current EPA Project Office,
S. L. Rakes, the previous Project Officer, P. P. Turner and R. P.
Hangebrauck.
x
-------
SECTION I
CONCLUSIONS
MINIPLANT PERFORMANCE
The combustor section of the Exxon pressurized fluidized bed coal
combustion miniplant is capable of operating at its basic design
conditions with close control of operating variables.
Operation was demonstrated at pressures up to 1020 kPa (10 atm),
superficial velocities up to 3.2 m/s (10.5 ft/sec), temperatures up
to 980°C (1800°F) and combustion intensities of 5 MW/m3 bed volume.
Coal feed rates of 155 kg/hr and combustion heat release rates of
1.3 MW have been demonstrated to date. These are somewhat lower than
specified in the design, but no difficulty is seen in operating at or
above the design conditions.
Sustained operation of the combustor was demonstrated for a 100 hr.
period.
COMBUSTION RESULTS
The pressurized fluidized bed combustion (FBC) process is capable of
meeting the current EPA standards for S02 and NO emissions. Both
limestone and dolomite are suitable sorbents under pressurized com-
bustion conditions although dolomite is a more effective S02 sorbent,
at least on a molar basis. A preliminary assessment of sorbent require-
ments based on data available to date indicates that the use of limestone
will require a calcium to sulfur molar ratio of about 2.0 to 3.0 while
dolomite use will require a calcium to sulfur ratio of about 1.0 to
1.5 under pressurized FBC operating conditions.
NOX emissions are in the range of 0.09 to 0.17 g (as N02)/MJ (0.2 to
0.4 Ibs/M BTU) compared to the current emission standard of 0.3 g
(as N02)/MJ (0.7 Ibs/M BTU).
CO emissions are low, normally in the range of 150 to 250 ppm. Poor
combustion conditions can cause an increase in CO emissions to levels
exceeding 1000 ppm.
It is possible, with the two stage cyclone separator system used on the
Exxon units, to reduce particulate emissions to the range of 0.5 to
0.9 gm/m3 (0.2 to 0.4 gr/SCF). This is equivalent to 0.4 to 0.8 lb/
M BTU and exceeds the emission standard of 0.1 Ib/M BTU. More effic-
ient cyclones could reduce these levels a bit further. However, even
with improved cyclones, a third stage particulate removal device will
be required to meet emission standards. The device will require a
removal efficiency in the range of 75 to 99+%.
1
-------
Once through carbon combustion efficiency ranges from 93 to 97% as the
excess air level increases from 20 to 100%. Increasing the combustion
efficiency to the target level of 98.5 to 99% will require recycle of
carbon fines from the first stage cyclone to the combustor or to a
carbon burnup cell.
Overall heat transfer coefficients between the fluidized bed and the
steam coils ranges from 320 to 470 W/m2 K (55 to 85 BTU/hr ft2 °F)
depending on the diameter of the cooling coils. Heat transfer coef-
ficients measured in the miniplant at conditions fairly representative
of those expected in utility size PFBC units range from 320 to 350 W/m2
K (55 to 60 BTU/hr ft2 °F).
Attrition rates measured for dolomites (Tymochtee, Pfizer No. 1337 and
Baker) are much higher than those measured for Grove No. 1359 lime-
stone. The attrition rate measured for Tymochtee dolomite varies with
the batch. Most of the dolomite attrition occurs upon calcination
during the heat up stage or the first few minutes of a run after coal
combustion has begun.
Little difference is seen in the combustion of high sulfur, caking
Eastern coal and low sulfur, noncaking Western subbituminous coal.
The batch and miniplant FBC units give comparable emission.;and com-
bustion results despite a three-fold difference in diameter.
(10 vs 32 cm)
EQUIPMENT DEVELOPMENT AND PERFORMANCE
A number of equipment and design related problem areas were uncovered
and overcome in the shakedown of the miniplant and batch units. Some
of these problem areas are described below.
The use of a pneumatic transport coal injection system is feasible,
but is sensitive to plugging and upsets. The injection system must
be designed without any internal discontinuities such as sudden
decreases in diameter to prevent plugging. In the coal feed systems
used in the Exxon units, it is also necessary to use cooled injection
probes with high velocity jets surrounding the inner coal/air core
to prevent plugging and overheating at the probe outlet. With these
modifications, the system performs satisfactorily. However, the suc-
cessful operation of the system is dependent on close control of a
fairly small pressure differential between the injector vessel and
combustor. Disturbances in the pressure differential result in upsets
in the coal feed.
-------
The temperature distribution in the Exxon combustors is influenced
by bed solids mixing rates. Closely packed horizontal cooling coils
impede solids mixing rates and cause a sharp temperature gradient in
the combustors. Use of vertically oriented coils which allow more
rapid axial mixing of the solids is necessary to obtain fairly uniform
temperatures. This factor must be considered in the design of larger
units.
The cooling coils in the Exxon combustors were also subject to damage
and distortion caused by mechanical stress. Damage usually occurred,
or, at least, was more severe after the fluidized bed had agglomerated
due to high temperature or a water leak into the bed. The coils in
the miniplant, especially the horizontally oriented colls, were part-
icularly subject to distortion which caused fracture of the tubes on
several occasions. Vertical coils are less susceptible to damage than
horizontal coils. The tubes showed no signs of corrosion as long as
metal temperatures were kept low. It is important in larger units to
protect the tubes from loss of coolant since the combination of the
mechanical forces caused by the fluidized bed and high metal temper-
atures may cause tube damage. Injection of water into the bed caused
by a tube break appears to result in agglomeration of the bed. No
indications of steam-carbon reactions were noted when a water leak
occurred. Calculations also indicate no significant reaction will
occur.
Care must be exercised to prevent condensation of moisture in the flue
gas cyclone diplegs during startup. If condensation occurs, solids in
the diplegs plug, causing a backup of solids into the cyclone, render-
ing the cyclone inoperable. As long as the dipleg is dry, the pulsing
system used to recycle solids from the first cyclone to the combustor
works satisfactorily.
Erosion rates are high in the flue gas piping and valving system
especially when the cyclones are not operating properly. A modified
pressure control system was developed to function in the presence of
high dust loading. A positive filtration system will be required to
prevent damage to flue gas turbines in larger FBC units since failure
of a cyclone could result in severe erosion damage in a very short
time.
Combustor bed temperature must be up to about 650°C before feeding coal.
At temperatures much below 650°C, the coal combustion efficiency is
very poor, resulting in high carbon loadings and CO concentrations in
the flue gas. This situation can also lead to burning of the coal
above the bed and in the cyclones. Pressure and flow surges, especially
during startup, can also result in burning in the cyclones.
-------
Heat removal from a fluidized bed combustor is dependent primarily on
the heat transfer surface area covered by the fluidized bed. The tem-
perature difference between the bed and a steam/water cooling medium
cannot be varied enough to have a significant effect on heat removal
rates. Therefore, the fluidized bed temperature is controlled by the
coal feed rate. If the expanded bed level is dropped to uncover heat
transfer surface and thereby reduce the heat removal rates, the flue
gas temperature will drop significantly. The heat transfer coefficients
between the heat removal surface and the flue gas in the dilute phase
above the expanded bed are fairly high, and a significant amount of
heat will be extracted from the flue gas.
Care must be taken to obtain representative samples of flue gas for
analysis. Condensation of water vapor, long residence time and ex-
posure to metal surfaces can affect flue gas composition by decreasing
SC>2 and increasing 803 concentrations.
-------
SECTION II
RECOMMENDATIONS
Potential operating problems were uncovered during the shakedown of
the miniplant combustor which should be considered in the design of
larger fluidized bed combustion units. The steam coil design must
permit free movement of the fluidized solids to obtain a uniform bed
temperature. Vertical arrangement of the steam coils is inherently
more conducive to free movement of the solids than horizontal coil
configurations. Horizontal configurations cannot be ruled out, but
the design of the coils should allow free movement of the solids.
Damage to the coils can also readily occur if they are not properly
supported. Horizontal coils are more susceptible to damage than
vertical coils. Large scale tests of tube bundles should be made in
"hot" combustors to assure the development of workable designs.
The pneumatic transport coal and sorbent feed system used on the Exxon
pressurized FBC units is a workable system. However, it is susceptible
to upsets and this should be considered in the design of coal feed
systems for larger units. The design of the coal and sorbent transport
system should minimize the opportunity for plugs to occur. Sudden
decreases in pipe diameter should be avoided, instead, gradual changes
using tapered connectors should be used. Multiple feed injection lines
are desirable to minimize the impact of line plugging. Alternate feed
lines should be provided to allow clean out of plugged lines. Gas flow
through the coal and sorbent transport lines should be high enough not
only to exceed the saltation velocity in the line, but also provide
easier entry of the solids into the fluidized bed. An air-jacketed
solids injection nozzle with high velocity air jets surrounding the
coal/air core is recommended to assure adequate penetration of the coal
into the fluidized bed combustor. Since the pneumatic transport system
relies on a relatively constant pressure differential between the
injector and the combustor to maintain steady coal flow, control systems
must be designed to prevent the possibility of sudden changes in the
pressure differential. Sudden changes in the pressure differential
can result in temperature runaways and bed agglomeration, or back up
of hot solids from the bed into the injector lines. Fast acting shut
off systems must be used to shut off coal flow if either of these
upsets develop.
Extremely high erosion rates will occur if the flue gas particulate
loading and velocity are high. This situation could occur in the flue
gas expansion turbine if a cyclone malfunctions. Therefore it is
recommended that a positive filtration device be used between the
cyclones and the gas turbine. The development of a continuous parti-
culate concentration monitoring device is also recommended to prevent
the possible occurrence of a sudden undetected increase of the parti-
culate loading in the flue gas.
-------
Special attention must be given to the start up of fluidized bed
combustors. A heat up system is needed which will raise the bed
temperature up to about 650°C while coolant is flowing through the
steam coils. If coal is admitted to the combustor at lower tempera-
tures, very poor combustion efficiency will occur which can lead to
burning above the fluidized bed. Coolant flow is required to pre-
vent damage to the steam coils during the heat up stage. Care must
also be exercised during startup to prevent sudden flow surges during
the period when the combustor is being brought up to operating pres-
sure and flow. Flow surges can cause burning above the bed and in
the cyclones. Steps must also be taken during startup to prevent
condensation of moisture in lines which handle solids. If condensa-
tion occurs, the line will readily plug with wet solids. Such lines
should be isolated or kept free of solids until the lines or the gas
passing through them are hot enough to prevent condensation.
Since the miniplant combustor is now fully operational, it should be
used in combustion and emissions related studies. It is capable of
operating at conditions of high velocity, pressure and temperature
in deep beds of sorbent with very high combustion intensities. It
can be used to study the environmental effects of designs and operating
conditions anticipated for larger units. It can also be used to test
performance of subsystems. One very important area requiring further
study is particulate control. The miniplant is capable of studying
and evaluating particulate control devices under conditions closely
approximating those expected in large FBC systems. Maximum use of the
miniplant is planned in environmental characterization of the fluidi-
zed bed combustion process.
Although the miniplant is now fully operational, additional improve-
ments can be made to certain components. It is recommended that a
modified coal and sorbent feed system be installed with three feed
lines. This will reduce the impact of plugs in the feed lines.
Larger coal feed vessels are also recommended. These will reduce the
frequency of the refilling operation which requires close attention
to prevent upsets in the coal feed rate. Further modification of the
vertical cooling coils now used in the miniplant is recommended to
lengthen coil life.
It is also recommended that the study of sulfated sorbent regeneration
be resumed. Additional regeneration studies should be made in the
batch unit. Following the scheduled schakedown of the miniplant re-
generator, a program should begin in the miniplant in which continuous
combustion and regeneration can be studied over extended periods of
time. Such regeneration studies are currently planned.
-------
SECTION III
INTRODUCTION
The pressurized fluidized bed combustion of coal is a new combustion
technique, which can reduce the emission of S02 and NOX from the
burning of sulfur-containing coals to levels meeting EPA emission
standards. This is done by using a suitable S02 sorbent such
as limestone or dolomite as the fluidized bed material. In addition to
emissions control, this technique has other potential advantages over
conventional coal combustion systems which could result in a more
efficient and less costly method of electric power generation. By
immersing steam generating surfaces in the fluidized bed, the bed tem-
perature can be maintained at low and uniform temperatures in the
vicinity of 800 to 950°C. The lower temperatures decrease steam tube
corrosion, allow the use of lower grade coals (since these temperatures
are lower than ash slagging temperatures), and also decrease NOX emissions,
Operation at elevated pressures, in the range of 600 to 1000 kPa, offers
further advantages. The hot flue gas from a pressurized system can be
expanded through a gas turbine, thereby increasing the power generating
efficiency even further.
In the fluidized bed boiler, limestone or dolomite is calcined and reacts
with S02 and oxygen in the flue gas to form CaS04 as shown in reaction (1)
CaO + S02 + 1/202 •*• CaS04 (1)
Fresh limestone or dolomite sorbent feed rates to the boiler can be
reduced by regeneration of the sulfated sorbent to CaO and recycle of
the regenerated sorbent back to the combustor. One regeneration system,
studied by Exxon Research and Engineering Company in the past, is the
so-called one step regeneration process in which sulfated sorbent is
reduced to CaO in a separate vessel at a temperature of about 1100°C
according to equation (2). S02 in the regenerator off gas is at a suf-
ficiently high concentration to be recovered in a by-product sulfur plant.
CO C02
CaSO, + H2 -> CaO + S02 + H20 (2)
A diagram of the pressurized fluidized bed combustion and regeneration
process is shown in Figure III-l.
Exxon Research and Engineering Company, under contract to the EPA, has
built two pressurized fluidized bed combustion units to study the com-
bustion and regeneration processes. The smaller of the two units, the
batch unit, was built under contract CPA 70-19 and was described in
-------
GAi
M\/V=
c
STEAM TURE
. —
CONDENSER ^J
TT
Vir:
5 TURBINE
F?*-
lr!
TO f
,TACK
DISCARD
3INE
^ I
PUMP
<^ Y"
kJ
^ SEPARATOR)
)
\
N
— ^
COAL AND »~
MAKEUP SORBENT
_.-•*— ^x_
IH/
\
|i
--^X-^S— --W_
1
//
\
4
AIR
COMPRESSO
SOLIDS
TRANSFER
SYSTEM
;;^:>
^^^^
^
,
/" —
-—
^
1
Ft
TO SULFUR
RECOVERY
SEPARATOR
DISCARD
BOILER REGENERATOR
Figure Ill-l
PRESSURIZED FLUIDIZED BED COAL COMBUSTION SYSTEM
-------
report EPA-650/2-74-001. That report also described the high pressure
regeneration studies carried out in the unit and the initial coal com-
bustion studies. The subsequent coal combustion studies carried out
in the batch unit over the period August 1, 1973 to July 1, 1975 under
Contracts CPA 70-19 and 68-02-1451 are described in this report. The
program was aimed at the development of equipment and operating tech-
niques, the study of the effect of process conditions on SC>2, NOX and
CO emissions, the measurement of combustion efficiency, particulate
and trace metal emissions and the measurement of heat transfer co-
efficients between the fluidized bed and steam tubes. Various coals
and sorbents were also tested.
The larger unit, called the miniplant, was designed under EPA Contract
CPA 70-19 and the design was described in a report to the EPA by Exxon
Research and Engineering Company (1). The miniplant was built under
Contract 68-02-0617. The shakedown of the combustor section of the
miniplant was funded under Contract 68-02-1312. This report describes
the combustor shakedown program which concluded in a 100 hr. continuous
run. It includes a description of the miniplant, performance during
the shakedown phase, equipment problems and corrective steps taken
and combustion results. The combustion results cover measurement
of SO,,, NO , and CO emissions, combustion efficiency and heat transfer
coefficients made during the shakedown program. The shakedown period
covered in this report is from October 1, 1973 to August 4, 1975.
-------
SECTION IV
MINIPLANT SHAKEDOWN
EQUIPMENT
The Exxon fluidized bed combustion miniplant is shown schematically in
Figure IV-1. Figure IV-2 is a photograph of the unit. The miniplant
consists of a refractory lined combustor vessel with provisions for
continuous feeding of coal and fresh sorbent and continuous withdrawal
of spent sorbent. A refractory lined regenerator vessel was built
adjacent to the combustor and, when in operation, will provide for the
continuous transfer of spent sorbent from the combustor to the re-
generator and the continuous return of regenerated sorbent to the com-
bustor. The combustor vessel is 9.75 metres high (32 ft.) lined to
an internal diameter of 31.8 cm (12.5 in) and is capable of burning
up to 218 kg/hr of coal (480 lbs/hr.). Cooling coils in the combustor
remove the heat of combustion and maintain the bed temperature in the
operating range of 800 to 1000°C (1470 to 1830°F). The maximum oper-
ating pressure is 1010 kPa (10 atm.); the maximum superficial velocity
is 3 m/s(10 ft/sec). The regenerator vessel is 6.7 metres high (22 ft.)
lined to an internal diameter of 21.6 cm (8.5 in.) and is capable of
operating at temperatures up to 1100°C (2000°F) at 1000 kPa (10 atm)
pressure.
The following discussion will focus in more detail on the major system
components which include: 1) solids feeding system, 2) fluidized bed
combustor with internal subcomponents, 3) combustor cyclones, 4) pres-
sure control and flue gas discharge system, 5) flue gas sampling and
analytical system, 6) process monitoring and data generation system,
7) combustor safety and alarm system, 8) fluidized bed regenerator
system, and 9) miniplant support structure.
Solids Feeding System
The solids feeding system, originally designed by Petrocarb, Inc. and
subsequently modified by Exxon Research is illustrated in Figure IV-3.
The system provides for uninterrupted solids feed (coal and limestone)
from the primary injector to the combustor while allowing intermittent
refilling of the primary injector (193 kg operating capacity) when its
charge is reduced below 102 kg.
Solids in the primary injector are continuously aerated at a controlled
pressure above that in the combustor. They exit the primary injector
through a 1.3 cm diameter orifice and are pneumatically conveyed by
a controlled stream of dry transport air through an s-shaped 1.3 cm I.D.
stainless steel pipe leading into the combustor. A short segment of
1.3 cm I.D. rubber hose is used to connect the injector to the transport
10
-------
ORIFICE
COOLING
WATER
MAIN
AIR
COMPRESSOR
(1400 SCFM
@150'PSIG)
LIQUID FUEL
STORAGE
FIGURE IV-I. Exxon Fluidized Bed Combustion Miniplant
-------
FIGURE IV-2
EXXON FLUIDIZED BED COMBUSTION MINIPLANT
L2
-------
Figure IV-3
COAL & LIMESTONE FEED SYSTEM
LIMESTONE BIN
CONTROLLERS
HIGH PRESSURE AIR
TC
1 h
••"> L_ V^V
r—?O——(Z_H
COMBUSTOR
1/2 S.S. PIPE
-------
line in order not to interfere with operation of the load cells, located
under the injector, which are used to monitor solids feed rate. Rate of
solids feed is automatically controlled in order to maintain a specific
operating temperature within the combustor. This is accomplished through
a series of controls involving the pressure differential between the
primary injector and combustor, the injector pressure, and the transport
air flow rate. Final entry of solids into the combustor is through a
1.3 cm I.D. probe located 28 cm above the fluidizing grid and horizontally
extending about 2.5 cm beyond the reactor wall. The tip of the probe
includes ten 0.79 mm diameter holes which surround the solids feed
opening. They are used to continuously inject an annular stream of
sonic-velocity air to assist penetration of the solids feed into the
fluidized bed and alleviate any tendency of the probe tip to become
blocked with bed solids.
The remainder of the solids feeding system is involved in generating a
specific feed ratio of coal to limestone, and refilling the primary
injector in response to an appropriate weight demand signal. Crushed
and sized coal and limestone are stored separately in 13.6 tonne and
1.8 tonne capacity storage bins, respectively. From here, they are
screw-fed at a preselected ratio through a blender into the feed injector
(91 kg operating capacity). Transfer from the feed injector to the pri-
mary injector is done penumatically.
Prior to initiation of a refilling operation, the primary injector, feed
injector, and pair of solids storage bins remain isolated from each other.
When the load cell under the primary injector detects a solids loading
of less than 102 kg, 91 kg of solids are automatically transferred from
the pressurized feed injector without interrupting feed to the combustor.
Refilling is usually completed in about 5 minutes. After refilling, the
feed injector is again isolated from the primary injector, vented, and
filled with solids from the storage bins. The feed injector is again
isolated and repressurized to await repetition of another cycle.
Process air for the solids feeding system is provided by an auxiliary
air compressor with a rated capacity of 200 SCFM at 200 psig (5.6
standard m3/min at 1380 kPa gauge). Prior to contacting solids, air
is dried using a Pall Heatless Regenerative dryer.
Fluidized Bed Combusjtor
The combustor consists of a 61.0 cm I.D. steel shell refractory lined
with Grefco #75-28 Litecast to an actual internal diameter of J1.8 cm
(see Figures IV-4 and IV-5). The 9.75 metre high unit is designed in
flanged sections and contains various ports to allow for material entry
14
-------
FIGURE IV-4
COMBUSTOR VESSEL
G> ~ /SO* STEEL K-f: SUP-QtJ FLAMGE.
HO jt4-l6"X.5&2"ivALi X6"tG. STEEL PIPE WITH
^STEE
fJo.36,38 • (B'X-ZSQ'WALL X &"Lq. STEEL P/f>t
WITH &'-3QC* ST£EL f?,f, SUP-ON FLQ
fJO,
&"-3C?C>* ST££L ff,f- SL/P-Qfil FLAtitiE
MATERIAL AfJ
PRESSURE VESSEL CODE.
TO 8£ TSS,T£0 AND A.S-M,£, COOED fOf? ISO P.
15
-------
o
o
a
CO
H HI
O H
90 §
CO
CO (—I
t-d <3
f I
H
21
H
2!
O
don
G HASB££fJ
-------
and discharge. Numerous taps are also provided along the length of
the combustor to monitor both pressure and temperature. Currently pres
sure measurements are made at five locations: immediately below and
above the fluidizing grid, at two points within the expanded bed, and at
one location above the bed. Temperatures within the combustor are con-
tinuously measured via protected thermocouples at twelve vertical loca-
tions extending from 15.2 cm above the grid to the top of the unit, as
well as at one point within the plenum below the grid. Combustion air
to the unit is provided by the miniplant main air compressor having a
capacity of 1400 SCFM at 150 psig (40 standard m3/min at 1030 kPa gauge).
Combustor Startup Burner
The bottom plenum section, where combustion air enters, also houses the
burner used for initially preheating the limestone bed during unit start-
up. Fuel to the burner is provided by a natural gas compressor with a
capacity of 20 SCFM at 200 psig (0.57 standard m3/min at 1379 kPa gauge).
The burner is shown schematically in Figure IV-6. A well-mixed stream
of natural gas and air is ignited as it exits the nozzle using a flame
generated by an auxiliary spark-fired pilot ignitor (see Section dealing
with burner performance for more details regarding the pilot ignitor).
The burner is water-cooled (both skin and cooling-water temperature are
continuously monitored) and is designed to avoid flash-back during opera
operation.
Once the fluidized bed temperature reaches 430°C, a liquid fuel (kero-
sene) system is used to heat the bed to the coal ignition temperature
(»650°C). This system is also used to maintain bed temperature above
650°C should coal feeding be interrupted during normal combustor opera-
tion. Liquid fuel entry is at a point 15.2 cm above the combustor
fluidizing grid. Injection is made through a 1.6 mm I.D. spray nozzle
mounted within a 1.6 cm I.D. pipe resulting in a 0.76 mm annular spacing
at the tip. Air flowing through the annulus provides cooling and
creates a high velocity stream to insure penetration of the fuel into
the bed as well as to avoid clogging of the nozzle.
Combustor Fluidizing Grid
The combustor fluidizing grid, located 71 cm above the combustor bottom,
separates the plenum from the main combustor chamber. The grid consists
of a 1.3 cm thick X 69 cm diameter stainless steel plate containing 776
equally-spaced holes of 0.16 cm diameter within a 31.8 cm circle (see
Figure IV-7). The design is such to give a grid pressure drop of
approximate 30% of the total pressure drop across a 2.5-3.0 metre dsef>
expanded bed. The unit is water-cooled with both the cooling water tem-
perature and metal temperature continuously measured.
17
-------
Cooling
Channel
Porous disc
1/8" x3.62" dia.
1/4" Coppertubing
for cooling water
Entire tube
filled with —
alumina beads
10" Flange
air
Brass grid containing 1900-1/32" holes
within 3.5" dia. circle
99999)0000
yvyyvT^^^^rv .^iv'
iiiifill I f I 11 111"
ooo
Igniter (combination gas flow
and electrode spark)
Burner nozzle, 3.5" sch.40
pipe, 7.5" long
. 40 mesh wire
•Baffle, 1/8" thick, containing
24 x 11/32" dia. holes with
1 x 3/4" dia. center hole
All material 316
stainless steel except
as noted
Fuel sparger
Gas
Figure IV-6 COMBUSTOR PREHEAT BURNER
18
-------
IfJSEKr 14 X /rf X p
n
o
CO
1-3
H
Q
O H
H <
Nl I
W ~J
a
o
pa
M
a
I IG47-&Z-D\
-------
Combustor Cooling Coils
Heat removal from the combustor is provided by cooling coils located in
discrete vertical zones above the grid. Each coil has a total surface
area of 0.55 w~ and consists of vertically-oriented loops constructed
of 0.5 in Schedule 40 316 stainless steel pipe (Figure IV-8). A coolant
distributor plate accommodates each pair of coils with one coil extend-
ing 45.7 cm above the plate while the other extends the same distance
below. The plates, in turn, are sandwiched between combustor flanges
located at 91.5 cm vertical increments beginning 91.5 cm above the
fluidizing grid. Although there is provision for ten coils as shown
in Figure IV-1, only four are currently installed and are located at
the first two flanges above the grid.
Demineralized cooling water is pumped from a storage tank through the
coils where it is partially vaporized. After exiting the combustor,
the steam/water mixture flows through a condenser prior to return to
the storage tank. Flow to each of the coils can be separately monitored
and controlled. Cooling water exit temperature from each coil is also
routinely recorded. In order to simplify measurement of heat transfer
coefficients, it is also possible to increase the water flow to any one
of the coils to prevent vaporization during the time of measurement.
Combustor Solids Rejection and Transport
Solids rejection from the combustor is required to maintain a steady-
state bed height whenever a mixture of coal and limestone is fed. Such
rejection is accomplished through a port located 230 cm above the fluid-
izing grid. From here, solids flow by gravity through a 15.4 cm I.D.
steel pipe, refractory lined to 5.1 cm, into a refractory-lined pulse
pot (see Figure IV-9). Next, they are pneumatically transported by
controlled nitrogen pulses to a pressurized lockhopper (Figure IV-10)
from which they are periodically dumped into metal drums.
The full operational mode of the miniplant involves simultaneous opera-
tion of both the combustor and regenerator. In such a scheme, limestone
would be continuously transferred from the combustor to the regenerator
and back again after appropriate regeneration. Since regenerator shake-
down has not been completed, such a transfer sequence has not been
tested, nor have the final connections between both units been made.
When such operation is carried out, solids transfer between both units
will utilize pulse pots in a manner similar to that currently being used
(and discussed above) for combustor solids rejection. Solids will
exit the combustor through another port located 230 cm above the fluid-
izing grid and will return from the regenerator through a port located
at the same height as the coal feeding point (28 cm above the fluidizing
grid).
20
-------
/&47-7S-D
o
o
CJ
t/3
H *d
O H
o
o
tr1
O 00
n
o
Cfl
-------
1
1
-
1
— -J
_J
\/
COUI^L iHGt \ "A
s
0 H
CO O
CH
hd £2
W H
W <
hd VO
O
H
-------
ijrr-
&UE.VATIOK1S
—asm.
TOP ~ IB" *r*D wJAi-U WfcUJ CAP
SHELL- is" o.t>, x: YA' wKuL
EfJDS WITH fSt>* i-^PTOJU
* POR.F ~ 3 - Q" ar'o fJAj_L F(Pe vJ»TM d"
(50** SI-lP OJO pLjKuSC *• BUljfr F
PORT - ^ - ^" i-P.S, H>>.l_F C-oopl.tOe
PORTS - Sj&^&rfjtQ,* y^ i-RS. HKUJS COUP
* NOTE; FOR POItT /Jo,3 USE J/ls" JM-fco
.S.M.E COt>Et> FOR |S"O
ftfc A.T faOO*F.
I I
LOCK.
ESSO RESEARCH «ID ENGINEERING COMPANY
MECHANICAL DIVISION
LINDEN, N.J.
en
O
r1
H
G
CO
O
M
w
n
a
M
f
o
n
o
TI
nj
@
hrj
M
O
M
-------
Combustor Cyclones
Flue gas and entrained solids (flyash and limestone) exit the top of the
combustor through a 45.7 cm I.D. steel pipe, refractory lined to 18.4 cm
I.D., and enter a two-stage cyclone system (Figures IV-11 and IV-12).
Both cyclones are refractory lined and designed to operate at
combustor temperatures and pressures. Solids (primarily limestone)
separated by the first stage cyclone drop through a 20.3 cm I.D. steel
dipleg, refractory lined to 10.2 cm I.D., and enter a refractory-lined
pulse pot (Figure IV-9). From here, they are pneumatically conveyed
back to the combustor using controlled nitrogen pulses and enter at a
point 66 cm above the fluidizing grid. Solids (primarily fly ash)
escaping the primary cyclone pass through a 30.3 cm I.D. steel pipe,
refractory lined to 14.5 cm I.D., and enter the more efficient second-
stage cyclone where a finer distribution of solids is captured. Col-
lected solids then pass through a 20.3 cm I.D. dipleg, refractory lined
to 10-2 cm I.D., and enter a pressurized lockhopper (Figure IV-10) from
which they are periodically dumped into metal drums. Although not yet
installed, a Ducon granular bed filter with appropriate modifications
will be placed after the second-stage cyclone to capture solids that
manage to escape the cyclone separation system.
Pressure Control and Flue Gas Discharge System
The technique used to control combustor pressure consists in dropping
the system pressure across an appropriately sized ceramic coated
nozzle located in the flue gas exit line. Back pressure is controlled
by regulating the flow of a secondary air stream to the nozzle inlet.
After exiting the second-stage cyclone, flue gas passes through 20.3 cm
I.D. steel pipe refractory lined to 10.2 cm I.D., and is expanded through
a converging nozzle. The 2.5 cm thick nozzle consists of a conical-shaped
entrance section with an initial diameter of 3.5 cm which converges at
an angle of 30° over a distance of 1.0 cm to a 2.3 cm diameter cylindrical
throat of 1.5 cm length. The nozzle is constructed of mild steel and
is mounted in a carbon steel flange. The exposed surfaces of the nozzle
and flange are covered by a 0.1 mm thick flame-sprayed coat of chromium
carbide. In addition, a subcoat of nickel aluminide was flame sprayed
to the base metal surfaces to provide a sound substrate for adherence of
chromium carbide. Future plans call for the entire nozzle to be con-
structed of silicon carbide instead of coated metal as described above.
Nozzle geometry will remain the same except that three different throat
diameters will be used to achieve the correct combustion pressure over
the range of operating variables planned.
The remainder of the combustor pressure control system is located 0.6 m
upstream of the nozzle and consists of a secondary source of high pres-
sure air which is metered through a 5.0 cm Kamyr ball valve equipped with
24
-------
ro
m
O
O
cj
U)
H
O
&
M
H
f«
cn
H
I
O
r1
o
SI
M
H
O
MHHI-PLAHT
f_CYCLONe
/&47-2O-O
-------
n
a
H
O
CO
tt
n
§
o
O
M
n
Kl
O
r<
i
w
M
O
H
<
M
KJ
I647-2I-D '
-------
a penumatic actuator and positioner. By superimposing a secondary flow
of air onto the primary flow of flue gas through the nozzle, control of
combustor pressure is maintained at the desired level which has been
typically 910 kPa,
Downstream of the nozzle, flue gas passes through a 2.5 m 316 stainless
steel schedule 40 pipe, water-jacketed by a 4 in, schedule 40-carbon
steel pipe, and enters a scrubber for final cleanup before venting to
the atmosphere. During transit from the upstream side of the nozzle to
the scrubber entrance, flue gas temperature is reduced from approximately
820°C to 260°C, with 35-50% of the reduction due to expansion through
the orifice and injection of the low temperature secondary air stream.
Flue Gas Sampling and Analytical System
Various problems have occurred in previous systems designed to sample
exiting flue gas. These are discussed in a subsequent section of the
report. Accordingly, discussion here will focus on the flue gas
sampling system currently installed.
Flue gas is sampled at a point about 7 metres downstream of the second-
stage cyclone exit (ca. 4.5 m upstream of the pressure control nozzle).
A schematic of the sampling system is shown in Figure IV-13. The system
is designed to produce a solids-free, dry stream of flue gas at approxi-
mately ambient temperature and atmospheric pressure whose gas composition
is unaltered, except for water content, from that of the original flue
gas. In order to allow for cross-checking and backup of the miniplant's
main analytical instrument train, the system includes provisions for
removing samples for wet chemistry determinations and also direct routing
of a stream to the batch fluidized bed combustion unit analytical
instruments.
The instrument package for analyzing combustor flue gas composition
consists of the following:
Gas Instrument Ranges
S02 Beckman 315A(Infrared) 0-3000 ppm
NO Beckman 315B (Infrared) 0-2500 ppm
0-1000 ppm
0-500 ppm
CO Beckman 315A (Infrared) 0-6250 ppm
0-1250 ppm
C02 MSA Lira M300 (Infrared) 0-25 %
02 Beckman 715 (Ft Electrode) 0-25 %
0-5 %
27
-------
Figure IV-13
FLUE GAS SAMPLING SYSTEM
VENT
tsJ
00
FLOW RATE
0.12 l/s
1
OFF
ANALYTICAL
GAS LINE TRAIN
FLOW RATE
0.3 l/s
NITROGEN J,
PURGE v
WET CHEMISTRY SAMPLING 1 |/s _^«
| 0-1000 kPa 0-150kPa
A 1 1 ^260°C O 0 i
t \ 4 q> V 9 Y i ^
COAXIAL BALLSTON PRESSURE PERMAPURE
HEAT EXCHANGER FILTER REGULATOR MODEL PD 1C
TPAQIMIFT RT^°r LJ 30/12 DOWNSTREAM
I GAS INLET -81 5 C TAPE HEATED PRESSURE VENT
TGAS OUTLET - 80-200°C BALLSTON T0 175-200°C 100 kPa A
FILTER T
20/80 p
TAPE HEATED
TO 175-200°C LtJa —
V
-f^n
u
FILTER
DRYER
100-24
0.48 cm ID
TEFLON TUBINt
30 m
FLOW -0.3 l/s
-------
Output from the instruments is recorded on strip charts every 15
seconds with the exception of 0~ data which are recorded continuously.
Process Monitoringand Data Generation System
With the exception of the output from the analytical system described
above, the remaining data characterizing total system operation are
handled in a common manner. Three multichannel recorders (24-channel
Honeywell Electronik 112) record output from various measuring instru-
ments. In addition, at one minute intervals, the same output is recorded
by a data logger system consisting of a Digitrend 210 data logger with
printer and a Kennedy 1701 magnetic tape recorder with the electronic
interface between the two designed by Automated Technology Corp. Output
from the analyzers will also be recorded by the data logger in the future.
Approximately 60 pieces of data are logged with three-quarters involving
temperature measurement while the rest deal with pressure and material
flow rate. The points logged are given in Appendix E.
Signals from the data logger appear as digital output on printed paper
tape and are also stored on magnetic tape. The magnetic tape, contain-
ing about 3600 items of data per hour of run time, is fed to a computer
which converts the logger output to flow rates, pressures, etc with the
proper dimensions. The data are then averaged and standard deviations
calculated over preselected time intervals (usually 10 min)- Other
quantities are also calculated. This includes average bed tempera-
ture, based on four thermocouple readings covering the 15-114 cm interval
above the fluidizing grid, superficial gas velocity, and excess air.
Combustor Safety and Alarm System
A process alarm system was designed to warn of impeding operational pro-
blems. Two general alarm categories exist. The first, dealing with less
critical situations, alerts the operator of the problem so that appro-
priate corrective action can be taken. The second class of more critical
alarms results in the immediate or time delayed shutdown of the complete
system or specific subsystems. A brief description of the alarms is
presented in AppendixD-
Fluidized Bed Regenerator System
Since shakedown of the regenerator system has not been completed,
discussion of the regenerator will be abbreviated.
The regenerator reactor, designed for operation at 1100°C and pressures
up to 1010 kPa, consists of a 45.7 cm I.D. steel shell refractory-lined
with 75-28 Grefco Litecast to an internal diameter of 21.6 cm (see
Figure IV-14). Numerous taps are provided along its 6.66 m overall
height to monitor both temperature and pressure, while appropriately
located ports allow for material entry and exit.
29
-------
FIGURE IV-14
REGENERATOR VESSEL
i, ALL fOKT f=~i.Ah/!j€ BOLT t-tOL&S To
STKADDLK tf£RTtCAL 4
M4itJ FLAMES BOLT HOL£& &M £
30
-------
Reducing gas for the regeneration step is produced just above the
fluidizing grid by the partial combustion of fuel (natural gas) which
is injected directly into the bed. The natural gas is provided by the
same compressor (20 SCFM capacity at 200 psig) which services the
combustor. Plans are underway to install an independent large capa-
city compressor (43 SCFM at 200 psig) to allow simultaneous operation
of both units at typical miniplant operational loadings. A burner
located in the regenerator plenum provides additional heat for the regen-
eration reaction. The burner is identical to that installed in the
combustor.
The water-cooled fluidized grid, located 50.8 cm above the regenerator
bottom separates the plenum from the main regenerator chamber. Its
design, similar to that used in the combustor consists of a 1.3 cm thick
X 53.3 cm diameter stainless steel plate containing 392 equally-spaced
holes of 0.20 cm diameter within a 21.6 cm circle.
Ports located 25 cm and 145 cm above the fluidizing grid will serve to
receive and discharge solids, when transport between combustor and
regenerator becomes operational. Desired bed levels in both units will
be achieved through a process control package involving differential
pressure transmitter circuits to monitor bed level coupled with appro-
priate regulation of the two pulse pot feeders operating between the
combustor and regenerator.
Provision is also made to inject a secondary source of air (or oxygen)
near the top of the bed to create an oxidizing zone. Past experience
with the batch unit has indicated that this may be required to mini-
mize the formation of CaS byproduct.
Gases exiting from the regenerator pass through a 20.3 cm I.D. steel
pipe, refractory lined to 5.7 cm I.D., and enter a single-stage refractory
lined cyclone (Figure IV-15). Collected solids drop through a 15.4 cm
I.D. steel dipleg, refractory lined to 5.1 cm I.D. and enter a pres-
surized lock hopper from which they are periodically discharged to metal
drums. Cyclone effluent gas is cooled by a heat exchanger and sent to
a scrubber for final cleanup before atmospheric discharge.
Regenerator off-gas is to be sampled after the pressure reducing valve,
will be filtered, dried and sent to analyzers at approximately ambient
temperature and pressure. Provision for cross-checking and backup of
the main instrument package will also be included. Analysis of off-gas
composition will be made by the following instruments!
31
-------
U)
ro
o
O
f
i
td
H
o
-------
Gas Instrument Ranges
S02 Beckman 315A (Infrared) 0-15 %
CO Beckman 315B (Infrared) 0-30 %
0-10 %
0-2.5 %
CO Beckman 315B (Infrared) 0-1.0 %
0-0.5 %
C09 Beckman 337B (Infrared) 0-20 %
0-5 %
0? Beckman 715 (Pt Electrode) 0-25 %
0-5 %
Temperature, pressure, and material flow rate data characterizing
regenerator operation" will be handled in the same manner as previously
discussed for the combustor using multichannel recorders and the data
logger system. The points logged are listed in Appendix E. Final
processing of data via computer will parallel the scheme previously dis-
cussed for the combustor.
Various safety alarm systems have been incorporated into the design of
the regenerator control system as was done for the combustor. Depending
on the particular problem, these either trigger an automatic system res-
ponse (immediate or time-delayed) or simply alert the operator so that
corrective measures can be initiated. Those conditions which actuate
alarms are given in Appendix D.
Miniplant Support Structure
The flanged steel beam support structure for the miniplant is 12.9 m high,
9.0 m wide, and 3.9 m deep. Three platforms at 2.4 m, 5.7 m, and 9.0 m,
with connecting stairways, are provided for servicing the miniplant. The
structure rests 18.4 cm off the ground on concrete support pillars. The
reactors are supported by the structure at the first platform. Thermal
expansion joints are provided at various locations to accomodate thermal
expansion of the reactors in the vertical direction. A five-ton bridge
hoist, mounted on top of the support structure, is used for assembly
and disassembly operations.
MATERIALS
Coal
Coal used during the miniplant combustor shakedown was a high volatile
bituminous coal obtained from Consolidation Coal Company. In Runs
2-12.2, the coal was obtained from the Arkwright Mine in West Virginia.
33
-------
In subsequent runs, the coal was obtained from the Champion preparation
plant in Pennsylvania. Both coals are Pittsburgh No. 8 seam coals and
have similar analyses. Grinding and sizing was done by Penn-Rillton
Company. Essentially all of the coal was less than 2380 ym (No. 8 U.S.
Mesh) in size. Actual size distributions used during the course of the
program are given in Figure IV-16. Two size distributions were used
during shakedown. The original batch used in runs 2 to 12.2 contained
all the fines. Batches prepared for subsequent runs had most of the
fines smaller than 40 mesh removed. This gave a distribution more
closely resembling that expected to be used in commercial FBC units.
Composition analyses are shown in Table IV-1.
Limestone
The only sorbent used for the miniplant runs was uncalcined limestone
obtained from Grove Lime Company (Stephen City, Va) designated as Grove
No. 1359- This material, in its calcined form, contained 97.0 wt. %
CaO, 1.2 % MgO; 1.1 % Si02, 0.3% A1203, and 0.2% Fe203. The uncalcined
limestone feed was screened to give a distribution with a minimum of 90%
between 2380 \im (No. 8 U.S. Mesh) and 841 ym (No. 20 U.S. Mesh). Actual
particle size distribution of the limestone bed during the course of a
run would be shifted into the smaller particle size range due to attrition.
TABLE IV-1. MINIPLANT COAL ANALYSES
Component
Coal
Run No.
Moisture
Ash
Total Carbon
Hydrogen
Sulfur
Nitrogen
Oxygen (by difference)
Chlorine
Higher Heat Value (BTU/lb)
Arkwright
1-12.2
1.0
8.1
76.5
5.3
2.6
1.5
5.0
0.1
Weight Percent
Arkwright
13.1-15.4
0.9
7.4
77.1
5.1
2.5
1.1
6.0
0.1
Champion
16.1-19.3
2.2
8.8
78.1
5.1
2.2
1.6
4.1
0.1
14100
13700
13600
34
-------
Figure IV-16
COAL PARTICLE SIZE DISTRIBUTION
LU
M
CO
LU
O
1-
o:
z
:r
H-
LU
_J
2
CO
0^
1-
-J-
o
LU
J UU
90
80
70
60
50
40
30
20
10
n
^^5-
^X* A .X**^
/* A^X?"
>"* S .*r
/ / •
• ^ XA/
/ss
//
7A y-
/ /
4 ;
- / - •/
- / /•-•
• X •
- i /
>/ A^B'* i i i i i i i i i
0 200 400 600 800 100012001400160018002000
1 1 1 1 III! 1 1
100 5040 30 20 18 16 14 12 10
• RUNS 1-12.2 (ARKWRIGHT COAL)
• RUNS 13.1-15.4 (ARKWRIGHT COAL)
A RUNS 16.1-19.3 (CHAMPION COAL)
i i i
2400 2800 3200 (MICRO
i i i |
8 7 1/8" 6 (MESH
PARTICLE SIZE
-------
PROCEDURES
Combustor Startup
Prior to initiating a run, a detailed checkout procedure is followed to
insure that the system is ready for operation. These include such things
as various equipment checks, calibration of flue gas analyzers, activation
of process monitoring and control systems, and turning on all cooling
water systems. All runs were begun with an initial bed of limestone in
the combustor. This consisted of either a fresh charge of uncalcined
stone or the bed from the previous run.
The first operation of startup involves preheating the limestone bed
using natural gas and then by kerosene. Prior to ignition of natural gas,
an air flow of about 350 SCFM (9.9 standard m^/min), or about half that
used at normal operating conditions, is fed through the burner while com-
bustor pressure is raised to 280 kPa gauge. Once ignition of the natural
gas occurs, this procedure maximizes incoming gas temperature under con-
ditions which allow good natural gas combustion and adequate bed fluid-
ization. Ignition begins by simultaneously feeding 20 SCFM (0.57 standard
m-Vmin) of natural gas through the burner while activating an ignition
electrode.
Because of the limited capacity of the gas compressor, natural gas is
used only to heat the bed to a temperature of about 430°C, sufficient to
insure self-ignition of kerosene. This generally requires 20-30 minutes.
At this point, kerosene is injected into the lower portion of the bed.
When rising temperatures indicate ignition of liquid fuel, natural gas
feed is discontinued to insure sufficient air for complete combustion of
kerosene. Approximately 10-15 minutes are required to raise the bed tem-
perature to 650°C which is sufficient to achieve self-ignition of coal.
Coal, usually mixed with limestone, is then fed to the combustor from
the primary injector. A steady stream of 60 SCFM (1.7 standard nH/min)
of transport air is used to convey coal into the combustor. Actual rate
of coal injection is determined by the pressure differential between the
injector and combustor. The rate is initially set at an appropriate value
based on past experience under similar operating conditions. Once ignition
of coal is verified by rapidly rising temperatures, kerosene flow is stop-
ped. At this time, the main combustion air feed line to the plenum is
opened allowing most of the air to bypass the burner, and both combustion
air flow rate and combustor pressure are rapidly increased to their
designated operating values. Flow of water to each cooling coil is
adjusted to maintain steam/water exiting temperatures of 138-150°C.
Once the desired bed temperature has been reached, it is held approxi-
mately constant by the automatic coal feed rate control system.
36
-------
Combustor Shutdown
A run is terminated by first discontinuing coal feed which results in a
rapid decrease in bed temperature. As temperature falls, fluidizing air
flow rate and combustor pressure are decreased stepwise. When tempera-
ture falls below 90°C, which generally requires 10-15 minutes, air flow
is halted and the combustor is depressurized. At this time, remaining
systems, including cooling water flows, are shut down and a nitrogen purge
is introduced into the combustor to prevent condensation of moisture.
UNIT PERFORMANCE
Length of Operation and Conditions Tested
During the shakedown phase a total operating time of 500 hours was
accumulated and 37 runs conducted. Many of the initial runs were
limited to 10 hours or less duration, but as the systems were improved
the run durations increased. Included among the longer runs were four
of 24 hours duration, one 50 hours in duration and a 100 hour contin-
uous run which culminated shakedown.
The miniplant was demonstrated to be capable of meeting its basic
peak design conditions. As indicated in Table IV-2, operating levels
such as pressures of 1020 kPa (10 atm.), superficial velocities of
3.2 m/sec (10.5 ft/sec) and temperatures of 980°C (1800°F) were
achieved. The coal feed rate and combustion heat release rates reached
to date are somewhat lower than designed due to the reduced heat trans-
fer surface area presently installed in the combustor (62% less than
the maximum).
High combustion intensities (approx. 5 MW/rn^ exp. bed) were achieved
while maintaining excellent temperature profiles and without signi-
ficant agglomeration problems in the combustion zone. Operation at
full design conditions and peak coal feed rates will be accomplished
at a future date.
Under typical test conditions the superficial gas velocity and bed
temperature were maintained at values somewhat lower than the maxi-
mum levels. The superficial velocity was normally in the vicinity
of 1.8-2.1 m/s. Temperatures were typically in the 870-950°C range.
The expanded bed depths were usually maintained at approximately 3 m
which, was sufficient to immerse the entire heat transfer surface in
the expanded fluidized bed. In some runs lower bed levels were
maintained to reduce the immersed cooling surface, thereby, reducing
the coal feed rate and enabling operation at higher excess air levels.
In this manner an excess air level as high as 130% was achieved. In
all runs an Eastern Pittsburgh Seam Coal was burned in the presence
of Grove No. 1359 limestone. Alternate coals and sorbents will be
tested in future runs. The test conditions for each of the runs is
given in Appendix G.
37
-------
TABLE IV-2. MINIPLANT FLUIDIZED BED COMBUSTION CONDITIONS
Pressure (kPa)
Temperature-lower zone
Superficial gas veloci
Excess air (%)
Bed depth - static (m)
Bed depth - expanded (m)
Coal feed rate (kg/hr)
Combustion heat release (MW)
MW
Combustion intensity (—* ;—:—;—r
mr expanded bed
Coal
Sorbent
Design
1010
'. (°C) 950
.ty (m/s) 3.0
15
1.2
m) 4.6
218
;6 (MW) 1.8
MW , 5. 2
Test
Range
405-1020
815-980
1.5-3.2
10-130
0.8-2.0
1.3-3.7
90-155
0.8-1.3
^5
Typical
930
870-950
1.8-2.1
10-30
1.4-1.6
^3
120-140
1-1.2
*>5
Pittsburgh seam
-7 +40 mesh
Grove No. 1359 limestone
-8 +20 mesh
38
-------
Summary of Operating Results
Details of the equipment performance and combustion results from the
miniplant shakedown are reported in other sections of this report.
This section briefly summarizes these results.
Good control of each of the operating variables was demonstrated for
sustained periods. Standard deviations typical during such times are
given in Table IV-3.
TABLE IV-3. CONTROL OF OPERATING VARIABLES
Variable Typical Value Stand. Dev.
Bed temperature 900 + 10°C
Coal feed rate 130 + 9 kg/hr
Combustor pressure 930 + 10 kPa
Gas superficial velocity 2.0 + .05 m/s
Bed height (as determined + 5% of total AP
by the pressure drop ~
across the bed)
Bed temperatures were well controlled using the cascade control sys-
tem described on page 43. In this system, coal feed rates are adjusted
to maintain constant combustor bed temperatures. Earlier attempts
to maintain a constant coal feed rate resulted in temperature varia-
tions in the combustor and this approach was abandoned.
The combustor pressure was well controlled using the fixed converging
nozzle with supplementary air addition as described on page 24. This
system replaced an earlier system which used a pressure control valve
in the flue gas line. The pressure control valve was susceptible to
damage by erosion and that approach was dropped. Steady gas velocit-
ies were maintained by achieving good control of the air flow rate,
pressure and bed temperature.
The combustor bed level was well controlled by continuously removing
solids from the combustor through a "pulse pot" as described on page
20. The pulsing rate was adjusted to maintain a constant bed level
by maintaining a constant pressure drop across the bed.
Very good temperature profiles were established after the installation
of vertical cooling coils. The temperature variation across the bed
was almost entirely due to a hot spot caused by the large volumetric
heat release in the lower zone of the combustor near the point of coal
™anC^°?' However> the hot sPQt temperature was generally less than
JO C higher then the average temperature across the remaining bed. Flue
gas exit temperatures were typically close to the temperatures at the
top of the expanded bed.
39
-------
Combustion efficiencies with flyash recycle from the first cyclone
ranged from 93 to 97% over a range of excess air levels from 20 to
110%.
A limited amount of SC^ emission data were obtained through the shake-
down period due to problems in developing a reliable S02 sampling sys-
tem. The data do, however, suggest that the EPA S02 emission standard
of 1.2 Ib SC>2/M BTU can be readily met with limestone bed material.
Prior to run No. 19.3, a reliable SC>2 sampling system was installed.
The S02 emission in run 19.3 for a Ca/S molar ratio of only 1.45 was
720-780 ppm or 1.3 Ib S02/M BTU, (60% removal) just slightly over
the allowable emission limit.
NOX emission levels were measured and correlated as an increasing
function of the excess air level. Emissions ranged from 50-250 ppm
over excess air levels from 10-110%. The maximum NOX level corresponds
to an emission of, 0.4 Ib N02/M BTU, well below the EPA standard of
0.7 Ib N02/M BTU.
CO emissions were typically below 100 ppm, indicative of good combus-
tion.
Performance Characteristics of the Miniplant
A fluidized bed combustor such as the miniplant, which relies on
immersed cooling surface cooled by water to remove heat and thereby
control the combustor temperature, has certain operating restrictions.
To begin with, the amount of heat removed by the cooling coil is
determined solely by the surface area of the immersed coils, the bed tem-
perature and the heat transfer coefficient on the outside of the coils.
A change in the flow of cooling water has very little effect on heat re-
moval. This is due first, to the fact that the resistance to heat
transfer inside the tubes is very low and does not influence heat transfer
rates significantly. Secondly, the temperature difference between the
combustor bed and the interior of the cooling coils is relatively
insensitive to changes in the flow of cooling water entering the coils.
As a result, at a given set of operating conditions i.e., temperature,
pressure, fluidization velocity, particle size, etc., the only para-
meter which affects heat removal rates is the surface area of the cool-
ing coils immersed in the fluidized bed. Once the surface area is
fixed, the heat removal rate is also fixed and the only parameter which
can be varied to control bed temperature is the coal feed rate.
If saturated or superheated steam is used as the cooling medium, the
situation is different. In that case, increasing the flow of steam
can increase the temperature difference between the bed and the cooling
medium. Therefore, varying the flow rate of the steam will exert some
40
-------
measure of bed temperature control. Since saturated or superheated
steam was not used as the cooling medium in the miniplant, this means
of temperature control was not available and changing the coal feed
rate was the method used to control the bed temperature.
It can also be seen that fluidization velocity and excess air are not
independent variables. Once the coal feed rate has been adjusted to
give the desired bed temperature, the air feed rate can be adjusted
either to the desired excess air level or the desired fluidization velo-
city. Fixing one sets the other. If it is desired to study the effects
of fluidization velocity and excess air independently, this can only be
done by changing the heat transfer area immersed in the fluidized bed.
This is best done by removing (or adding) cooling coils. It can also
be done by dropping the expanded bed level so only a portion of the cooling
coils.are covered by the expanded bed. This is a less desirable method
since it results in a change in bed depth. It also results in a lower
flue gas outlet temperature which, in some cases, may not be desirable.
Once the heat transfer surface area has been changed, fluidization velo-
city and excess air level will still be dependent, one on the other,
but at different values than before.
EQUIPMENT PERFORMANCE
Coal Feeding
The performance of the pneumatic transport coal injection system was
originally unsatisfactory and had to be improved considerably (see Figure
IV—3 (Equipment Section)). The system was sensitive to plugs, par-
ticularly at the orifice of the primary injector. The orifice assembly
was redesigned to have no internal discontinuities such as sudden
decreases in diameter. Coal had to be dry and be free of any foreign
material which might plug the orifice. There was also difficulty in con-
trolling Ap between the primary injector and combustor during refilling
of the primary injector. This problem was alleviated by modifications
in design and attentive monitoring of the unit during the primary in-
jector filling operation.
The original orifice assembly contained a 3/4 inch (nominal) ball valve,
a 0.95 cm diameter orifice, and a tee for mixing coal with transport
air. A number of different sized pipe fittings were used in the assembly,
which resulted in a momuniform internal diameter. Performance was
erratic and unsteady coal flow and plugging occured frequently. The
entire orifice assembly was therefore redesigned, using a larger valve
and piping. A sketch of the new assembly is given in Figure IV-17.
Care was taken to be sure that the diameter of the orifice assembly was
uniform from the bottom of the hopper to the orifice. This diameter
was the inside diameter of the one inch ball valve (2.1 cm). At the
orifice, the diameter decreased to 1.3 cm. Although this orifice was
larger than the original 1.0 cm orifice, coal feed rates could be kept
41
-------
BUSHINGJAPERED
TRANSPORT
AIR
COAL VESSEL
BALL VALVE, 1 INCH
UNION WELDED TO VALVE
UNION (3/4 INCH)
pyfyfy///////////////////////~frfr)fy?' > '^'Vswv'*
\ \
MIXING TEE CONNECTOR, TAPERED
Figure IV-17
COAL FEEDER ORIFICE ASSEMBLY
42
-------
the same by increasing slightly the flow of transport air and/or
decreasing the AP from the primary injector to combustor. The larger
orifice was much less susceptible to plugging. In fact, performance of
the new orifice assembly was significantly improved over the original
assembly; coal feeding was steady and plugging became very infrequent.
When plugging did occur, it was usually caused by foreign material,
such as a piece of scale from the coal hopper, which had become lodged
in the orifice.
After flowing through the orifice, coal dropped into the mixing tee,
where it was picked up by a stream of transport air and conveyed through
the injection line into the combustor. Again, it was of the utmost
importance that the change in diameter from the mixing tee to the
injection line occurred gradually. It was found best to avoid sharp
turns in the transport line because doing so both decreased the chance
of plugging and minimized erosion in the line itself. A tee was used
to prevent erosion where a sudden turn was necessary (Figure IV-18);
however, tees were eventually replaced with gradual bends in the trans-
port line (1/2 inch pipe).
During most of the work, the orifice assembly was vibrated fairly
intensely to minimize plugging, however this caused sections of the
orifice assembly and transport piping to eventually come loose or some-
times even fracture from fatigue. After several runs were made with
the larger orifice, the vibrator was turned off, causing no apparent
deleterious effect on coal feeding. As a result, vibration of the ori-
fice assembly is no longer being used.
When the primary coal injector was nearly empty, coal was transferred to
it from the feed injector. Since coal feed rate to the combustor was
very sensitive to changes in AP between the primary injector and com-
bustor, AP had to be carefully controlled and monitored during transfers.
AP is normally controlled by venting air from the primary injector at a
constant rate and adding air through a control valve. However, coal
entering the primary injector during transfers carried with it transport
air, so that there was a tendency for overpressuring the injector. There-
fore, in order to relieve some of this pressure, additional air was vented
during transfers through a second valve. The system worked satisfactorily
if this second valve was sized properly and did not plug. With regard to
plugging, a ball valve was found to be much better than the globe valve
which was originally used. However, attention by an operator was usually
necessary to make sure that the AP stayed within the desired range. Some
changes to the transfer system are planned in order to reduce the degree
of attention required by the operator.
Temperature Control
The miniplant combustor is operated at bed temperatures from 815
to 955°C (1500 to 1750°F), with temperatures monitored at many elevations
within the combustor. The combustor temperature is primarily controlled
by adjustment of the rate of coal injected into the fluidized bed.
43
-------
FLOW-
PACKED COAL PROTECTS TEE
Figure IV-18
COAL FLOW THROUGH A TEE
-------
A thermocouple In the lower zone of the combustor (at port No. 7) 47 cm
(18 inches) above the fluidizing grid is used as the sensor to con-
trol temperature of the combustor, and the set point temperature at this
location regulates the coal feed rate to the combustor.
As explained on page 10, the coal feed rate to the combustor is regulated
by the pressure differential between the primary coal vessel and the
combustor - the greater the pressure difference, the greater the feed
rate. A sketch of the arrangement is presented in Figure IV-19.
Temperature control is accomplished by a cascade type control loop using
two controllers, one for temperature set point and another for AP set
point. A deviation of the desired (set point) temperature from the
actual combustor temperature (at port No. 7) causes a signal to be
transmitted by the temperature controller to the AP coal feed rate con-
troller. This error signal actually resets the set point of the AP
controller so that a different AP will be established between the coal
vessel and the combustor.
This change in the pressure difference between the coal vessel and com-
bustor causes a change in the coal feed rate which will tend to return
the bed temperature to the desired (set point) value.
Proper tuning of the controllers was necessary for optimum reaction to
system perturbations and anticipation of changes in bed conditions. This
control system has performed very satisfactorily and provides excellent
temperature control and response.
Cooling Coils
The miniplant was operated with cooling coils of two different designs
during shakedown and a third design is currently being used.
Each of the original coils had a horizontal serpentine configuration
with a 5.7 cm horizontal pitch (Figures IV-20 and IV-21). The materials
of construction were 3/4 inch type 316 seamless tubing (0.049 inch
wall) for the straight sections and 3/4 inch type 316L stainless steel
tubing for the 180° U-bends. Ten coils, each covering 0.45 m in the
vertical dimension were originally installed. They extended from a
height of 0.45 m above the fluidizing grid to an elevation of 5m. The
surface area of a coil was 0.58 m^.
These coils were susceptible to fatigue at the inlet and outlet piping
and were readily deformed when bed agglomeration occurred. Modifications
in the number of coils, their size, and orientation were made after a
number of instances of cooling coil damage. Because of their configura-
tion the coils had a high packing density in the cross sectional area
of the combustor, limiting the vertical movement of solids and resulting
in poor temperature profiles.
45
-------
COAL INJECTION
AP CONTROLLER
II I
r*
| AP
| INPUT
| SIGNAL
I
, I
ERROR SIGNAL
ADJUSTS
AP
CONTROLLER
SET POINT
PRESSURIZING
AIR
1
TEMPERATURE
CONTROLLER
-
-
t.-
E.
'
~
1 I 1
TEMP.
INPUT
SIGNAL
PRIMARY
COAL
FEED
VESSEL
1
^
f*
\
TRUL TRANSPORT
PRESSURE/ A,R
1
1
1
1
A P
CELL
^
^ — v.
r-/~X
y ^
\ /
\ /
\
N
/
~u
THERMO-
COUPLE \
PORT #7
^^_ \
^^
CO
Al
c
0
M
U
S
T
0
R
,
'—'&•''
* •
t
4<
t *
\ 2ocr
Figure IV-19
COMBUSTOR TEMPERATURE CONTROL SCHEMATIC
NOZZLE
CM
GRID
-------
con ASSEM&L
ru
o
H H
> O
O
r1 <
H I
^ NJ
o o
n
o
-------
FIGURE IV-21
HORIZONTAL COOLING COILS
48
-------
The second generation coils were constructed with a vertical configura
tion to achieve better top to bottom mixing (see Figures IV-22 and
IV-23). These coils were installed after Run 15.4. The loops of the
0.45 m high coil were distributed as uniformly throughout the cross
section of the combustor as possible to promote uniform fluidization
patterns. The area of one of these coils was 0.55 m2. The materials
of construction were 3/4 inch 316 tubing for the straight sections and
3/4 inch 316L tubing for the U-bends. Schedule 40 stainless steel pipe
used for the inlets and outlets gave increased support. In addition,
reinforcing braces and bands were used to hold the coil rigid." Because
of the desire to operate at moderate conditions with all coils immersed
in the fluidized bed only the four lowest coils were installed.
The use of these coils did result in significant improvements in the
temperature profile, however, the coils were still susceptible to damage.
Third generation coils of a vertical configuration were constructed
from heavy wall pipe for added sturdiness and installed after Run 18.3.
The design is discussed on page 20.
The configuration, number and orientation of the coils used in the various
miniplant runs is documented in Table IV-4.
Temperature Distribution
A representative coal feed rate for the miniplant has been 145 kg/hr
(320 Ibs/hr), which implies a combustion heat release level of 1.2 MW
(4,200,000 BTU/hr). The major portion of the heat liberated occurs in
a combustion zone near the coal inlet port and this heat must be imme-
diately transferred to the internal heat transfer surfaces by the rapidly
circulating fluidized solids in order to achieve a uniform bed
temperature.
In freely fluidized beds, the recirculation rates are generally rapid
enough to achieve minimal temperature gradients, even with highly exo-
thermic reactions. However, it was found that the miniplant horizontal
coil configuration constrained the solids mixing and recirculation to
a degree that adversely affected the temperature profile. Localized
high temperatures and large temperature gradients were experienced with
the original serpentine horizontal coils (Figure IV-24). This non-
uniform temperature distribution limited coal feed rates, caused low
flue gas temperatures, created temperature control difficulties and
tended to promote bed agglomeration in the high temperature zone.
The coils when viewed from the top were optically dense which suggested
they were severely hindering the movement of solids and bubbles in the
vertical direction. Therefore some of the coils were shortened and
orientated on their side to act as vertical coils and to open the cross
sectional area. A dramatic improvement in the profile was observed and
subsequently the vertical coils shown in Figure IV-22 were installed.
The temperature gradient decreased to 14°C/m or less as illustrated in
Figure IV-25. The increasing temperature profile in the freeboard volume
shown in Figure IV-25 occurred because the high circulation rates in the
49
-------
HI
H
O
tr*
M
3
O
n
o
M
H
O
ro
-------
FIGURE IV-23
VERTICAL COOLING COILS
51
-------
TABLE IV-4. MINIPLANT COOLING COIL MODIFICATIONS
Runs In Use
I - 4.3
5.1 - 5.4
6.1 - 9.1
10.1 - 13.3
14.1 - 15.4
Configuration
Horizontal Serpentine
Horizontal Serpentine
Horizontal Serpentine
Horizontal Serpentine
Horizontal Serpentine
Size, Orientation and Number
10 coils, each with an area of
0.58 m2
Coils 1A-4B, 5B, each with an area of
0.58 m2. Coil 5A removed
Coil 1A - 0.29 m2, Coil IB - 0.27 m2
Coils 2A-4B, 5B - 0.58 m2 each
Coil 5A removed
Coil 1A - 0.29 m2, Coil IB - 0.27 m2
Coil 2A - 0.47 m2, Coil 2B - 0.56 m2
Coils 3A, 3B - 0.58 m2 each
Coils 4A, 4B, 5A removed
Coil 5B - no water flow
Coil IB was orientated with the
straight tubes in a vertical direc-
tion to act as a vertical coil
Coil 1A - 0.29 m2, Coil IB - 0.27 m2
Coil 2A - 0.30 m2, Coil 2B - 0.31 m2
Coil 3A - 0.23 m2, Coil 3B - 0.58 m2
Coils 4A, 4B, 5A removed
Coil 5B - no water flow
All coils except 3B & 4A were
orientated with the straight tubes
in a vertical direction to act as
vertical coils
-------
TABLE IV-4. (Continued) MINIPLANT COOLING COIL MODIFICATIONS
Runs In Use
16.1 - 17.1
18.1 - 18.3
19.1 - 19.3
Configuration
Vertical (second generation)
Vertical (second generation)
Vertical (third generation)
Size, Orientation and Number
Coils 1A, IB, 2A, 2B - 0.55 m2/coil
Coils 1A, IB - 0.55 m2/coil
Coils 1A, IB, 2A, 2B - 0.55 m2/coil
Note: The coil number indicates the flange the coil was mounted on. Flanges 1,2,3,4,5
are located at heights above the fluidizing grid of 0.915 m, 1.83 m, 2.74 m,
3.66 m and 4.57 m, respectively. Coils A extend below the flange, coils B extend
above the flange.
-------
Figure IV-24
BED TEMPERATURE PROFILE
HORIZONTAL COILS
IUUU
900
800
700
600
500
400
300
200
100
n
i
IMMERSED
COILS
(EXPANDED BED HEIGHT)
—
—
l 1
TEMPERATURE AVERAGE FOR 10 MINUTE INTERVAL
RUN #7 DATE 8/1/74
COMBUSTOR PRESSURE: 910 kPa
SUPERFICIAL VELOCITY: 1.89m/s
SETTLED BED DEPTH: 1 .52 m
•l TEMPERATURE GRADIENT: 92°C/m
• """^ — •__
EXPOSED
COILS
l I
FREE BOARD
l 1 1 1
o
o
LU
o:
<
D£
LU
Q_
LU
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
HEIGHT ABOVE THE FLUIDIZING GRID , m
-------
Figure IV-25
BED TEMPERATURE PROFILE
VERTICAL COILS
o
o
uT
i-
tf
&
LU
Q_
2
LU
h-
-* \s \s \s
900
800
700
600
500
400
300
200
100
o
• • t
•'•*•' • •— •
—
IMMERSED
COILS
—
FREE BOARD
TEMPERATURE AVERAGE FOR 10 MINUTES
RUN # 19.3 DATE 8/4/75
COMBUSTOR PRESSURE: 930 kPa
SUPERFICIAL VELOCITY: 1.9 m/s
SETTLED BED HEIGHT: 1.58 m
TEMPERATURE GRADIENT: 14 °C/m
i i i l i i i i
0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9
HEIGHT ABOVE THE FLUIDIZING GRID, m
-------
bed caused the elutriation of some partially combusted coal. Flue gas
temperatures comparable to the combustion zone temperature were thus
achieved. In addition the circulation rates in the combustion zone were
high enough to avoid a hot spot problem.
A transparent cold model unit with a diameter of 15.2 cm was used to
investigate qualitatively the effect of the baffle configuration on
solids mixing. Motion pictures were taken of the mixing in the presence
of simulated horizontal and vertical coils.
In the cold model, the top to bottom mixing in a bed with horizontal
baffles (simulated coils) was found to be relatively poor. Because of
the small diameter of the unit, the bubbles which formed filled much
of the cross-section and a dampened slugging action ensued. Much of
the vertical movement proceeded in small steps corresponding to the
periodic passing of a bubble. Eddies of solids near the wall and on the
baffles were relatively motionless between slugs. The transport of
solids from the well fluidized section of the bed below the coils to
the upper bed was poor. This was due to the slugs which impacted on the
bottom of the lower baffle impeding the solids movement to the upper
bed. This boundary may in part explain the hot spots observed in the
combustion zone of the miniplant when horizontal coils were used.
The slugging action and top to bottom mixing was far more vigorous with
vertical baffles. At velocities lower than 1.5 m/sec entire portions
of the bed were carried up the column before the slug disintegrated.
At higher velocities the slugging became much more frequent and the
slugs were smaller and disintegrated faster. The mixing behavior
approached that anticipated in the turbulent regime where the movement
is very swift and random and the slugging action dissipates.
The cross-sectional area of the miniplant is greater than the area of
the cold model unit by a factor of 4.3. In the miniplant, mixing will be
influenced to a lesser degree by slugging action. The size of the cold
model is closer to that of the batch unit and the mixing patterns may
be more representative of what is occuring in the batch unit.
A model was developed to predict temperature profiles in fluidized beds
with internal baffles (2). Estimates of the solid recirculation rates
and of the cross flow rates between the bubble wake and emulsion phase
were obtained using the concepts of the bubbling bed theory. One con-
clusion was that the length of the free path of bubbles traveling through
a baffled volume is of critical importance. The solids transport in the
vertical direction is carried in the wake of bubbles. A bubble which
impacts on a baffle undergoes destruction and rebirth, and in the pro-
cess the wake is also disrupted. With a coil configuration which has
a small vertical spacing between tubes, the wake is frequently dis-
rupted, creating high solids cross flow rates, short mixing lengths
and poor top to bottom solids recirculation rates. The model was able
56
-------
to reproduce fairly well the temperature profile obtained in the
miniplant with horizontal serpentine coils using a mixing length of
7-8 cm, the tube spacing in the vertical direction. Reasonable agree-
ment also was obtained for a vertical coil configuration when the
mixing length was made equal to the vertical distance between coil
bends, i.e., the vertical coil length.
Cooling Coil Damage
The cooling coils used to extract the heat of combustion were damaged
in a;number of instances as listed in Table IV-5. The causes were
varied, but include fatigue, deformation, erosion and erosion/corrosion.
The original horizontal coils were susceptible to fatigue. These coils
were constructed of 3/4 inch O.D. 316 stainless steel tubing and were
of a 0.46 m high horizontal serpentine configuration (see Figure IV-26).
They were supported solely by the inlet and outlet tubing extending
through the refractory lined flanges. The strength of the tubing was
not sufficient to hold the coil rigid. The forces exerted by the weight
of the coil and the slugging bed were thus able to fatigue the inlet and
outlet lines and resulted in the failures during Runs 9.1 and 13.3.
In a number of instances the water entering the bed after a coil failure
caused agglomeration of a portion of the bed. The agglomerates would
impact with considerable force onto the coils causing severe deformation.
Figure IV-26 shows the damage which coils 3A and 3B sustained during Run
13.3. The impact of the agglomerated bed forced the coils to deform in a
direction perpendicular to the support rods. The buckled inlet where the
failure occurred is to the left of center in the picture.
Samples from the damaged coil removed after Run 4 were sent to Exxon
Engineering Technology Department, Materials Engineering Division for
metallurgical examination. The coils had been in service for a total of
60 hours. No evidence of corrosion or deterioration of the coil was
detected. The circulating cooling water maintained the metal temperature
of the coil at an adequately low temperature to prevent sensitization and
corrosion attack from the flue gas.
Instances of coil damage due to a high temperature erosion/corrosion
attack were observed after the bed heat up procedure was modified. In
Runs prior to No. 15.2 cooling water flow was always maintained to the
coils, even while heating up the bed with a natural gas burner located
in the plenum.. In Runs 15.2-17.1 the cooling water flow was turned off
during startup because the natural gas supply was found to be inadequate
for heating up a deep bed while the coils were extracting heat. The bed
was typically heated to 620-650°C before coal feeding was started and
the water flow was initiated. In some circumstances, however, the coils
experienced even higher temperatures.
57
-------
TABLE IV-5. MINIPLANT COOLING COIL DAMAGE
Ul
00
Run No.
4.3
5.4
9.1
13.3
15.4
17.1
18.3
Coal
Configuration
Horizontal
Horizontal
Horizontal
Horizontal
Horizontal
Vertical
Vertical
Damage
Coil 1A was compressed and pushed to one side, some bed
agglomeration occurred.
High velocity impingement of particles from the first stage
cyclone pulsed return eroded a hole in coil 1A.
Coil 2A fatigued at the inlet.
The inlet of coil 3A buckled. Bed agglomeration and
deformation of other coils resulted.
The U-bends of coil IB were thinned and contained several
dimples and holes. Bed agglomeration and deformation of
other coils resulted. Cause was diagnosed as high tem-
perature erosion/corrosion.
The U-bends of coils 2A & 2B were thinned and contained
dimples and holes. Much of the bed was agglomerated and
deformation of the coils resulted. Cause was again
diagnosed as high temperature erosion/corrosion.
Erosion of the sockets connecting the U-bends to the
straight sections of tubing.
-------
FIGURE IV-26
DAMAGED HORIZONTAL COOLING COILS
59
-------
Inspection of the colls after Run 15.4 revealed that coil IB had a num-
ber of dimples, 1 to 3 cm long in the upper bends and that two of the
dimples contained holes 0.3 to 0.6 cm in diameter. Many of the U-bends
were thinned and some had a shiny polished appearance. A portion of the
bed had agglomerated and deformed several of the other coils.
The coils of a vertical configuration which replaced the damaged coils
(see Figure IV-22) were subject to damage of a similar nature in Run
17.1. Figure IV-27 shows a photograph of coil 2A, one of the coils dam-
aged during the run. Note the puncture on the top wall of two of the U-
bends and the warping in several of the straight sections of tubing.
A metallurgical examination was performed on samples from the damaged
coils. The presence of a scale was detected on the outside surface of
all of the tubes and the microstructure of the tubes indicated that the
coils had experienced temperatures above 870°C. The high temperature
resulted in oxidation/sulfidation attack which was aggravated by erosion
of the corrosion scale. The attack was for the most part selective to
the U-bends.
After the second occurrence of high temperature coil damage, the heat up
procedure was again modified. The natural gas supply was supplemented
with liquid fuel so that the cooling water flow to the coils could be
maintained at all times. This has prevented further high temperature
damage during startup.
Two vertical coils were constructed to replace those damaged in Run
17.1. These were used in Runs 18.1, 18.2, and 18.3. One of these coils
developed a leak during Run 18.3 and was badly damaged by the high tem-
peratures which occurred after the cooling water flow was stopped.
The one remaining coil, while intact, was severely eroded at the sockets
connecting the U-bends to the straight sections of tubing. It is
probable that the first coil failed due to erosion of the sockets. The
erosion may have occurred because the sockets of the U-bends projected
from the perimeter of the tube into the path of the vertically directed
fluidized particles.
The third generation cooling coils, were built from heavy wall pipe
(1/2 inch, Sch. 40, 316 stainless steel) to protect against fatigue and
deformation. For additional support the adjacent coils are reinforced
by support rods. In addition, precautions will be taken to ensure that
these coils are never without a flow of cooling water. The new coil
design is shown in Figure IV-8.
Flue Gas Discharge System
Up to about 470 sdm3/s (1000 SCFM) of flue gases were discharged from
the combustor under maximum operating conditions and piped to a scrubber
where residual S02 and particulates were removed before the gas was
discharged to the atmosphere. Gas leaving the combustor was hot (about
60
-------
FIGURE IV-27
DAMAGED VERTICAL COOLING COILS
61
-------
800-900°e) and at elevated pressure (800-1000 kPa) so that it had to be
cooled and reduced in pressure before entering the scrubber. Because
of poor operation of the secondary cyclone prior to its modification,
the discharge gas had high loadings of solids (flyash and sorbent). This
caused serious erosion problems in the discharge line and pressure con-
trol valve, and also made cooling the gases difficult.
Operation of the shell and tube off-gas cooler was often unsatisfactory.
A cooler was needed because a control valve, good for temperatures only
up to 250°C, was used to regulate back pressure in the combustor. The
tubes of the cooler would plug with solids and gas flow would be res-
tricted. Leaks in the heat exchanger tubes also occurred and once a
tube separated from the tube sheet.
Erosion of the pressure control valve was a recurring problem because of
the high solids loading in the off-gas. Erosion was so severe that the
valve usually failed completely after only 24 hours of operation.
Attempts were made to increase the valve's resistance to erosion by
flame spraying a coating of aluminum oxide or tungsten carbide. The
aluminum oxide coating was eroded away at the point of greatest impact
after less than one day's operation. Tungsten carbide showed somewhat
better resistance to erosion.
The off-gas piping was also eroded at elbows in the line. However, this
problem was solved by replacing elbows with tees. Solids quickly filled
the "dead" leg of the tees and provided barrier to erosion of the pipe.
Because of valve erosion and cooling problems, it was decided to control
pressure by an alternate technique. Instead of using a control valve,
pressure was reduced by expanding the gas across a ceramic-lined sonic
nozzle. Pressure was controlled by adding secondary air to the nozzle
inlet. This system had the advantages that the nozzle could be flame
sprayed for erosion protection much more, easily than a valve and that
the nozzle could be designed for high temperatures. This meant that the
troublesome heat exchanger could be eliminated. Thus, the exhaust
gas heat exchanger was replaced with a 20 cm .(8 inch) refractory lined
pipe with a 10.2 cm (4 inch) diameter bore and "the pressure control :valve
was replaced with a converging nozzle with a source of high pressure-air.
This secondary air was metered through a 5.1 cm (2 inch) ball valve equip-
ped with a pneumatic actuator and positioner. The nozzle was flame
sprayed with nickel aluminide and tungsten carbide. The nickel aluminide
provided a sound substrate for the tungsten carbide to adhere to. The
inside surfaces of the pipe downstream of the nozzle were also flame
sprayed with the same materials. The thickness of the coatings were about
0.20-0.28 mm.
The combustor pressure was well controlled once the correct size nozzle
was found, by the sonic nozzle and air supply. However, the tungsten
carbide coating on the nozzle peeled off, probably because of high tem-
peratures. Chromium carbide was substituted for the tungsten carbide
and showed much better resistance to wear.
62
-------
High gas temperatures downstream of the nozzle caused flange gaskets to
fail and flanges to leak. The non-refractory lined pipe also warped
due to thermal expansion. Also, the flue gas had to be cooled before
reaching the scrubber. Because of the problems encountered in using
a shell and tube heat exchanger to cool flue gas which was heavily
loaded with solids, it was decided to use a different method of
cooling. A water cooling jacket of 10 cm (4 inch) pipe was installed
around the 6.4 cm (2-1/2 inch) off-gas piping from the discharge side
of the sonic nozzle to the scrubber, a total run of about 15 m. The
flue gases were cooled from about 700°C, as they exited the nozzle,
to about 250°C before they entered the scrubber. No plugging problems
have occurred with this system, which is effectively a simple double-
pipe heat exchanger.
Prior to the use of the double-pipe heat exchanger, another cooling
system was tried but eventually abandoned. This consisted in the direct
spray of water into the flue gas lines between the secondary cyclone and
the pressure control nozzle. Although this technique effectively cooled
the flue gas, the flue gas line became plugged with flyash that was
deposited near the water spray nozzles. The plugs could be cleared by
turning off the water spray nozzles for a time, allowing the flyash
deposits to dry and then blow away. However, this caused sudden pres-
sure changes in the combustor which, in turn, caused erratic coal
feeding.
Cyclone Operation
Problems with the miniplant's two cyclones can be divided into three
categories: poor collection efficiency, plugging of solids in the dip-
legs, and combustion in the cyclones themselves. The last problem was
not caused by the cyclones per se, but rather by upsets in the combustor
which caused large amounts of carbon to be entrained from the bed.
However, changes in cyclone design were made to minimize the damage
resulting from such upsets. These three problem categories are discussed
below, beginning with the problem of cyclone damage and concluding
with the problem of low efficiency.
During some upsets in the combustor, especially when superficial velocity
reached excessive levels, unburned carbon was blown from the bed into the
cyclones. For example, in one run bed velocity increased to 3.2 m/sec
during an upset and unburned carbon reached the first and second cyclones
and ignited. Temperatures in the cyclones of 1100-1300°C resulted.
When the cyclones were inspected it was found that the ceramic gas out-
let tubes had broken. Pieces retrieved from the cyclones and diplegs
indicated that the tubes had probably been broken for some time, although
several fractures were new. The alumina gas outlet tubes in both cyc-
lones were replaced with new sections of heavy wall stainless steel
63
-------
pipe. While the pipes were not susceptible to cracking, they did melt
when fires again occurred. However, it was soon appreciated that such
fires were the result of abnormal operating conditions, and that they
could be avoided by more careful attention to operation of the combustor,
particularly when changes in operating conditions were made. Changes
in operating procedures were then made which have prevented fires in
the cyclones. Under these conditions, the stainless steel gas outlet
tubes have given satisfactory service.
Plugging of both the first and second stage cyclone diplegs occurred
frequently during early stages of the combustor shakedown. These pro-
blems were virtually eliminated by enlarging the diameter of the dip-
legs, removing sudden reductions in diameter, and keeping the dipleg
piping as nearly vertical as possible. It will be recalled that the
dipleg from the first cyclone returned solids to the combustor, at a
rate controlled by a "pulse-pot" located at the bottom of the dipleg,
and that the dipleg of the second cyclone discharged solids into a
lock hopper.
Use of a gas purge flow (combustor off-gas) in the second stage cyclone
dipleg was tried for awhile as a means of preventing plugging. Both
upward and downward flows were tried. At first it seemed that plugging
was reduced and cyclone collection efficiency improved, particularly
with the downward purge. However, further operating experience indicated
that the use of a purge flow may have actually aggravated plugging by
causing water condensation in the dipleg. The problem was alleviated
by modification of the dipleg geometry to facilitate flow of the solids.
By eliminating the purge flow, condensation was avoided and solids that
were removed from the second cyclone dipleg were invariably dry.
Poor cyclone collection efficiencies were caused largely by low cyclone
inlet velocities. This was partly the result of the cyclones having
been designed before any operating data from the miniplant were avail-
able, i.e., entrainment rates, particle sizes and size distribution,
and particle densities. However, the situation was made worse because
the miniplant was being run with lower air flow rates than were orig-
inally planned. Hence, at the usual combustor operating condition of
870°C, 2.1 m/sec, and 900 kPa, the inlet velocity was only 6.2 m/sec to
the first cyclone and only 10.4 m/sec to the second.
Poor cyclone collection efficiency caused high solids' loading of the
combustor off-gas and serious erosion problems in the off-gas piping.
To correct this situation in the simplest and least costly manner, the
secondary cyclone was modified in order to improve its efficiency. The
area of the gas inlet was decreased by three in order to increase
the inlet velocity to about 33 m/sec. Figure IV-28 gives the original
dimensions of the secondary cyclone and indicates the changes that were
made. Only the size of the inlet and outlet pipes were changed, not the
64
-------
Figure IV-28
CHANGES IN SECONDARY CYCLONE DIMENSIONS
h-15-H
19
10
1
HI
J._4 i
•8
•H
r8.4
-32
ORIGINAL DESIGN
INLET AND OUTLET
MODIFIED INLET
AND OUTLET
ALL DIMENSIONS IN CM
70
67
65
-------
size of the barrel or cone. This resulted in a cyclone of unconventional
geometry but produced a substantial improvement in collection efficiency
with a minimum of changes to the cyclone. Also, the walls of the cyclone
were found to be rough and were reworked to a smooth surface.
Figure IV-29 is a typical particle size distribution for material col-
lected by the primary cyclone. The particle size cut-off was about 70y,
i.e., 90 percent of the collected particles were larger than 70]i.
Figure IV-30 shows the particle size distribution for solids collected by
the secondary cyclone during the last run before the cyclone was modified.
Also shown is the size distribution
-------
Figure IV-29
100
90
80
70
60
50
40
30
PARTICLE SIZE DISTRIBUTION
PRIMARY CYCLONE COLLECTION
LU
LU
O
cc
LU
a.
I-
o:
Lu
§
LU
H
o 20
10
RUN NO. 17.1
0
10,000 5,000
2,000 1,000 500
PARTICLE SIZE, MICRONS
200
100
50
-------
Figure IV-30
100
o
COMPARISON OF PARTICLE SIZE DISTRIBUTION FOR
SOLIDS COLLECTED BY SECONDARY CYCLONE
BEFORE AND AFTER MODIFICATION
Run No. 19.2
D Run No. 19.3
Before Modification
10 -
1000800 500
200 100 80 50
PARTICLE SIZE (MICRONS)
-------
present. In addition, a 2 in. stainless steel pipe was inserted in the
line to prevent agglomerated solids from clinging to the refractory.
Fluidizing Grid
The fluidizing grid first used in the combustar was a water-cooled
stainless steel plate continaing 128 caps which extended above the
grid, each with eight holes of 0.20 cm diameter. Service life of these
caps was short because they were not cooled sufficiently. The holes
would gradually enlarge and the top of the cap would eventually fracture
at the holes.
Because of these problems, a new fluidizing grid was designed which was
similar to the type that had been successfully used in the batch com-
bustor. The new grid (Figure IV-7) was also a water-cooled stainless
steel plate but it did not contain any caps. Instead of caps, the
grid had 776 holes of 0.16 cm diameter. This number and size of holes
was such that the grid pressure drop was equal to about 30% of the
typical pressure drop across the fluidized bed. This is the design
criterion for which uniform distribution of air flow across the grid is
expected. Also, the holes were too small for bed material (limestone)
to fall through them into the plenum. A small fraction of the holes
plugged occasionally, but this was not a problem. In fact, no problems
have been experienced in over 200 hours of operation since the grid was
installed.
Bed Agglomeration
Although the bed material remained freely fluidized in most runs, there
were occasional episodes of agglomeration. Operational difficulties
which resulted in a portion of the bed seeing temperatures greater than
1150°C would cause agglomerates to form. In some instances an agglome-
rate -would develop on the refractory wall opposite of the coal feeding
port due to localized hot spots. Much of the agglomerated material was
loosely packed and would disintegrate with time when exposed to ambient
conditions. At times a portion of the agglomerates would be composed
of fused bed material and flyash. This indicated that temperatures had
been well above 1150°C at some points in the combustor, as calcium car-
bonate and calcium oxide melt at temperatures of 1340°C and 2570°C,
respectively, calcium sulfate melts at about 1350°C and flyash softens
at temperatures of 1260-1370°C.
69
-------
Bed agglomeration also occurred when a break developed in a cooling coil
allowing water to enter the combustor. The bed in the vicinity of the
break apparently absorbed some water before it vaporized. The wet
clinging solids probably formed permanent agglomerates once they were
dried in other portions of the combustor. In several instances a large
portion of the bed was agglomerated in this manner. The agglomerates
usually became enmeshed in the cooling coils and caused severe deforma-
tion. Agglomerates formed in this fashion also had a different appear-
ance and structure compared to agglomerates formed at high temperatures.
The latter were hard, fused solids obviously formed by a high temperature
melting process. The former were softer, could be easily broken by
hand, were not fused but maintained the appearance of a cluster of
particles, each particle maintaining its original shape.
Burner
The burner shown in Figure IV-6, was used to preheat the bed above the
self-ignition temperature of coal. No problems have been encountered
with the burner itself; however, because of an insufficient fuel supply
(natural gas), there was difficulty in heating the bed to the proper
temperature when water flowed through the cooling coils. This necess-
itated the use of supplemental fuel addition as discussed previously.
Also, the ignitor had to be correctly aligned in order to light the
pilot. The ignitor and pilot tube were eventually combined into one
unit which required no alignment. The combined ignitor-pilot is
shown in Figure IV-31.
Flue Gas Sampling
There have been problems in developing a satisfactory flue gas sampling
system. The function of this system is to deliver to the instruments
a clean, dry, low pressure, ambient temperature sample. This is a
demanding task because the flue gas leaving the combustor is heavily
loaded with particulates, contains water vapor, is at high pressure
and is hot. Moreover, because the combustor is at a considerable dis-
tance from the analytical instruments, the residence time of gas in the
sample lines is too long to preclude the possibility of changes in com-
position of the gas. However, an acceptable system was developed and is
now in use on the miniplant based on the system used in the batch com-
bustor unit (see page 27).
70
-------
FUEL INLET
ELECTRODES
CERAMIC SHEATH
12.7 mm STAINLESS STEEL TUBE
Figure IV-31
IGNITOR-PILOT
71
-------
The original sampling system used on the miniplant consisted of a water
knockout, stainless steel filter, refrigeration dryer, and glass micro-
fiber filter. The sample was taken downstream of the second stage cyc-
lone. No pressure reducing device was included; sample pressure was
dropped across a manual flow control valve just before each analyzer.
This system did not work well for several reasons. The stainless steel
filter rapidly became clogged with wet solids, reducing the sample flow.
SC>2 was lost in the knockout and filter, and also in the sample lines.
Residence time in the sampling system exceeded two minutes. Changes in
concentration of the sample (SC>2 and NOX) probably occurred.
In order to reduce the residence time of the sample, the flowrate was
increased and a portion of the total flow was vented before the analyzers.
However, this caused the filter to plug even more rapidly. The sample
line connecting the filter with the combustor off-gas pipe also plugged.
Residence time can also be reduced by lowering the pressure in the sample
line. This can be done by using a corrosion resistant pressure regulator
which is heated above the dew point of the sample to prevent condensation.
Removing a sample directly from the low pressure section of the combustor
off-gas piping was not considered to be practical because of the method
used to control combustor pressure. Pressure is reduced by expanding
the off-gas through a sonic orifice and control is obtained by adding
varying amounts of air just upstream of the orifice. Hence, the low
pressure off-gas has been diluted with a varying amount of air and its
concentrations have been changed.
The new sampling system was described in a previous section (see page 27) .
The flue gas sample is taken downstream of the secondary cyclone and
enters a cooler designed to cool the stream to 200°C or less. A sample
of gas can be withdrawn after the cooler for analysis by wet chemical
techniques. The balance of the gas stream passes through glass micro-
fiber type filters. After the filters, a pressure regulator reduces
pressure to slightly above atmospheric, in order to minimize the resid-
ence time of the sample in the lines. Following the pressure regulator,
a dryer of the permeation distillation type removes water from the flue
gas. This type of drier uses a membrane permeable only to water. This
reduces the losses of SC>2 than would occur if a refrigerated drier were
used. The temperature of the sample is maintained above its dew point
upstream of the drier. The dried gas is then sent to the analyzers
through teflon lines. A portion of the flow is vented before the
analyzers, to reduce residence time further.
This system has given acceptable performance on the miniplant. It can
deliver sample gas for extended periods, but is still subject to occas-
ional plugs and must be cleaned periodically. Analysis of the flue gas
for SC>2 by the IR analyzers agrees reasonably well with analysis by the
wet chemical method. Since the sample for wet chemical analysis is
taken close to the combustor before the sample has been filtered, pres-
sure reduced, dried and passed through long lines to the IR analyzer,
72
-------
it appears that the flue gas sample preparation system does deliver a
representative sample to the IR analyzers. However, additional changes
are expected to be made to the system to improve operation even more.
Included in the improvements will be the installation of a ultraviolet
adsorption analyzer to be used along with the infrared analyzer for
SO 2 measurement.
COMBUSTION RESULTS,
S0 Emissions
The measurement of 309 emissions during most of the shakedown period was
hampered by SOo sampling problems. A reliable sampling system was only
developed at the end of the shakedown period. In addition, a number of
the runs were made before the bed solids rejection system was operable,
and either a captive bed was used with no limestone addition or lime-
stone was added and the bed level was allowed to increase during the run.
In either case, interpretation of the SOo emission data was made more
difficult. However, the concluding run in the shakedown phase was made
under steady state conditions with continuous addition and rejection of
limestone and a functioning flue gas sampling system. This run was made
at a Ca/S ratio of 1.45, a temperature of 870°C, a pressure of 930 kpa
and a superficial velocity of 1.9 m/s. The S02 emissions averaged 750
ppm. This is in line with results from the batch combustion unit ob-
tained at comparable operating conditions. In general, SOo emissions
during the shakedown ranged from 150 to 800 ppm, usually in the 300 to
500 ppm range.
NO Emissions
The NO emissions measured during the shakedown runs were less suscepti-
ble to sampling problems than S0« emissions and a number of runs yielded
data of acceptable quality.
NO emissions were largely NO. Only 10 to 15% of the NO emissions were
present as N0~ and this was quite possibly formed in the sample lines.
NOX emissions were found to correlate with per cent excess air as shown
in Figure IV-32. In Figure IV-32, NOX emissions are expressed as Ib
(as N02)/M BTU coal fired. Data were obtained at pressures of 900 kPa,
temperatures in the range of 900 to 950°C and at excess air levels from
0 to 140%. As seen, the data generally group well around a single cor-
relating line. The emissions also fell well below the current EPA
emission standard of 0.7 Ibs N02/M BTU even at the highest excess air
levels standard.
73
-------
Figure IV-32
NOX EMISSIONS
OQ
CM
O
to
CO
0.6
0.4
to
1 0.2
ui
X
O
0
PRESSURE: 900 kPa
TEMPERATURE: 900-950°C
0
20
40
60
80
100
120
140
EXCESS AIR
-------
Combustion Efficiency
Carbon combustion efficiency measured in the miniplant were in the range
of 92 to 98%. Combustion efficiency is a function of excess stoichio-
metric air as shown in Figure IV-33. The miniplant was operated at
superficial velocities of 1.8 to 2.2 m/s at temperatures of 800 to 950°0
and at a pressure of 900 kPa. Two other factors may influence combustion
tion efficiency: bed temperature and the efficiency of the carbon par-
ticulate recycle system. In these tests, the effect of these variables
could not be determined. However, there were some indications that
increasing temperatures would increase combustion efficiency as expected.
The efficiency of the carbon particulate recycle system was a variable
during a number of the runs due to plugging of the first stage cyclone
dipleg or problems with the system which injected the solids from the
first stage cyclone back into the combustor. The effect of these
variables may be responsible for the data scatter seen in Figure IV- 33.
at low excess air levels.
The target combustion efficiency of 99% could not be attained even at
high excess air levels in the shakedown runs. Therefore, a more effic-
ient first stage cyclone recycling the unburned carbon fines to the
combustor, operation of the combustor at higher temperature or possibly
a separate carbon burn-up cell may be required to reach 99% combustion
efficiency. Additional work will be done to improve combustion effic-
iency and attain the target level of 99%.
The loss of combustion efficiency is due chiefly to unburned carbon
particles entrained from the bed. This accounts for about 98% of the
loss. The balance is due to CO formation.
CO Emissions
Carbon monoxide emissions were generally quite low as long as conditions
conducive to high combustion efficiency were maintained. Typically,
levels of less than 500 ppm were observed at excess air levels exceeding
157» and average bed temperatures of ca. 900°C. A sufficient amount of
data at widely varying operating conditions was not obtained during
miniplant shakedown to allow development of meaningful correlations.
However, in addition to the importance of excess air level and average
bed temperature, preliminary results indicate that steady coal feeding
and high combustor freeboard temperatures favor maintenance of low CO
levels. For example, during runs 17.1 and 18.1, under relatively steady
conditions of coal feeding, with temperatures throughout the combustor
in the 850-920 C range and excess air ranging from 15-60%, measured CO
levels were 150-250 ppm.
Heat Transfer Coefficients
Heat transfer coefficients were measured during selected runs by main-
taining the cooling water flow through the coils in the liquid state.
The flow rate to each coil was measured with an orifice meter. The
75
-------
100
Figure IV-33
COMBUSTION EFFICIENCY
95
>
o
LLJ
O
n 90
CO
CD
S
o
o
85
80
0
20
PRESSURE: 900 kPa
SUPERFICIAL VELOCITY: 1.8-2.2 m/s
TEMPERAT URE: • 880-900°C
• 900-940°C
_L
40 60
EXCESS AIR
80
100
120
140
-------
inlet and outlet temperature of the cooling water were measured with
thermocouples inserted within the 3/4 in. O.D. tubing. The flows and
temperatures were recorded on magnetic tape at one minute intervals.
These data were used to calculate the average coefficient and standard
deviation for a 10 minute interval. The results are given in Tables
IV-6 and IV-7.
The coefficients measured for the vertical coils during Run 19.2 compare
closely with those measured in runs 14.1, 14.2 and 15.1 for horizontal
serpentine coils oriented vertically. This result implies that as
long as the configuration and/or orientation of the heat transfer
surface allows good mixing, the coefficients will be similar. Unfortu-
nately coefficients are not available for the horizontal serpentine
coils oriented in a horizontal direction. However, it is likely the
coefficients would have been somewhat lower due to poorer mixing.
The coefficients measured for the lower-most coil, 1A, are consistently
slightly lower than those measured for coil IB. The cause may be poor
fluidization at the boundary between the lower-most coil and the com-
bustion zone. Insufficient data are available to determine if there are
other variations in the coefficient as a function of axial position.
Component Balances
Complete material balances were routinely made only for sulfur and
calcium. Even for these particular components, various problems were
encountered which led to questionable data reliability. For example,
flue gas sampling problems introduced some error in SC>2 determinations.
In addition, S03 levels in flue gas were not determined. Problems
associated with collection, sampling, and analysis of solids were of
even more importance. These included such things as poor cyclone effic-
iency, some inaccuracy in determining the weight of the final bed solids,
difficulty in obtaining representative samples of collected and bed
solids, and, most importantly, large sensitivity in the overall sulfur
balance to relatively small errors in solids analysis. The net result
was that sulfur balances fluctuated from 50-120%. Calcium balances
were better, but were still subject to similar sources of error with
regard to collected and bed solids. Typical balances were in the range
of 85-115%. Appendix Tables G3 and G4 give more details. Work is
continuing to improve these balances, tentatively to 90% or higher with
a higher degree of precision. Preliminary indications are that these
levels have been reached in the most recent runs.
77
-------
TABLE IV-6. MINIPLANT OVERALL HEAT TRANSFER
COEFFICIENT MEASUREMENTS - RUN 19.2
Coil No.
1A
IB
2A
Average Coeff.
W/m2K
319
353
330
Standard Deviations
of 10 Measurements
Obtained At One
Minute Intervals
W/m2K
7.7
7.2
11.3
Surface Area of a Coil, m
Coil Heat Flux, W/m
3
Combustion Intensity, W/m Bed
Heat Removed by Cooling Coils in
Bed, % of Coal Heat Input
Calculated Overall Heat Transfer
Coeff., from Heat Balance, W/m K
Coil Configuration
0.551
280,000
5,250,000
57
358
Vertical
78
-------
TABLE IV-7. MINIPLANT OVERALL HEAT TRANSFER
COEFFICIENT MEASUREMENTS - RUNS 14.1, 14.2, 15.1
Standard Deviations
of 10 Measurements
Obtained At One
Run No. Coil^No^ Average Coeff. Minute Intervals
W/m2K
8.3
7.4
3.1
6.6
3.9
3.8
5.7
Coil Configuration - Horizontal Serpentine Half Coils Oriented
in the Vertical Direction
14.1
14.1
14.2
14.2
14.2
15.1
15.1
1A
IB
1A
IB
IB
1A
IB
W/m2K
316
327
321
336
348
317
338
79
-------
SECTION V
BATCH COMBUSTOR STUDIES
EQUIPMENT, MATERIALS, PROCEDURES
Fluidized Bed Coal Combustion Unit
A schematic diagram of the Exxon batch fluidized bed combustion unit is
shown in Figure V-l. Figure V-2 is a photograph of the unit. The
combustor is equipped with a continuous coal feeding system. The sorbent
is added batch-wise. The primary components of the unit are (1) the
coal feeding system, (2) the fluidized bed combustor, and (3) the gas
handling and analytical equipment.
Coal Feeding System - Figure V-3 shows the Petrocarb Model 16-1 ABC
injector. The main features are a conical-bottom tank that holds solids
to be fed and an orifice and mixing tee assembly that mixes solids with
carrier gas. Solids in the tank are aerated by a controlled stream of
air at a selected pressure. Aerated solids flow through the orifice at
the bottom of the tank into the mixing tee assembly and are picked up by
a controlled stream of carrier gas (air). Solids are pneumatically con-
veyed through a transport line into the combustor. The feed rate of
solids is controlled by pressure in the feed tank, carrier or injection
air flow rate, and pressure differential between the feed tank and com-
bustor.
The Petrocarb solids feeder was modified to feed ground coal (-16 mesh)
at rates of 3-13 kg/hr into the combustor against pressure of up to
approximately 1000 kPa. The feeder, as it is supplied by Petrocarb, can
handle only much higher feed rates. The diameter of the orifice was
reduced to 0.71 cm and the Petrocarb injection hose was replaced with a
0.58 cm diameter (I.D.) X 6.10m long stainless steel tube. In order to
make the feeder work satisfactorily with the batch combustor, the feeder
to combustor pressure differential had to be held constant. This was
accomplished with automatic controls which maintained the pressure in
the feed tank above the pressure in the combustor by the desired amount.
The entire feeding assembly was mounted on a platform scale which
measured the coal feed rates.
Fluidized Bed Combustor - A schematic diagram of the batch fluidized
bed coal combustor is given in Figure V-4. The vessel was constructed
from four sections of 25 cm (10 inch) diameter standard wall carbon
steel pipe and refractory lined with Grefco Litecast No. 7528 to an
inside diameter of 11.4 cm. The height of the vessel, above the fluidiz-
ing grid, was about 4.9m. Below the grid was a 61 cm long burner section
80
-------
CYCLONES
CONDENSER
DRAIN
CITY WATER
FLUIDIZING
GRID
HEATUP
DEMINERALIZER BURNER
AIR FROM,
COMPRESSOR
PROPANE
WATER SAMPLING SYSTEM
STARTUP
HEATER
FILTER
PRESSURE
CONTROL
VALVE
INJECTION AIR
PLATFORM C
SCALE
VENT
X
OFF-
GAS
LLER
KNOCKOUT
T
Figure V-l
BATCH FLUIDIZED BED COAL COMBUSTION UNIT
-------
FIGURE V-2
BATCH FLUIDIZED BED COAL COMBUSTION UNIT
REGENERATOR
-------
EXHAUST VALVE
PRESSURE RELIEF VALVE
FILLING VALVE
TANK
PRESSURE
GAUGE
TANK PRESSURE
CONTROL VALVE
TANK PRESSURIZING
VALVE
TANK
PRESSURE REGULATOR
AIR INLET
FLOWRATOR
INJECTION
LINE
TOCOMBUSTOR
AIR /
SHUTOFF
VALVE
LINE PRESSURE GAUGE
AIR PRESSURE GAUGE
AIR PRESSURE
REGULATOR
AIR CONTROL VALVE
AIR HOSE
ORIFICE ASSEMBLY
Figure V-3
PETROCARB COAL INJECTOR
83
-------
TH F RMO~
COUPLE 2.7 M
071 M
0 71 M
r.oni ING 1
INLET
0£.l n/i
. 0 1 IVI
PROBE i
.
0.69 M
i
n
u
^^•nH
SfSSS
-•_••»
[ fc
^flf,""
X'TSH* MM
^
•• .
••• ••
••• ••!!
••• ••
•«g=a
•••» ••
sanssai
•n™" ••
1^
i
\
383
^
\_
^
I—1
^
^
^___l
s
f ™
Ull
•nm
III
Illl
LuU
n
^
^^
>Q
^85
J
X
J^^
'./
/
%
L
GREFCO LITECAST #7528
REFRACTORY
11.4 CM DIAMETER
< 10" STD WALL STEEL PIPE
j^^\
<^ SOLIDS CHARGING PORT
HORIZONTAL COILS
(NO WATER FLOW)
^ r~i A M r* cr
^^ rLANut
x/rrnTipAi pnni IK
Vt-Kllw/AL. Ov/Ul—IIM
> COILS
1 COOLING WATER OUTLET
\^> SOLIDS REMOVAL PORT
^^^
P
— I ^ r LU IUIZ.IIM b bKIU
(WATER COOLED)
,, PI1PPAI ITT RFF RAP TORY
BURNER
Figure V-4
COMBUSTOR VESSEL
84
-------
lined with Grefco" Bubbalite. The fluidizing grid, which was made of
stainless steel, had 80-0.16 cm diameter holes to distribute the fluidiz-
air and was water cooled. It is described in Figure V-5. A propane
burner, located at the bottom of the burner section, was used to preheat
the unit to above the self-ignition temperature of propane (505, C). At
this point propane was added directly into the bed just above thefluid-
izing grid to raise the temperature of the bed above the self-ignition
temperature of coal. A schematic diagram of the burner is shown in
Figure V-6. The maximum operating temperature and pressure of_the
batch combustor were approximately lOOO'C and 1000 kPa respectively.
The combustor had three vertical cooling coils made of 0.95 cm diameter
(0 D.) stainless steel tubing. The locations of the coils within the
combustor are shown schematically in Figure V-7. They extended from
27 to 141 cm above the fluidizing grid and each had a surface area of
0 060 m2. Each coil had its own rotameter and control valve and the
water flow into each coil and the temperature of the steam issuing
from the coils was easily controlled. Initially, the combustor con-
tained ^ horizontal serpentine cooling coils of 0.093 m2sU**« «"
each The bottom four coils were removed and replaced by three vertical
coils as detailed on page 98. Two horizontal coils, 180-249 cm above
the fluidizing grid were left in place but were not cooled. Thermocou-
ples were located 15 cm apart in the lower section of the combustor and
30 cm apart in the upper section.
Sorbent was loaded into the combustor through a charging port located
in the upper section. Solids could be removed through a port in the
lower section or, alternatively, transferred directly to the adjacent
batch regenerator by blowing them through a 5 cm diameter pipe supplied
for this purpose.
Coal entered through a "sonic" coal injection probe which was connected
to the end of the 6.10m X 0.58 cm diameter (I.D.) coal injection tube.
Figure V-8 gives a schematic of the "sonic" probe. The inside diameter
of the probe was 0.77 cm. The stream of flowing coal was surrounded by
seven sonic air jets. The primary intent of the high velocity air jets
was to improve coal feeding by clearing a path through the bed of
fluidized solids in the combustor. Air flow through the annulus of
the probe helped to cool the probe. This flow was about 1.89 dm /s
and compares to an air flow of about 3.30-3.78 dm3/s used to transport
the coal.
Gas Handling and Analytical Equipment - Flow of air and fuel into the
combustor, and system pressure, were under automatic control. Gases
leaving the combustor first passed through two cyclones, which removed
85
-------
Figure V-5
BATCH COMBUSTOR
FLUIDIZ1NG GRID
2-0.635 CM DIA. CHANNELS,
180°APART
80-0.159 CM HOLES
ON 0.953 CM SQUARE
PITCH
11-
0.476 CM DIA.
CHANNELS
INSERT
12.7CM
SQUARE
0.953 CM
OD 316SS
TUBING
31.8 CM DIA,
1.27 CM THICK
86
-------
BURNER HEAD
(BRASS)
GASKETS
7.6 CM DIA.
400-0.8
HOLES
COOLING COIL
WATER COOLING
CHANNEL
BAFFLES
MM 4
19.0
POROUS METAL DISK
5 CM DIA. S.S. TUBE, 14 CM LONG
ENTIRE TUBE FILLED WITH
ALUMINA BEADS
FLANGE
AIR& FUEL
Figure V-6
PREHEATER BURNER
87
-------
6.4 CM ;
71 CM
6.4 CM
61 CM
V
27 CM
_L
\s
r\ r\
11 CM
HORIZONTAL BAFFLES
183-254 CM ABOVE GRID
32 CM
"4.4CM-
34 CM
141 CM
34 CM
102 CM
8 CM
T
GRID
VERTICAL COILS-316 SS TUBING, 6.4 MM O.D., 0.89 MM WALL
EACH COIL = 0.060 M2 SURFACE AREA
Figure V-7
LOCATION OF COOLING COILS
88
-------
Figure V-8
COAL INJECTOR
AIR
13 MM OD, 0.89 MM WALL
5.1 CM
6.4 MM OD X 0.89 MM WALL
1 I } \
9.5MM ODX 0.89 f
t/IM WALL
* A ° t
CM
2.5
ci\r
i
5.1
CM
COAL IN
TRANSPORT
AIR STREAM
ALL MATERIAL 304 STAINLESS STEEL
7-0.4 MM HOLES,
EQUALLY SPACED
-------
entrained sorbent and fly ash. An off-gas cooler, which followed the
cyclones, reduced the temperature of the off-gas to the desired level.
The off-gas then entered a coil of 2.5 cm diameter stainless tubing
which was electrically heated during startup to raise the temperature
of the gases above the dew point. A 3.81 cm diameter Aerotec cyclone,
following the heater, was used to remove particulates during startup
of the combustor, when water vapor condensation in the first two cyclones
caused them to operate at reduced efficiency. Fine particulates were
removed with a Pall Model MEC-800-18-C filter, located upstream of the
back-pressure control valve. This filter had a mean pore size of 165
microns and an area of 0.37 m2. Before being vented, the off-gas
entered a chiller and knockout to remove moisture so that the water
content of the gas could be determined. A small portion of the off-gas
was diverted after the back-pressure control valve and sent to the gas
sampling system. A schematic of the sampling system is shown in
Figure V-9. The sample gas passed through a small length of heated
stainless tubing before entering a Balston Model 33 filter for final
particulate cleanup. The gas then went through a 1.8 m section of
heated Teflon line before passing through a permeation drying tube
(Perma Pure Dryer Model PD-1000-24S) for moisture removal. Downstream
of the dryer the line was unheated Teflon tubing to the analytical
instruments. The analytical instruments used in the system were:
1. Beckman 865 S02 Analyzer (Infrared) 0-2000 ppm
0-3000 ppm
0-10,000 ppm
2. Beckman 864 CO Analyzer (Infrared) 0-1000 ppm
0-2500 ppm
0-5000 ppm
3. Beckman 864 C07 Analyzer (Infrared) 0-5 %
Z 0-10 %
0-20 %
4. Beckman 715 Oxygen Analyzer (Polarographic) 0-5 %
0-25%
5. Thermo-Electron 10B NO/NO Analyzer (Chemiluminescence)
8 ranges from 0.01-10,000Xppm
Coal - Three different coals were used in the batch fluidized bed coal
combustion program. The majority of the runs were made using a high
volatile (A) bituminous coal from Consolidation Coal Company's Arkwright
mine in West Virginia. It was ground to -16 mesh by Penn-Rillton Co.
The specified particle size distribution is given in Table V-l. A
measured particle size distribution is presented in Figure V-10 and
corresponds very closely to the specified distribution. Runs were also
90
-------
TO
PRESSURE |{ OFF
CONTROL 1* CHIL
VALVE I ,*.
ABSORBERS
f y •
FILTER
XER
r++*n
N2 PURGE
1
^
t I
i PERMEATION UNHE/
HEATED STAINLESS \ DRYING TUBE
STEEL LINE HEATED TEFLON LINE
TO ANALYTICAL INSTRUMENTS
UNHEATED TEFLON LINE
Figure V-9
FLUE GAS SAMPLING SYSTEM
-------
Figure V-10
PARTICLE SIZE DISTRIBUTION OF ARKWRIGHT COAL
s l-°
CO
UJ
_i
o
0.9
2 0.8
z
<
' 0.7
CO
UJ
0.6
G °-5
U_
i- 0.4
LU
UJ
0.3
0.2
3 0.1
O
0
J_
±
J_
JL
J_
_L
_L
10 20 30 40 50 60
70 80 90
MESH SIZE
100 110 120 130 140 150 160 170
-------
made with a low sulfur Western sub-bituminous coal and a high sulfur
Illinois bituminous coal. The Illinois coal was provided by Argonne
National Laboratory. A proximate and ultimate analysis was made on
each of the coals and the results are presented in Table V—2.
TABLE V-l. COAL PARTICLE SIZE DISTRIBUTION
PENN-RILLTON CO. GRIND B-2 SPECIFICATION
U.S. Mesh Size 10 20 30 40 100 200 pan
Wt. Fraction on Screen 0 4;5 15.5 14 35.5 12.5 18
TABLE V-2. COMPOSITION OF COALS USED IN BATCH
FLUIDIZED BED COAL COMBUSTION PROGRAM
Source - Consolidation Coal Co. (Arkwright Mine - West Virginia)
Proximate Ultimate
Moisture 1.00 wt. % Moisture 1.00 wt.
Ash 8.11 Ash 8.11
Volatiles 36.86 Total Carbon 76.26
Fixed Carbon 54.03 Hydrogen 5.30
Sulfur 2.66
Nitrogen 1.49
Chlorine 0.07
Oxygen(1) 5.11
Higher Heat Value = 14,045 Btu/lb
Source - Peabody Coal Co. (Mine 10, Seam 6 - Illinois)
Proximate Ultimate
Moisture 6.0 wt. % Moisture 6.0 wt,
Ash 12.7 Ash 12.7
Volatiles 40.5 Total Carbon 62.9
Fixed Carbon 40.8 Hydrogen 4.5
Sulfur 4.1
Nitrogen 1.2
Chlorine 0.0
Oxygen CD 8.5
Higher Heat Value = 11,300 Btu/lb
93
-------
Source - Carter Oil Co. (Wyoming)
Proximate Ultimate
Moisture 2.2 wt. % Ash 7.9 wt. % (Dry)
Ash 7.9 Total Carbon 68.0
Volatiles 44.0 Hydrogen 5.0
Fixed Carbon 45.9 Sulfur 0.7
Nitrogen 0.9
Chlorine 0.0
Oxygen(l) 17.9
Higher Heating Value = 11,850 Btu/lb (Dry)
(1) By difference
Sorbents - Grove limestone (BCR No. 1359) and Tymochtee dolomite were the
primary sorbents used in the experimental studies. The stones and their
properties are given in Table V-3. The particle size of these materials
was generally in the 8 X 25 mesh range. Typical particle size distri-
bution curves for the limestone and dolomite are shown in Figure V—11.
Baker dolomite and Pfizer dolomite (BCR No. 1337) were also tried but,
because of their high attrition rates, their use was discontinued.
TABLE V-3. PROPERTIES OF SORBENTS USED IN
BATCH FLUIDIZED BED COAL COMBUSTION PROGRAM
Stone Chemical Analysis, wt. %
Designation Quarry Source Type CaO MgO Si02 A12Q3 Fe203
1359 Grove Lime Co. Limestone 97.0 1.2 1.1 0.3 0.2
(Stephen City, Va.)
Tymochtee C. F. Duff & Sons Dolomite 53.8 38.7 5.3 0.9 1.2
(Huntsville, 0.)
Experimental Procedures - Operation of the batch fluidized bed combustor
can be divided into four phases: startup, ignition and pre-heating, coal
feeding, and shutdown. Startup consisted of those activities preliminary
to ignition of the propane burner. These activities included checking
equipment to make sure it was ready for a run, checkout of the analyzer
calibration, charging sorbent, turning on electrical circuits and the
air compressor, turning on all cooling water systems (fluidizing grid,
burner, steam coils, condenser) and purge air systems (pressure taps,
sight-glasses, AP cells).
94
-------
Figure V-ll
PARTICLE SIZE DISTRIBUTION OF SORBENTS
TYMOCHTEE
DOLOMITE
, GROVE
LIMESTONE
MICRONS
500
1000
PARTICLE SIZE
1500
2000
2500
-------
To ignite the propane burner, air and fuel flows were set and the
ignition electrode was activated. Safety devices shut down all flows
if ignition was not obtained within ten seconds or if a flame-out
occurred afterwards. A safety interlock prevented startup for 3 min-
utes after an automatic shutdown to assure adequate purging of the
combustor. Subsequent to ignition, combustor air flow and pressure
were adjusted to the values desired for making the run. All gas flows
and pressure were controlled automatically. After the bed temperature
reached the ignition temperature of propane, propane flow to the bed
was initiated to heat the bed up to the coal ignition temperature.
Preparation of the coal feed system for a run consisted of setting the
flow of injection air and activating and adjusting the coal feeder-to-
combustor AP control system. Coal injection could be started only
after the temperature in the combustor was high enough for self-igni-
tion of the coal to occur. Propane flow was stopped automatically at
the same time that feeding of coal was started. An automatic safety
circuit would shut down coal injection if the combustor temperature
dropped too low to ensure combustion of the coal or if the feeder to
combustor AP dropped below a pre-set minimum (about one psi).
Data on the weight of the coal feeder vs. time was taken so that the
feed rate of coal could be determined. Another method of estimating
the feed rate was to observe the oxygen concentration in the off-gas
from the combustor. A rapid rise in oxygen concentration was usually
the quickest way of determining that a problem was developing with the
coal feeding system.
Temperatures in the combustor could be controlled by regulating the
amount of coal burned. The feed rate of coal could be adjusted by
changing the flow of injection air or coal feeder to combustor AP.
To shut down the combustor routinely, the coal feed valve was shut,
fluidizing air was stopped, and nitrogen purge was started to pre-
serve the solids. Flow of injection air was kept on for several
minutes so that the coal feed line could be cleared of coal. All
water flows were reduced. Solids could be discharged from the reactor
(by blowing them out of a port located just above the fluidizing grid)
after the bed cooled overnight.
EQUIPMENT AND TECHNIQUE DEVELOPMENT
Coal Feeding
A detailed discussion of the early phases in the development of the
modified Petrocarb coal feed system used on the batch unit was dis-
cussed in a previous report (3). In the earlier study an extensive
96
-------
program was undertaken to test the suitability of the Petrocarb feeder
for supplying coal to the batch combustor. The effect of feed tank
pressure, feed tank-combustor pressure differential, injection
line diameter, and injection air flow rate on the coal feed rate were
thoroughly investigated.
As a result of numerous coal combustion runs since then, additional
insight into the nature of the coal feeding system has been gained.
Other factors such as coal probe design, orifice assembly design, and
moisture content of the coal have been shown to have a pronounced
effect on the ability of the Petrocarb system to deliver a steady and
dependable coal flow.
Observations of the mixing of solids in Exxon's cold model test unit
were also made. The depth of the injection air jet penetration into
the fluidized bed was also observed in this unit to obtain qualitative
information about coal feeding by pneumatic transport into fluidized
beds. The solids mixing in the vicinity of the coal injection point
was observed to be poor, especially with deep beds. The poor mixing
was due to the horizontal orientation of the coils. Only when the
settled bed depth was 0.30 m or less and the superficial velocity was
1.83 m/sec or more did the solids mix well. Slugging occurred in the
space between the fluidizing grid and the bottom of the simulated cooling
coils. Also, the downward moving solids tended to be densely packed at
the reactor walls. Penetration of air jets into the bed from a simu-
lated coal injection probe was very low. These observations suggested
that the bed solids were impeding the flow of coal into the bed. At
this point a significant improvement was made to the batch unit coal
feeding system by modifying the coal injection probe. The water-cooled
probe which had been used up to this point was replaced with a new air-
cooled probe which had sonic flow air jets surrounding the central air/
coal stream. The primary intent of the high velocity air jets was to
improve coal feeding by clearing a path through the bed of fluidized
solids.
The sonic jet probe was then tested in the cold model unit. It was
observed that the bed was penetrated to a much greater depth by the jets
issuing both from the sonic flow holes and the central coal transport
air stream. The system was then installed on the combustor and the
coal feed rate became much steadier with significantly fewer stoppages
due to plugs in the coal feeding line.
The coal feeding system was improved further by redesigning the orifice
assembly at the bottom of the injector vessel. The dependability of the
coal feeding system had often been poor due to the numerous instances in
which the coal had plugged the orifice of the feed vessel. Some of the
plugs appeared to be due to the bridging of fine coal at the orifice or
by a caking of moist coal on the orifice walls. To alleviate these pro-
blems an orifice assembly was designed in which the inside bore was
97
-------
wou TTh With1n° SUdden ChangeS in dia^ter. Thus, no shoulders
would exist where coal would tend to accumulate. Figure V-12 shows a
schematic of the orifice assembly presently used. A£O! to insurlthat
dry coal was used, it became a routine practice to store the coal in a
hot room for several days prior to its use. The coal had
apparently become wet during extended storage. After installation of
the sonic_jet probe and the improved orifice assembly, and the use of
the additional drying step, coal feeding became very steady and depend-
cluJLG •
Bed Temperature Distribution - Cooling Coils
A serious problem which occurred during the early testing of the com-
bustor was a poor temperature profile in the bed, characterized by
severe hot spots at the coal.feed point. Temperature drops of over
330°C were measured across expanded beds 0.9 m high. The cause of this
problem was believed to be combustion of the powdered coal occurring at
a much faster rate than solids could mix and distribute heat throughout
the bed. In the 11.4 cm diameter reactor, bed heights of 60-180 cm were
used, corresponding to length-to-diameter ratios of 5-16. Top-to-bottom
mixing of solids was poor because of the high L/D ratio of the bed and
also because the tightly wound horizontal serpentine cooling coils
hindered movement of solids. The intent of this coil design was to
break up slugs of solids and promote better mixing between gas and solid
phases. Visual observation in the cold model unit confirmed that
slugging was absent in the coil section but that top-to-bottom mixing
was very poor.
In order to Improve mixing, the coils were changed from the horizontal
serpentine design to vertical loops. Four of the six horizontal coils
(0.093 m2 area each) were replaced by three vertical coils of the same
surface area per coil. The top two horizontal coils remained in place.
Figure V-13 is a photo of a vertical coil. Figure V-14 is a comparison of
temperature profiles obtained with horizontal and vertical coils. It
is clear that vertical coils produced a dramatic improvement; hot spots
were nearly eliminated and the temperature drops across the bed were
reduced to about 50°C.
Vertical coils had another effect on the operation of the combustor. The
first set of vertical coils had the same surface area as the horizontal
coils which they replaced. However, in the first run made with vertical
coils, there was some difficulty in maintaining bed temperatures even
though the coal input was higher than normally used with horizontal coils.
It was quickly apparent that vertical coils produced higher heat transfer
coefficients than horizontal coils, and an explanation was that solids
were moving more freely with vertical coils. Indeed, measurements of heat
transfer coefficients indicated that vertical tube coefficients were sig-
nificantly higher (see page 123). Eventually, the first set of vertical
coils had to be replaced with coils with less surface area (0.060 m2 per
coil). The coil tubing O.D. was also increased from 0.64 to 0.95 cm.
98
-------
Figure V-12
COAL FEEDER ORIFICE ASSEMBLY
32 MM TAPERED BUSHING
MACHINED SMOOTH
25 MM SMITH BALL VALVE
TO WELDED
19 MM UNION
— 6 MM ORIFICE
— 6 MM PIPE THREAD
MATERIAL: ALL STAINLESS STEEL
99
-------
FIGURE V-13
VERTICAL COOLING COIL
100
-------
Figure V-14
O
O
o
UJ
cc
LU
1000
950
900
850
800
COMPARISON OF BED TEMPERATURE PROFILES FOR
HORIZONTAL AND VERTICAL COOLING COILS
• HORIZONTAL
SERPENTINE
A VERTICAL LOOPS
750
20
40 60 80 100
HEIGHT ABOVE FLUIDIZING GRID, CM
120
-------
Sampling System
To assure reliable, accurate flue gas analyses the sampling system used
on the batch fluidized bed combustion unit was continually modified and
improved. The system has undergone three distinct phases and each will
be discussed below.
1. Phase I - This system was the initial one used on the batch unit. A
small portion of the off-gas was drawn off from the gas handling system
and diverted to a refrigerator which lowered the dew point of the gas to
about 2°C before it was sent to the gas analysis equipment. This system
had several drawbacks which may have prevented an accurate analysis of
the flue gas. With this system the possibility of SC>2 dissolving and
being removed by liquid water existed and the analyzer response times to
changes in concentration of the off-gas were slower than desired.
2. Phase II - This system consisted of a steam-heated pressure regulator,
a glass microfiber filter (Balston Type 33) and a self regenerative per-
meation drier. The sampling line was stainless steel and was heated up
to the drier. The sampling port on the combustor was located downstream
of the second cyclone and before the off-gas cooler. Another major
difference between the Phase I and II sampling systems was the sampling
pressure. The Phase I sampling was done at atmospheric pressure and the
Phase II sampling was done at 800 kPa.
The sampling system was tested by burning coal in an inert bed and then
comparing the measured and calculated (by material balance) S0£ concen-
trations. The inert material used was alundum (38X, Size 20, supplied
by Norton Company). On the first test, an 0-ring in the drier failed at
about 700 kPa, even though the unit was rated for 800 kPa. For this
test only, the drier was replaced with a refrigerator. The average
measured S02 concentration was only about one-third of the calculated
concentration. Also, the S0£ concentration increased when the sample
flow rate was increased; the maximum concentration was about two-thirds
of that calculated.
In the second test, the self-regenerative drier was used and the sample
gas flow was increased. The residence time in the sample, line was about
45 seconds. The measured S02 concentration averaged about one-half of
the calculated concentration. Increasing the sample flow further caused
only a slight increase in the indicated concentration.
In the third and fourth tests, the flue gas was analyzed for S02 by
absorption in a 3% hydrogen peroxide solution. This is a modification of
the EPA test method (titrimetric). In both cases, the measured S02 was
less than half of the calculated values. The sampling points were dif-
ferent for each test. In test no. 3, the sample to the peroxide bubbler
was removed downstream of the drier, after passing through the heated
stainless steel sample line and filter. In test no. 4, the sample was
102
-------
removed from the combustor and connected with polyethylene tubing to a
heated glass wool filter and the bubbler. Hence, the sample did not
contact any metallic surfaces after leaving the combustor.
As a result of these tests, it was determined that this sampling system
was inadequate for reliable S0£ measurements. A possible reason for the
S02 loss might have been the condensation of H2S04 in the sampling line.
There were some signs of corrosion in the sampling system which supports
the possibility of condensation and S03 formation. Another drawback of
this system was the high residence time of the gas in the lines.
3. Phase III - This system is the one presently used on the batch unit.
The location of the sampling port was changed to a position downstream
of the combustor pressure control valve. The major advantage of sampling
at low pressure is that the residence time in the lines is substantially
reduced. The gas sample passed through a small length of stainless steel
line, electrically heated to about 200"C to prevent condensation of 112804,
a heated glass microfiber filter, and a permeation drying tube. A 1.83 m
length of heated Teflon tubing connected the filter and dryer. Unheated
Teflon line was used downstream of the dryer.
Checks were made on the sampling system and the results indicated that
the system worked well. A series of tests were made in which a S02/N2
gas mixture was metered into the combustor along with air and the SC>2
was analyzed in the off-gas. The results showed that the actual vs.
expected measurements were within the expected range of error. Several
tests were also conducted in which the S02/N2 gas mixture was injected
into the combustor while the preheat burner was operating. The infrared
analyzer read 25 to 38% lower than the calculated concentration.
Measurement by the wet chemical method was 10-12% higher than calculated.
Additional testing will be continued to aid in the development of an
even more reliable sampling system.
Cooling Coil Damage
Damage to the 316 SS cooling coils has occurred several times. It is
difficult to attribute failures to one cause; however, it was probable
that tube failures occurred because of one or more of the following
conditions: excessively bigh tube metal temperatures, corrosion, and
mechanical causes.
The first failure of a horizontal serpentine coil occurred after about
50 hours of operation. This coil had been subjected to conditions that
were much more severe than would be encountered in normal operation.
Because of hot spots in the fluidized bed, temperatures external to the
coil frequently exceeded 1100°C. A lack of water pumping capacity
caused steam temperatures inside the coil to be about 500-600°C. These
conditions may have caused tube temperatures to be higher than the metal
could withstand, particularly with regard to corrosion.
103
-------
A specimen from this coil was analyzed by optical and electron microprobe
techniques. The conclusion was that the coil was in a "sensitized" con-
dition and had suffered severe sulfur attack on its outer surface.
"Sensitized" means that chromium carbides had precipitated at the grain
boundaries, leaving regions close to the grain boundaries low in chromium
and thus less resistant to intergranular corrosion. This would have been
particularly detrimental during "downtime" when sulfur acids and chlorides
present in condensed water could produce intergranular cracking. Type
316 SS is susceptible to sensitization when it is heated to temperatures
of 450-650°C. The external surface of the tube showed sulfur attack down
to a depth of about 25 ym (wall thickness = 890 ]m) . Corrosion took the
form of a surface scale of Cr-Fe oxide and an area beneath the scale
rich in Fe, Ni, and S, but low in Cr. The internal surface contained
deposits of Mg, Ca, and Si and showed slight erosion. No erosion was
noted on the outside of the coils.
A second horizontal coil which failed showed similar damage. A section
of coil located near the steam outlet had heavy sensitization and inter-
granular attack; but, a section near the water inlet had negligible
damage. All the coils had a scale deposited on the outer surface which
consisted primarily of carbon and sulfur. Beneath this scale was the
metal surface and intergranular scale.
Failure of a vertical coil has also occurred; in this case a 20 cm
length of tubing near the water inlet had melted. Surprisingly, the
remainder of the coil appeared to be in good shape.
Cooling coil failures appeared to be related to excessive metal tempera-
tures. However, analysis showed that tube metal temepratures should be
quite low, under normal operating conditions. The tube wall temperature,
Tw, is given by the following expression:
To + r Ti
w 1 + r
where To = temperature outside coil
Ti = temperature inside coil
r = jp = ratio of inside to outside heat
0 transfer coefficients
Since hi is at least 5500 W/m2 °C and ho is about 425 W/m2 °C, r is at
least about 13. For a fluidized bed temperature of 870°C and a steam
temperature of 120°C, Tw = 174°C. This seems too low a temperature
to cause damage to the coil except, perhaps, by corrosion. For example,
at some regions of the tube, especially near the water inlet, metal
temperatures could be low enough to permit condensation of 1^504 on the
outside wall. Corrosion could result, particularly if the metal had pre-
viously been sensitized. This might be the mechanism by which cooling
coils were damaged so that they eventually failed.
104
-------
Another explanation for coil damage that was considered was that the
critical or "burnout" heat flux was exceeded. Above this flux, the
metal temperature rises sharply as a consequence of a drop in the inside
film coefficient. Under normal conditions, the actual heat flux should
be safely below burnout. However, it is possible that on occasion the
burnout flux was exceeded at certain regions of the coils.
There has been no evidence of erosion to the cooling coils. Slight
mechanical damage (bending) did occur on one occasion when the bed had
agglomerated because of high temperatures.
Bed Agglomeration
A problem which periodically occurred in the batch combustor was bed
agglomeration. This was a problem not only because of the high tem-
perature situation which it creates, but also because agglomerated
solids can cause a deformation of the cooling coils. Bed agglomeration
normally occurs when the ash in the coal begins to soften. This happens
at temperatures in excess of 1150°C. As the stone begins to agglomerate,
the solids mixing becomes poorer resulting in poorer heat transfer to the
cooling coils, and an uncontrolled temperature condition. However in
runs where bed agglomeration occurred, temperatures in excess of 1150°C
were usually not observed. This implies that localized hot spots which
are not picked up by the thermocouples may develop and cause the
agglomeration process to begin. These high temperatures could be caused
by a sudden surge of coal into the combustor. A sudden surge in the
coal feed to the combustor could be caused by an upset in the system
which controls the pressure differential between the coal injector and
the combustor or by an upset in the injection air rate. Upsets of
these types have been observed at times.
COMBUSTION RESULTS
SO2 Emissions
S02 emissions measured during batch runs varied with time during the
run as the bed of sorbent became more highly sulfated. At the beginning
of a run, the S0£ emissions were low. After a period of time, S02
levels increased and then, if the run lasted long enough, reached a
maximum determined by the sulfur content of the coal. Runs were made
at various S02 concentrations in the flue gas, the bed analyzed for
and the equivalent calcium to sulfur molar ratio calculated as
Ca/S = fraction S02 removed/fraction of calcium sulfated
The emissions at the end of the run were then plotted against the equiv
alent calcium to sulfur ratio calculated for conditions at the end of
the run. It had been planned to calculate equivalent calcium to sulfur
ratios for periods before the end of a run by calculating a sulfur mass
balance and using the final sulfation level as an anchor point.
105
-------
1600
Figure V-l 5
S02 EMISSIONS-LIMESTONE SORBENT
\
1400
SORBENT: GROVE LIMESTONE
PRESSURE: 800 kPa
TEMPERATURE: 850-920°C
SUPERFICIAL VELOCITY: 1.1-1.5 m/s
1200
s
Q.
a.
o
LJ
Z3
_l
U-
•z.
1000
- 800
z
o
K-
UJ
o
•z.
o
o
Cvl
o
CO
600
400
200
0
0
CALCIUM TO SULFUR MOLAR RATIO
106
-------
Figure V-16
S02 EMISSIONS-DOLOMITE SORBENT
900
750
Q.
S2.
co
<
o
LU
600
SORBENT: TYMOCHTEE DOLOMITE
PRESSURE: 800 kPa
TEMPERATURE: 840-970°C
SUPERFICIAL VELOCITY: 1.1-1.5 m/s
= 450
uj 300
z
o
o
CM
O
to
150
0
J_
0123
CALCIUM TO SULFUR MOLAR RATIO
4
107
-------
However, this approach was not workable since a portion of the bed was
lost by attrition and entrainment during the run and corrections for
bed loss could not be made accurately.
S02 emissions for runs using Grove limestone are shown in Figure V-15 as
a function of the equivalent calcium to sulfur ratio. Data are also
tabulated in Appendix H. Data were obtained at a pressure of 800 kPa,
at superficial velocities of 1.1 to 1.5 m/s over a temperature range of
850 to 920°C. Figure V-16 shows S02 emissions for runs using Tymochtee
dolomite made over the same range of conditions. As expected, emissions
were lower when dolomite was used.
S03 Emissions
S03/S02 levels were determined during several runs using a modified ver-
sion of the EPA test method. A flue gas sample was extracted from the
ducting just downstream of the combustor pressure control valve. The
S03 levels were much higher than expected and additional measurements
will be made to determine the source of the 803 formation, the combustor
itself, the flue gas outlet piping or the sample system. Possible errors
in the sampling and analytical techniques will also be investigated.
NOX Emissions
NOX emissions measured during batch unit runs are shown in Figure V-17 as
a function of percent excess air. The data were obtained at pressures of
800 kPa, and at temperatures generally in the range of 820 to 975°C. The
data were fairly represented by a single correlating line as seen in Fig-
ure V-17 despite some variation in temperature. There were some indica-
tions of a temperature effect, but additional data must be obtained to
determine the magnitude of the effect. Data are also shown in Figure
V-17 for three runs made with inert alundum beds. As seen, the NOX emis-
sions were very similar to, if not a little lower than emissions measured
from runs with limestone or dolomite beds. A similar results was repor-
ted by Argonne National Laboratory at 800 kPa pressure (4).
As in the case of the miniplant data, the NOX levels were well below the
EPA emission standard of 0.7 Ib N02/M BTU. At excess air levels of 15 to
20%, the range anticipated for commercial FBC units, the emissions were
in the range of 0.2 to 0-3 Ib N02/M BTU.
The data shown in Figure V-17 were measured using vertical cooling coils.
Earlier runs made with horizontal cooling coils showed higher NO levels,
although the levels were still well below 0.7 Ib NO2/M BTU at excess air
levels in the range of 15 to 20%. It is believed the higher NO levels
measured with the horizontal coils were caused by the higher bed tem-
peratures and combustion intensities (heat release rate per unit volume)
occuring near the coal feed point when the horizontal coils were used.
The NOX produced by the batch unit combustor was predominately NO. Less
than 5% was present.as"N02 and this was probably formed in the sampling
system, since the equilibrium concentration for N02 formation is very low
at the high temperatures occurring in the combustors.
108
-------
Figure V-17
NOX EMISSIONS
DQ
CM
O
co
QQ
CO
Z
O
CO
CO
LU
X
o
0.6
0.4
0.2
0
O
PRESSURE:800 kPa
TEMPERATURE: 820-975°C
• LIMESTONE OR DOLOMITE BED MATERIAL
O INERT BED MATERIAL
0
20
40 60
EXCESS AIR
80
100
120
140
-------
Trace Metal Emissions
One of the advantages claimed for fluidized bed combustion, compared
with conventional combustion techniques, is that a larger fraction of
trace elements present in the coal remain in the solid constituents
leaving the boiler. In other words, due to the flower combustion tem-
perature, smaller amounts of trace elements are vaporized and emitted
with the flue gas as a vapor. Elements which are of particular environ-
mental interest include antimony, arsenic, cadmium, chromium, cobalt,
manganese, mercury, nickel, selenium, tellurium, vanadium, beryllium
and lead.
In order to determine the capabilities of the fluidized bed combustion
system for retaining trace elements, several samples of solids were
submitted to General Activation Analysis (San Diego, California) for
neutron activation analysis. The samples analyzed were W. Virginia
coal (Arkwright), Tymochtee dolomite, and specimens of bed and overhead
solids from run no. 3675-2C. The overhead solids consisted of a com-
posite of the material collected in the off-gas cyclones and filter.
The temperatures in the first two cyclones were relatively low, around
300-400°C, and the third cyclone and filter were at even lower tem-
peratures, around 150°C. Therefore, no conclusions could be drawn
about the physical state of the trace elements leaving the combustor
since collection was at temperatures much lower than combustor tempera-
ture.
Table V-4 gives operating conditions for run no. 3675-2C and the results
of chemical analysis of the bed and overhead solids. Results of the
neutron activation analysis are given in Table V-5. In general, the
highest concentrations were found in the overhead solids which contains
flyash and some entrained stone and unburned carbon. For many elements,
actual concentrations could not be determined; instead, upper limits, or
the maximum amounts of elements that could be present in the sample,
were given. These are listed in Table V-6.
The analytical procedure used in the neutron activation analysis was as
follows: The samples were irradiated for 30 minutes in the TRIGA Mark I
Nuclear Reactor at a flux of 1.8 x 1012 n/cm2 sec. After a decay of one
hour, one day, eight days, and twenty days they were counted on a Ge(Li)
detector coupled to a multi-channel gamma-ray spectrometer. A second
portion of the samples were irradiated for 15 seconds in the TRIGA Mark I
Nuclear Reactor and counted in a Na(Tl) detector coupled to a 400 channel
gamma-ray spectrometer for short-lived isotopes. The data thus acquired
was examined and the detected elements identified. The data was then
processed in a UNIVAC 1108 computer.
Material balances were calculated for each of the elements for which
concentrations were determined in the coal, dolomite, bed, and overhead.
The balances accounted for elements present in these solid phases only.
Material which had vaporized and left the filter in the gaseous state
110
-------
TABLE V-4. DATA FOR RUN NO. 3675-2C FROM WHICH
SAMPLES WERE ANALYZED FOR TRACE ELEMENTS
Charge: Tymochtee dolomite, 7.6 kg (0.6 m settled bed) pre-caleined
prior to introduction of coal.
Coal; W. Virigina (Arkwright), 6 kg/hr fed for about 3 hours.
Average Run Conditions; T = 940 + 60°C
Sup. Vel. = 1.9 m/s
Pressure = 640 kPa
Excess Air « 61%
Analytical Results;
CaSO,
CaC0
CaO
MgO
Si°
Other C+ Error)
Bed
Weight %
37.7
2.6
18.8
28.2
4.4
7.7
4.0
-3.4
Mole %*
18.0
1.7
21.9
45.6
2.8
8.3
1.6
—
CaSO,
4
CaC03
CaO
MgO
A1203
Si°2
Fe2°3
C
Other
Overhead
(+ Error)
Weight % Mole Z*
15.8
4.2
5.2
18.4
7.0
14.1
4.9
19.0
11.4
4.4
1.6
3.5
17.4
2.6
8.9
1.2
60.2
—
* "other-free" basis
111
-------
TABLE V-5. ELEMENTS DETECTED BY NEUTRON ACTIVATION ANALYSIS
(All concentrations in ppm)
Element
Al
*Sb
Ar
*As
Ba
Br
*Cd
Cs
Cl
*Cr
*Co
Dy
Eu
Fe
La
*Mn
K
Rb
Sm
Sc
Na
Sr
*Te
Th
U
Arkwright Coal
27000
*
29.
3.
12.
2.
*
1050
17
2.
„
13800
2.
20
599
3.
2.
784
205
1.
*
+
253 +
1 +
7 +
4 +
03 +
349 +
+
+
74 +
171 +
+
74 +
+
+
74 +
09 +
+
+
75 +
652 +
5400
.063
6.1
.78
2.6
.41
.091
220
3.4
.57
.036
2800
1.1
4
120
1.3
.42
160
44
.35
.30
Tymochtee Dolomite
10900
.0527
13.2
.566
6.75
.439
447
4.23
1.03
.0598
3240
42
2180
12.2
.658
.952
303
130
2.81
.58
2.23
+ 2200
+ .015
+ 2.9
± -17
+ 1.4
+ .091
+ 96
+ .85
+ .21
+ .013
+ 650
+ 8.4
+ 440
+ 2.5
+ .13
+ .19
+ 61
•f 29
+ .63
+ .12
+ .45
27000
18
8
107
2
787
52
3
19600
102
1280
9
2
364
278
3
1
3
Bed
+
.501 +
.9 +
.02 +
+
.75 +
.403 4-
+
.9 +
.05 +
.135 +
+
+
+
.17 +
.29 +
+
+
.6 +
.58 +
.13 +
5400
.11
4.2
1.6
24
.75
.094
170
11
.61
.028
3900
20
260
2.1
.46
73
60
.91
.32
.67
71300
55
9
2
627
133
6
39900
105
4040
31
7
3200
690
6
4
Overhead
+
.606 +
+
.4 +
.04 +
+
+
.81 +
.459 +
+
H-
+
.1 +
.63 +
+
+
.71 +
.41 +
14000
.13
12
2.0
.42
170
27
.14
.098
8000
21
820
6.5
1.50
640
150
1.3
.94
* Of particular environmental interest
-------
TABLE V-6. NEUTRON ACTIVATION ANALYSIS - UPPER LIMITS
(All concentrations in ppm)
Element^
Ca
Ce
*Cu
Er
*F
Gd
Ga
Ge
Au
Hf
Ho
I
In
Ir
Kr
*Pb
Lu
Mg
*Hg
Mo
Nd
Ne
*Ni
Nb
N
Os
0
Pd
*P
Pt
Pr
Re
Rh
Ru
Sm
*Se
Si
Ag
S
Tg
Tb
Tm
Ti
Arkwright Coal
<12000
33
100
14
9500
16
130
.034
.35
-49
3.1
2.1
.075
47
.095
18000
.82
35
6.8
72000
140
1700
<.77
6.0
12
76
Tymochtee Dolomite
<56000
12
46
5.8
5300
6.2
46
.011
.38
.21
1.6
1.5
.0093
23
.038
81000
.27
12
2.9
220000
160
610
<.15
3.3
30
33
Bed
<68000
34
60
6.2
9400
12
72
.028
.29
.35
4.2
3.6
.10
31
.1
61000
.7
31
6.2
45000
220
1000
<.40
5.4
43
44
.35
290
1.2
2.8
2.4
320000
.34
.29
.39
4300
.039
1000
.52
.33
1.1
_t_
.086
.095
.11
11000
.15
250
1.1
1.7
1.7
610000
.28
.35
.18
4100
Overhead
<17000
130
380
26
27000
30
340
4
.98
1.8
5.9
4.5
.28
130
59000
2.
120
23
140000
590
5400
16
93
300
.95
810
4.2
8.3
4.0
440000
1.0
1.2
,94
8300
113
-------
TABLE V-6. (Continued) NEUTRON ACTIVATION ANALYSIS - UPPER LIMITS
(All Concentrations in ppm)
Element Arkwright Coal Tymochtee Dolomite
Bed
Overhead
w
*v
Xe
Yb
Y
*Zn
Zr
3.3
140
110
.26
4800
180
1300
1.7
120
51
.068
3700
82
530
2.1
160
130
.21
6400
130
1400
13.0
420
400
.90
9300
790
7100
* Of particular environmental interest
114
-------
TABLE V-7. MATERIAL BALANCES FOR TYPICAL COMPONENTS
In Out
Element
Al
Sb
Ar
As
Cs
Cl
Cr
Eu
Fe
Mn
K
Rb
Sc
Na
Sr
Th
U
Ca
S
Coal
•i
4.85 x 10 kg
-6
4.54 x 10
5.22 x 10~4
6.67 x 10~5
6.26 x 10~6
1.89 x 10~2
-4
3.05 x 10
3.07 x 10~6
_i
2.48 x 10
-4
3.59 x 10
_2
1.08 x 10
_5
6.71 x 10
3.76 x 10~5
_2
1.41 x 10
0
3.68 x 10
3.14 x 10~5
c
1.17 x 10
recovery (based on
recovery (based on
Dolomite
_2
8.16 x 10 kg
_7
3.97 x 10
9.93 x 10~5
4.26 x 10~6
3.31 x 10"6
3.37 x 10~3
_5
3.18 x 10
4.50 x 10~7
-2
2.44 x 10
-4
3.16 x 10
_2
1.64 x 10
_5
9.21 x 10
7.17 x 10~6
_3
2.28 x 10
-4
9.80 x 10
4.37 x 10~6
_5
1.68 x 10
analysis by atomic
Bed
_2
7.08 x 10 kg
-6
1.32 x 10
4.99 x 10~5
2.11 x 10~5
1.06 x 10~6
2.07 x 10~3
-4
1.39 x 10
3.55 x 10~7
o
5.17 x 10
-4
2.69 x 10
_3
3.37 x 10
_5
2.41 x 10
6.03 x 10~6
-4
9.57 x 10
-4
7.30 x 10
4.15 x 10~6
-6
8.26 x 10
absorption) = 69.5%
wet chemical analysis, includes sulfur
Overhead
_1
2.99 x 10 kg
-6
2.54 x 10
2.31 x 10~4
3.95 x 10~5
8.57 x 10~6
2.63 x 10~3
-4
5.58 x 10
1.93 x 10~6
1
1.67 x 10
-4
4.40 x 10
-2
1.70 x 10
-4
1.31 x 10
3.20 x 10~5
_2
1.34 x 10
_3
2.89 x 10
2.82 x 10~5
_5
1.85 x 10
in flue gas) = 94%
Recovery
65.3
78.2
45.2
85.9
100.0
21.2
207.0
64.8
80.5
95.5
74.8
97.2
85.1
88.1
77.6
90.4
93.9
-------
was not included. Table V-7 gives the results. Recoveries are quite
high, particularly when two factors are taken into account. First, it
is estimated that 10-15% of the solids (bed and overhead) were lost
either because they remained in the combustor or were spilled prior to
weighing. This would, of course, reduce the recoveries. Secondly, as
can be noted in Table V-7, the concentrations reported are accurate
only to about + 20 percent.
The important conclusion to be drawn from Table V-7 is that the
fraction of trace elements retained by solid constituents was generally
high. There are several exceptions, particularly chlorine; however,
chlorine might be expected to volatilize at lower temperatures than
the other elements. The recovery of chromium, 207 percent, was probably
caused by addition of this element from chromium alloys used in the
materials of construction of the combustor (e.g., cooling coils).
In addition to the recovery, or the fraction of elements retained in
solid phases (stone + flyash), it is also of interest to know the fraction
of each element fed with the coal that was retained by the dolomite only
(not flyash). This term was called the "retention." Retention could not
be determined precisely because assumptions had to be made. One assump-
tion was that the dolomite which was found in the overhead had the same
concentration of a particular element as did the bed. Also, it was esti-
mated that 1.82 kg out of 4.20 kg total overhead collected was entrained
stone.
Table V-8 gives the retentions as the percentage of element fed with the
coal that was retained by the dolomite. The retentions are quite vari-
able with the highest (not counting chromium) being 47 percent for
arsenic. Retentions for the alkali metals were negative. This means
that there was a net loss of these elements from the dolomite during
the run. In general, retention by the bed was low, indicating that
most of the trace elements were lost from the bed as flyash particulates
(or possibly as a vapor in the case of some elements.)
This study did not indicate the point where the trace elements present in
coal and limestone will be removed from the combustion system, since
that determination must be made by sampling at more representative
cyclone temperatures. However, it does indicate that the trace elements
at some point, will be largely in the form of particulates, and particulate
removal from the flue gas must be sufficient to prevent environmental
problems. The fate of the more volatile elements such as mercury or
alkali metals must be studied separately since different sampling and
analytical techniques are required.
Argonne National Laboratory reported mass balances and recoveries for
trace elements calculated from neutron activation analysis (5).
Table V-9 compares these results with those reported above. The
agreement is quite good.
116
-------
TABLE V-8. RETENTION BY STONE OF ELEMENTS PRESENT IN COAL
Element Retention, %*
Al 7.9
Sb 40.3
Ar -2.9
As 47.1
Br -17.2
Cs -24.3
Gl 0.7
Cr 66.6(1)
Eu 4.9
Fe 25.3
Mn 38.4
K -99.1
Rb -76.5
Sc 8.1
Na -4.1
Sr 7.1
Th 8.5
U -24.8
* r, *.- Weight of element present in coal picked up by stone (bed + entrained stone) Tr __
** Retention — TT, ^^A-^^J^TI •& -Luu
Wt. of element fed with coal
(based on assumption that entrained stone has same concentration of elements as bed) .
(1) May be high due to high material balance. See Table V-7
-------
TABLE V-9. COMPARISON OF EXXON AND ARGONNE N.L.
DATA ON TRACE ELEMENT RECOVERIES (a)
Recovery,
Element
As
Br
Fe
K
Mn
Na
Sc
Exxon
86
21 (Cl)
80
75
96
88
85
ANL
85
18
100
90
130
96
97
(a) Recovery = percentage of element
entering coinbustor that
can be accounted for in
solids leaving combustor.
118
-------
CO Emissions
CO emission levels were measured during the batch unit coal combustion
runs. Although the data are not conclusive, it appears that the CO
level is a function of the steadiness of the coal feed rate, the average
bed temperature, and the excess air level. The average CO level for all
the runs was 946 ppm. However, in some cases, it was obvious that the
feed rate was not steady or the CO levels were inordinately high for
other reasons. The data were then treated statistically, and data were
rejected which were greater than the mean by more than twice the error
limit. This treatment gave an average CO level of 409 + 276 ppm (Is
limit) for all runs. For the most recent runs made, this treatment
indicated the average CO concentration was approximately 220 ppm. As
further tests are performed, correlations will be developed to relate
the CO level to the bed temperature and excess air level under more
closely controlled conditions of coal feeding.
Particulate Emissions
Particulate Loadings - The average bed outlet particulate loading
obtained by summing the particulates captured in the cyclones and filter
was 8.5 gr/scf. The average outlet loading from cyclone #1 was 2.0
gr/scf and from cyclone #2 was 0.9 gr/scf. However, in some cases, it
was obvious that the cyclones were not operating properly or loadings
were inordinately high for other reasons. The data were than treated
statistically, and data were rejected which were greater than the mean
by more than twice the error limit. This treatment gave the following
results:
Bed outlet loading - 6.9 +_ 1.4 gr/scf (Is limit)
Cyclone #1 outlet loading"- 0.70 + 0.35 gr/scf (Is limit)
Cyclone #2 outlet loading -0.33+0.26 gr/scf (Is limit)
Cyclone Efficiency - Overall collection efficiencies for the two stage
cyclone system have generally been over 90%. Low efficiencies for a
given run could usually be traced to a plugged cyclone dipleg. The
average collection efficiency based on the average grain loadings cal-
culated in the previous section was 90% for cyclone #1 and 53% for
cyclone #2, giving an overall combined efficiency of 95%.
Particle Size Distribution
A sieve analysis was used to determine the particle size distribution
of samples collected from the overhead from run no. 3675-46C. This run
was made with Grove limestone. The samples studied were collected in the
first stage cyclone and on the off-gas filter. Chemical analysis indica-
ted that the amount of carbon present was 51.4% for the material col-
lected by the first stage cyclone and 30.8% for the filter solids. Fig-
ure V-18 shows the results of the sieve analysis for the first stage
119
-------
Figure V-l 8
PARTICLE SIZE DISTRIBUTION
OVERHEAD SAMPLES
LU
to
LU
O
1—
Q.
•2.
<
X
H-
to
to
LU
I
M«J
^O
1—
_L-
LU
LU
| —
J__
<
_J
ZD
ID
O
inn
90
80
70
60
50
40
30
20
10
— • - *
^
^^•-""•""^
^^K^^
*s^
/
S* CYCLONE # 1
• ^
/•
y*
/
- /
•
—
-
i I i i l i ii
1
100 200 300 400 500 600 700 800 900
, , , . . , . . , MICRONS)
230 140 100 70 50 40 30 25 20
(MESH SIZE)
PARTICLE SIZE
-------
cyclone sample. The particle size for the 50% point was about 130
microns. The size distribution measured for the filter sample was dis-
torted by prilling of the fine particulates during sieving.
A particle size distribution for particulates collected in the overhead
during run no. 3675-23C using optical microscopy was also determined.
Microphotographs of the solids collected showed that two classes of
solids were distinguishable, a white "popcorn" solid and a black solid.
The black solids were more prevalent in the filter sample and are
probably carbon and ash, while the white material is stone, possibly
sulfated. The measurements indicated that the particulates ranged from
1 to 30 microns in the cyclone samples and 1 to 20 microns in the fil-
ter sample.
A problem in determining the size distribution using this technique is
that the particles are skeletal in nature and not easily defined.
Agglomerates may be present and the range of sizes present is very
large. The particles also appear to be fragile and break into very
small pieces upon handling. These factors may have caused a smaller
average particulate size for the first stage cyclone sample compared
to that measured by sieving.
Combustion Efficiency
Carbon combustion efficiencies measured in the batch unit varied from
87 to 98%. Combustion efficiency is a function of excess stoichiometric
air and temperature as shown in Figure V-19. The batch unit was
operated at superficial velocities of 1.1 to 1.5 m/s, at temperatures
of 750 to 970°C, and at pressures of about 800 kPa.
Another condition which affected combustion efficiency was the steadi-
ness of coal feeding. When coal feeding was unsteady, combustion
efficiencies were consistently lower. For the runs shown in Figure
V-19, coal feeding was fairly steady, but variations in coal feeding
were probably responsible for some of the data scatter apparent in
Figure V-19.
Low bed temperatures were also found to cause low combustion efficien-
cies. At temperatures lower than 840°C, combustion efficiency decreased
significantly as seen in Figure V-19. At the lower temperatures a sharp
increase in CO emissions occurred. At temperatures below 600-650°C, it
was difficult to maintain coal combustion.
Changing the coil design from horizontal to vertical orientation did
not appear to have any significant effect on combustion efficiency.
The results shown in Figure V-19 are consistent with those reported
previously from miniplant runs. A more detailed comparison of results
from the two units is made in Section VI.
121
-------
100
Figure V-19
COMBUSTION EFFICIENCY
95
ro
N3
O
z
UJ
O
u.
LL.
UJ
90
PRESSURE: 800 kPa
SUPERFICIAL VELOCITY: 1.5-1.5 m/s
TEMPERATURE: A 750°C
• 840-880°C
• 910-970°C
CO
:D
CD
2
O
O
85
80
0
20
40
60
80
100
120
140
EXCESS AIR
-------
As In the case of the miniplant, the loss of combustion efficiency was
due largely to unburned carbon particles entrained from the combustor.
CO loss accounted for only 2% of the carbon losses.
Heat Transfer Coefficients
Overall heat transfer coefficients were measured during coal combustion
for vertical coils no. 1 and no. 2 (27-62 cm and 67-102 cm above the
fluidizing grid, respectively). Tubing O.D. for these tests was 0-.64 cm.
Results of measurements made during three runs are given in Table V-10.
During the time data were taken, a sufficient flow of water was pumped
into each coil to prevent any phase change. Temperature rise of the water
and flowrate were carefully measured to determine the heat absorbed by
each coil. Measurements were repeated several times to achieve good ac-
curacy and to determine variability. Overall coefficients averaged 395
W/m2°c for coil 1 (lower coil) and 452 W/m2°C for coil 2 (upper coil) .
The reason for the higher coefficient for coil 2 is not certain but may
be due to slightly more vigorous mixing of solids higher in the combustor.
Heat transfer coefficients were also measured while the bed was heated
using the pre-heat (propane) burner. Table V-ll gives these coeffic-
ients for four vertical and two horizontal serpentine coils. Three of
the vertical coils were always immersed in the bed whereas both hori-
zontal coils were always in the freeboard. Operating conditions for
the combustor are also given. Bed-to-tube coefficients during pre-
heating were somewhat lower than those given in Table V-10 for coal
combustion because bed temperatures, and hence the radiation component
of heat transfer, were lower. Coefficients for coils in the freeboard
averaged 106 W/m^°C. This is large enough for the coils located in the
freeboard to remove enough heat to cause a sharp drop in the flue gas
temperatures.
The overall coefficients for the original horizontal serpentine coils
(located in the bed) were estimated to be 230-310 W/m2°C under coal
combustion conditions at temperatures of 850 to 900°C. This was lower
than the coefficients for vertical coils and is probably due to the
much slower motion of bed solids with horizontal coils.
It should be appreciated that the overall coefficients reported in this
section are nearly the same as the film coefficients outside of the
tubes, a consequence of the fact that the inside film coefficients are
very high. The outside coefficients are probably not more than 10 per
cent higher, at most, than the overall coefficients.
Coal and Sorbent Effects
Three types of coal have been burned in the batch combustor. Most work
has been with West Virginia Coal (Arkwright), which is highly caking;
the other coals tested were slightly caking Wyoming coal and moderately
caking Illinois No. 6 coal. The foremost reason for using these coals
was to determine if the temperature profile in the bed was affected by
123
-------
TABLE V-10. HEAT TRANSFER COEFFICIENTS (BED TO TUBE)
MEASURED DURING COAL COMBUSTION
Vertical Coils
0.64 cm O.D.
Overall Coefficients
Run No.
3675-28C
-31C
-32C
Coil 1 Coil 2 Average Bed
'(W/m2°C) (W/m2°C) Temperature (°C)
395 + 23
392 + 12 432 + 5
399 +3 472 + 11
840
895
885
Superficial Bed
Velocity (m/s)
1.1
1.6
1.6
Coil Positions in Combustor
Coil 1 27-62 cm above fluidizing grid
Coil 2 67-102 cm above fluidizing grid
124
-------
TABLE V-ll. HEAT TRANSFER COEFFICIENTS MEASURED DURING BED PRE-HEATING
Coil
No.
1
2
3
4
5
Overall Coefficients (W/m2°C)
Coil Orientation
Vertical
Vertical
Vertical
Vertical
Horizontal Serpentine
Horizontal Serpentine
Run No. 1
340
320
300
300
120
Run No. 2
300
330
280
200
100
100
Run No. 3
330
280
280
280
100
110
Run No. 4
350
350
300
110
100
Operating Conditions
Run
No.
1
2
3
4
Superficial
Inside Coil Velocity tn/sec.
Water-steam
Water-steam
Water- steam
Water only
1.16
1.06
1.62
1.15
Average Bed
Temperature °C
600
530
550
455
Expanded
Bed Height, m
1.65
1.30
1.75
1.27
Pressure
kPa
810
490
500
500
Coils
Immersed
In Bed
1,2,3,4
1,2,3
1,2,3,4
1,2,3
Coil O.D. = 0.64 cm
-------
coal type, in particular by the tendency of the coal to cake when
heated. The coal comparison study was made early in our work when the
combustor was fitted with tightly packed horizontal serpentine coils.
There was a severe longitudinal temperature gradient across the bed
at the time.
There was a slight improvement in temperature profile, i.e., a reduction
in the gradient, when Wyoming coal was substituted for West Virginia
coal. Figure V-20 shows the profile for the run made with Wyoming coal
and an average profile for two runs made with W. Virginia coal. A 0.6 m
(settled) bed of half-calcined Tymochtee dolomite was used for all three
runs. The energy input was slightly higher in the run with Wyoming coal,
which could indicate that the improvement shown in Figure V-20 might
have been even better with equivalent feed rates. It should be noted,
however, that even though the use of low caking coal improved the tem-
perature profile somewhat, there was still a temperature drop of about
280°C over the expanded bed. The effect of coal type on temperature
profile was very small when compared to the effect of changing the cool-
ing coils from a horizontal serpentine to a vertical configuration.
The run made with Illinois coal resulted in a highly non-typical
temperature profile because much of the coal burned above the bed. A
number of factors could have caused this but a characteristic of the
coal is not believed to have been one of them.
The effect on temperature profile of varying particle size of the coal
was also investigated. Since larger particles burn slower, a run was
made in which fines smaller than 70 mesh (about 40% of the coal) were
removed from Arkwright coal. The result was a slightly flatter profile
with the screened coal. This run was also made with horizontal coils.
Again, the effect was small when compared to the effect of cooling coil
conf iguration.
In addition to different coals, several sorbents were also tested,
including Grove limestone (BCR No. 1359), Tymochtee dolomite, and Pfizer
dolomite (BCR No. 1337). Stones varied with respect to sulfur removal
abilities (discussed on page 135).and resistance to attrition. The
resistance to attrition was indicated indirectly by the amount of sorbent
entrained from the bed during a combustion run. Entrainment rates are
inversely proportional to particle size, and since the sorbents were all
screened initially to the same size distribution, the entrainment rates
are a relative measure of the degree of attrition. For the stones
tested, the average entrainment rates were:
Grove limestone (BCR 1359) 15 wt. percent/hour
Tymochtee dolomite 20-25 wt. percent/hour
Pfizer dolomite (BCR 1337) 50 wt. percent/hour
Rates are given as the weight percentage lost per hour of material
remaining in the fluidized bed. For example, if a = fractional loss per
hour (a = 0.15 for Grove limestone), then W = Woe-at, where W =
126
-------
Figure V-20
COMPARISON OF TEMPERATURE PROFILES-WYOMING AND VIRGINIA COAL
1100
1000
o 900
LJ
UJ
Q.
800
700
600
WYOMING COAL
W.VIRGINIA COAL
(AVG.2 RUNS)
1 I I I I I I
I I I
0 10 20
30 40 50 60 70 80
HEIGHT ABOVE GRID, CM
90 100 110
-------
weight of bed remaining in batch reactor after time t (hours), and Wo =
initial bed weight. These rates are fairly high, especially for the
dolomites. However, lower rates are expected for continuous operation.
This will be studied in future runs in the miniplant.
Samples of sulfated bed material were also screened after runs and the
particle size distribution compared to that for the starting materials.
Table V-12 shows results for runs made with Grove limestone and
Tymochtee dolomite.
TABLE V-12. PARTICLE SIZE DISTRIBUTION OF SULFATED SORBENTS
„_ _ P£R CENT LESS THAN SCREEN SIZE
US Screen Size 6 8 12 16 20 25 30 40 50
Grove Limestone
New ! — 98 48 24 10
Sulfated — 100 92 67 35 17 7 4
Tymochtee Dolomite
New 2 ~~ 10° 60 31 10
Sulfated 97 93 84 67 50 39 24 9 3
1 Run 675-11C
2 Run 675-8C
The particle size whose terminal velocity corresponded to the fluidiza-
tion velocity of the runs was 25 to 30 mesh. Therefore, a good part of
the particulates below that mesh size were entrained from the bed.
However, the data in Table V-12 still show that significant attrition
did occur and the dolomite was more suceptible to attrition than the
limestone.
It was observed that even stone from the same quarry can display variable
attrition resistance. Tymochtee dolomite had always been a satisfactory
stone with regard to attrition resistance but a recently arrived ship-
ment showed very high entrainment rates. It was assumed that a change
in properties of the stone since the prior shipment was responsible;
perhaps the stone was being mined from a different portion of the quarry.
Component Balances
Sulfur and calcium balances were made for a number of batch unit runs.
The results are presented in Tables V-13 and V-14. The weight of sul-
fur into the combustor was determined from the coal feed rate, the run
length, and the sulfur content of the coal. The calcium input was
determined from an initial bed analysis. For the solids output streams
(material collected in the cyclone diplegs, off-gas filter, and the
final bed), the quantity accumulated was weighed and then analyzed for
128
-------
TABLE V-13. SULFUR BALANCES FOR BATCH COMBUSTOR
(All weights in kg)
Sulfur Out
Run No.
145C
2C
3C
4C
5C
17C
18C
23C
25C
28C
29C
31C
32C
46C
56C
Sulfur
In Coal
0.631
0.468
0.527
0.159
0.331
0.536
0.633
0.513
1.027
0.880
0.884
0.717
1.003
0.386
0.719
Flue
Gas (1)
0.177
0.050
0.159
0.014
0.204
0.007
0.144
0.014
0.003
0.051
0.132
0.019
0.031
0.120
0.101
Overhead
Solids
0.091
0.154
0.095
0.132
0.036
0.058
0.086
0.081
0.128
0.183
0.123
0.271
0.226
0.178
0.057
Final
Bed
0.254
0.232
0.100
0.023
0.027
0.375
0.530
0.269
0.311
0.523
0.829
0.384
1.328
0.235
0.437
Total
0.522
0.436
0.354
0.169
0.267
0.440
0.760
0.364
0.442
0.757
1.084
0.674
1.585
0.533
0.595
% Sulfur
Balance
83
93
67
106
81
82
120
71
43
86
123
94
158
138
83
Avg. 91
+ 25%
(excluding
run 32C)
(1) Based on average S02 concentration
129
-------
TABLE V-14. CALCIUM BALANCES FOR BATCH COMBUSTOR
U)
o
Run
No.
145
2C
3C
4C
5C
Charge
(stone, kg)
TD, 7
TD, 7
G, 7
TD, 7
TD, 7
.54
.54
.72
.54
.54
Input
(kgCa)
1.58
1.58
3.02
1.58
1.58
Wt. Bed
Recovered,
kg
2.63
2.63
4.72
0.82
4.65
Wt. fr.
Ca In Bed
0.220
0.256
0.433
0.285
0.265
Wt. Ca
in Bed
kg
0.58
0.67
2.05
0.23
1.23
Wt. Overhead
Solids
Recovered, kg
5.04
4.20
4.23
4.72
1.81
Wt. fr.
Ca in
Overhead
0.170
0.101
0.090
0.218
0.081
Wt. Ca in
Overhead
kg
0.86
0.42
0.38
1.03
0.15
Total
Output
(kgCa)
1.44
1.09
2.43
1.26
1.38
<9
/o
Calcium
Balance
91
69
80
80
87
Avg. 81
± 8%
TD - Tymochtee dolomite, 20.9 wt. % Ca
G - Grove limestone, 39.1 wt. % Ca
-------
sulfur, sulfate, and calcium. The quantity of sulfur in the flue gas
was determined from data obtained by continuously monitoring the
off-gas for SC>2 using a Beckman infrared analyzer.
After neglecting data outside che +2s limits, the sulfur balances
averaged 91 + 25% (Is limit) and the calcium balances averaged 81 +_ 8%
(Is limit). Possible causes of the low balances are uncertainties in
the analysis of the solids and loss of solids during removal of the
bed and flyash. In the case of the sulfur balances, loss of SC>2 through
863 formation is not likely to have a major effect on the sulfur balances
because the fraction of total sulfur entering the system that exits in
the flue gas is normally small.
131
-------
SECTION VI
DISCUSSION OF RESULTS
S02 EMISSIONS
S02 emission data measured in the batch, unit using limestone sorbent
were compared to data reported by the National Coal Board (NCB) (6)
and Argonne National Laboratory (ANL) (7). The data previously shown
in Figure V-15 were replotted in Figure VI—1 as percent reduction
in S02 along with the NCB and ANL results. NCB measured emissions at
a lower velocity, 0.6 m/s, compared to 1.1 to 1.5 m/s used in the batch
combustor program. The pressure was also lower in the NCB study,
600 kPa vs. 800 kPa and the limestone particle size was smaller, 350 ym
median diameter vs. 1400 1-im. These factors should have made desul-
furization more effective in the NCB study, but as seen in Figure VI-1,
the emissions reported by NCB were equivalent to those measured from
the batch combustor. The two studies were made with different coals
and limestones and this factor may be partly responsible for the
anomaly.
Results reported by ANL were measured at a velocity of 1.1 m/s, and
at a pressure of 800 kPa using limestone with a median particle size
of about 700 ym. The coal and limestone were from the same sources
as those used in the batch unit studies. Again, the results of the
ANL study and the batch unit results compare very well. However, in
this case, the data were obtained under more comparable conditions
and good agreement would be expected.
The emissions measured from the miniplant after the sample system was
operating properly as mentioned previously corresponded very well with
data from the batch unit. This is also shown in Figure VI-1.
From the results shown in Figure VI-1, a calcium to sulfur molar ratio
of 2.0 appears to be sufficient with limestone to meet the EPA emission
standard of 1.2 Ibs S02/M BTU for the 2.6% sulfur coal used in this
study. A comparison was also made between S02 emission data reported by
ANL (4) and data obtained from the batch combustor using Tymochtee dolo-
mite sorbent. The comparison is shown in Figure VI-2. In this case the
same coal and sorbent were used in both studies and the operating
conditions were comparable except for sorbent particle size. Both
studies were carried out at pressures of 800 kPa and at superficial
velocities of 1.1 to 1.6 m/s. However, the ANL study used a dolomite
with a median particle diameter of 750 ym compared to 1400 ym used in
the batch unit. The curve in Figure VI-2 represents the best line
through the batch combustor results. As seen, the data from the two
combustors compare very well. The NCB study reported much lower emis-
sions using dolomite sorbent, but this is not surprising considering
the difference in operating conditions. From the results shown in
Figure VI-2, a calcium to sulfur molar ratio of 1.5 appears to be
132
-------
100
Figure Vl-l
COMPARISON OF S02 REMOVAL
RESULTS - LIMESTONE SORBENT
90
80
70
UJ
cr
to
50
40
• EXXON BATCH UNIT
A NCB
• ARGONNEN.L.
a EXXON MINIPLANT
30
600-800 kPa
750-920°C
20
10
_L
I
2 3
Ca/S (MOLE/MOLE)
133
4
-------
Figure VI-2
COMPARISON OF S02 EMISSIONS
FROM BATCH UNIT AND ARGONNE N.L. STUDY
DOLOMITE SORBENT
100
80
60
cc
CM
O
C/)
LU 40
o
a:
UJ
a.
20
0
BATCH UN IT DATA
ANL DATA
_L
J_
0123
CALCIUM TO SULFUR MOLAR RATIO
134
-------
sufficient with, dolomite to meet the EPA S0£ emission standard for the
2.6% sulfur coal used in this study- This corresponds to an S(>2 con-
centration in the flue gas of 550 to 600 ppm.and a percent S(>2 retention
of 68%.
Emissions resulting from the use of limestone and dolomite in the
batch unit were compared and the results are shown in Figure VI-3.
As expected, emissions resulting from the use of dolomite were lower
than those obtained with limestone. However, the emissions differ by
only about 300 ppm at a Ca/S ratio of 2. NCB and ANL reported a much
larger difference, on the order of 500 to 600 ppm. The reason for
this difference requires further investigation.
NOV EMISSIONS
24.
NOX emissions measured in the miniplant and batch combustors are com-
pared in Figure VT-4. These data were shown previously in Figures IV-32
and V-17. As seen, the comparison is very good and the data from both
units can be correlated by a single curve. In addition, the range of
NOX data reported by NCB (8) is also shown and also agrees well with
miniplant and batch unit results. The NOX levels measured in this
study are well below the current EPA emission standard of 0.7 Ib
NO£/M BTU even at excess air levels as high as 140%. The range of
excess air levels anticipated for commercial FBC units is 15 to 20%.
In this range, it is expected the NOX emission levels will be on the
order of 0.2 to 0.3 Ib NO£/M BTU. These levels are significantly lower
than those reported previously for atmospheric pressure fluidized bed
combustion (9). In this case, NOX emissions on the order of 0.4 to
0.5 Ib N02/M BTU were reported.
COMBUSTION EFFICIENCY
A comparison of combustion efficiencies measured in the batch unit
and the miniplant is given in Figure VI-5- The data agree well
although the miniplant results do not show a temperature effect as do
those from the batch unit. However, as shown in Figure IV-33 the
miniplant results at excess air levels less than 40% exhibit a fair
degree of data scatter. The data obtained in this range also were
obtained at a lower temperature, so a true measure of the effect of
temperature has not as yet been established in the miniplant. Also,
these early miniplant results were affected by operating problems
with the cyclones. Further operation of the miniplant under more
closely controlled conditions is necessary before a true measure of
combustion efficiency can be established in this unit.
Some difficulties also occurred in the miniplant carbon recycle system.
This system recycles fines collected in the first stage cyclone back
to the combustor to improve combustion efficiency. Since this recycle
system was not always operating properly during early runs, the results
shown in Figure VI-5 are probably more indicative of a once through
system than a carbon recycle system.
135
-------
Figure VI-3
COMPARISON OF S02 EMISSIONS
FROM LIMESTONE AND DOLOMITE SORBENTS,
BATCH UNIT DATA
900
750
Q_
- 600
CO
<
o
UJ
u_
5 450
z
o
I—
<
LIMESTONE
DOLOMITE
300
o
z
o
o
CM
o
CO
150
0
0
CALCIUM TO SULFUR MOLAR RATIO
136
-------
Figure VI-4
0.8
COMPARISON OF NOX EMISSIONS
0.6
CM
o
0.4
• MINI PLANT
• BATCH UNIT
-- NCB -DATA RANGE
1
1
20
40
60 80
EXCESS AIR (%)
100
120
140
-------
I—
w
00
CO
D
GO
S
O
o
Figure Vl-5
COMPARISON OF COMBUSTION EFFICIENCIES
100
910-970°C
85
840-880°C
BATCH UNIT
MINI PLANT
80
_L
0
20
40 60
EXCESS AIR
80
100
120
140
-------
Results reported by ANL (4) show combustion efficiencies generally
in the same range as those given in Figure VI-5- However, the NCB (8)
reports much higher combustion efficiencies, usually in the range of
98.5 to 99.5%. These higher combustion efficiencies may be caused by
longer residence time in the NCB combustor.
A target combustion efficiency of 98.5 to 99% has been suggested by
Westinghouse Research Laboratory (10). The results obtained in the
Exxon units to date have not reached that level. A more effective
carbon recycle system, operation at higher temperatures, or use of
a carbon burn up cell, may be necessary to reach this target.
HEAT TRANSFER COEFFICIENTS
Heat transfer coefficients measured in the batch combustor were about
30% higher than those measured in the miniplant. This is shown in
Table VI-1.
TABLE VI-1. COMPARISON OF HEAT TRANSFER COEFFICIENTS
Unit
Batch
Miniplant 1.91
Coil
Diam.
(cm)
0.64
Coil
Position(l)
1A
IB
Bed
Temp.
(°C)
840-895
840-895
Superficial
Vel. (m/s)
1.1-1.6
1.1-1.6
1A
IB
2A
870-950
870-950
870-950
1.8-2.1
1.8-2.1
1.8-2.1
Heat Transfer
Coefficient
(W/m2K)
395
452
318
340
330
(1) Coil positions are in ascending order from the bottom
The reason for the higher coefficients measured in the batch unit is
probably due to the smaller diameter of batch unit coils, partly
offset by a lower superficial velocity (11) .
NCB also reports heat transfer coefficiencts of about 400 W/m2K (8) .
However, the particle size used in the NCB study was much less and the
superficial velocity was also much lower than used in the present
work.
139
-------
SECTION VII
REFERENCES
1. Skopp, A., Nutkis, M. S., Hammons, G. A., and Bertrand, R. R.,
"Studies of the Fluidized Lime-Bed Coal Combustion Desulfurization
System." Report to U.S. Environmental Protection Agency, Exxon
(Esso) Research and Engineering Company. December 31, 1971.
2. Hodges, J. L., "Predicting Temperature Profiles in Fluidized Bed
with Internal Baffling," worked performed under NSF Grant
HES75-05125, Exxon Research and Engineering Company, August 1975.
3. Hoke, R. C., Nutkis, M. S., Ruth, L. A., and Shaw, H., "A
Regenerative Limestone Process for Fluidized-Bed Coal Combustion
and Desulfurization." EPA Report No. EPA-650/2-74-001, Exxon
Research and Engineering Company, January, 1974.
4. Jonke, A. A., et al., "Reduction of Atmospheric Pollution by the
Application of Fluidized-Bed Combustion," EPA Report No.
EPA-650/2-74-104, Argonne National Laboratory, September, 1974.
5. Vogel, G. J., et al., "A Development Program on Pressurized
Fluidized-Bed Combustion." Annual Report to ERDA Office of
Fossil Energy. Argonne National Laboratory. July, 1975.
6. Roberts, A. G., et al., "Fluidised Combustion of Coal and Oil
Under Pressure," Institute of Fuel Symposium Series No. 1 (1975)
Proceedings Vol. 1 p D4-1, D4-11.
7. Vogel, G. J., et al., "Application of Pressurized, Fluidized-Bed
Combustion to Reduction of Atmospheric Pollution." Institute of
Fuel Symposium Series No. 1 (1975) Proceedings Vol. 1 p D3-1,
D3-11.
8. "Pressurized Fluidised Bed Combustion" R&D Report No. 85, Interim
No. 1 Report to OCR by National Research and Development Corp.
September, 1973.
9. Ehrlich, S., "A Coal Fired Fluidized-Bed Boiler" Institute of
Fuel Symposium Series No. 1 (1975) Proceedings Vol. 1 p C4-1,
C4-10.
10. Keairns, D. L., et al., "Fluidized Bed Combustion Process
Evaluation" Phase II - Pressurized Fluidized Bed Coal Combustion
Development. EPA Report No. EPA-650/2-75-027-C, Westinghouse
Research Laboratories, September 1975.
11. Kunii, D., and Levenspiel, 0., Fluidization Engineering, John
Wiley and Sons, Inc., New York (1969).
140
-------
SECTION VIII
LIST OF PUBLICATIONS
1. Hoke, R. C., Ruth, L. A., Shaw, H., "Combustion and Desulfuriza-
tion of Coal in a Fluidized Bed of Limestone," Presented at the
IEEE-ASME Joint Power Generation Conference, Miami Beach, FL,
September 15-19, 1974.
2. Hoke, R. C., Ruth, L. A., Shaw, H., Combustion 46, No. 7,
pp 6-12, January 1975.
3. Hoke, R. C., Workshop on Regeneration of Sulfated Limestone/
Dolomite for Fluidized Bed Combustion, ERDA, Fossil Energy.
Washington, B.C., March 3-4, 1975.
4. Nutkis, M. S., "Pressurized Fluidized Bed Coal Combustion,"
Presented at the International Fluidization Conference, Pacific
Grove, CA, June 15-20, 1975.
5. Ruth, L. A., "Combustion and Desulfurization of Coal in a
Fluidized Bed of Limestone," Presented at the International
Fluidization Conference, Pacific Grove, CA, June 15-20, 1975.
6. Hoke, R. C., Bertrand, R. R., "Pressurized Fluidized Bed
Combustion of Coal," Institute of Fuel Symposium Series No. 1
(1975) Proceedings Vol. 1. September, 1975.
7. Hoke, R. C., Bertrand, R. R., "Combustione Di Carbone Su Letto
Fluido Sotto Pressione," Associazone Termotecniea Italiana,
Fluidized Bed Combustion Symposium, Cagliari, Sardinia
September 27, 1975.
8. Nutkis, M. S., Bertrand, R. R., "Operation and Performance of
the Pressurized FBC Miniplant," Presented at the Fourth
International Conference on Fluidized Bed Combustion, McLean, VA,
December 9-11, 1975.
9. Ruth, L. A., "Regeneration of CaSO^ in FBC," Presented at the
Fourth International Conference on Fluidized Bed Combustion,
McLean, VA, December 9-11, 1975.
10. Hoke, R. C., "Emissions from Pressurized Fluidized Bed Coal
Combustion," Presented at the Fourth International Conference on
Fluidized Bed Combustion, McLean, VA, December 9-11, 1975.
141
-------
PATENT MEMORANDA SUBMITTED
1. Ruth, L. A., Hoke, R. C., A Calcining Zone to Permit Use of
Limestone in Pressurized Fluidized Bed Combustion.
2. Ruth, L. A., Hoke, R. C., Steam Coils for a Fluidized Bed
Coal Combustor.
3. Siminski, V. J., Positive Blow-Back Granular Bed Filter.
142
-------
SECTION IX
APPENDICES
Page
A. MATERIALS OF CONSTRUCTION 144
B. REPORTS OF METALLURGICAL EXAMINATIONS 152
C. SAFETY CONSEQUENCES OF A STEAM-COIL BREAK.
STEAM-CARBON EQUILIBRIUM 162
D. MINIPLANT ALARM ANNUNCIATORS 164
E. MINIPLANT DATA LOGGER POINTS 166
F. ANALYTICAL TECHNIQUES 169
G. MINIPLANT RUN SUMMARIES 170
H. BATCH UNIT RUN SUMMARIES 198
143
-------
APPENDIX A
MATERIALS OF CONSTRUCTION
Presented here is a brief review of materials which are candidates for
use in future generations of high-temperature, fluldized bed, coal
combustion-limestone regeneration units. Two classes of materials are
discussed: refractories (ceramics) and metals. The approach will be
to outline selection criteria while emphasizing underlying basic princi-
ples. Some examples of specific materials will be given but there has
been no attempt to present a broad review of the published literature
or manufacturers' brochures. Results will be given of our experience
with several materials.
A- Refractories
1. Function of Refractories
A fluidized bed combustor or regenerator would normally be constructed
with a steel outer shell, lined with a refractory material. The
refractory limits heat losses and also protects the steel shell from
excessive temperatures and corrosive materials. Temperatures inside
the combustor are typically about 1600°F, and in the regenerator,
2000°F. The refractory should limit the temperature at the steel shell
to 240-400°F. At this temperature the shell retains most of its room
temperature strength and is still hot enough to prevent condensation of
water vapor on its surface, thereby reducing corrosive attack. The
shell is protected against hot corrosive gases because gas is cooled
as it migrates through the refractory.
2. Classes of Refractories
The primary components of most refractories are aluminum oxide (alumina)
and silicon dioxide (silica). The physical forms most commonly encoun-
tered are brick, castables, plastic-refractories, and ceramic fiber
materials. Castable refractories are mixtures of raw and calcined
clays of carefully chosen size and particle size distribution. They
also contain compounds which hydrate on addition of water, usualy cal-
cium aluminate. Plastic refractories are mixtures of ground fireclays
with chemical binding agents and are similar to castables except that
they are shipped wet from the manufacturer and are usually applied by
ramming with pneumatic hammers. They can be formulated to harden on
exposure to air or when heated. Refractory brick consists of a mixture
of clays which have been fired to a high temperature to drive off
volatiles and form strong ceramic bonds between component particles.
Bricks, castables, and plastics of similar composition have almost
identical properties after firing at high temperatures. Castable
refractories, as a class, are particularly well suited to fluidized
144
-------
bed reactors because they may be used for any shape where installation
of forms is possible. Monolithic materials also tend to be slightly
less permeable than their brick analogs. The fourth class of refract-
ories, ceramic fiber board, have temperature limits which are too low
(<1800°F), limited strength, and relatively poor erosion resistance,
and will not be considered further.
3. Selection Criteria
Use Temperature
Use temperature is probably the single most important consideration in
selecting a refractory. Castables and plastic refractories can be
obtained with ratings as high as 3300°F for continuous service. High
alumina brick is available with ratings up to about 3500°F. In general,
maximum use temperature increases with alumina content.
It is not desirable to specify refractories, especially castables and
plastics, with maximum use temperatures far above anticipated operating
conditions. In the case of monolithic linings, ceramic bonds do not
form until the lining is thoroughly cured near maximum use temperature.
At low temperatures, which occur near the cool face of the lining,
strength is usually provided by hydraulic bonds. If a section of mono-
lithic material is heated sufficiently to destroy hydraulic bonds but
not enough to form ceramic bonds, the material is relatively weak. This
condition is often found near the center of a monolithic lining. The
presence of a relatively weak center makes the structure somewhat more
flexible than brick construction, and may allow it to better resist
thermal shock. However, if the hot face of the lining is never thor-
oughly cured by soaking at near maximum temperatures, the entire struc-
ture will be weak, and likely to spall. Even in brick, ceramic bonds
have not been completely formed during manufacture. Use of brick at
temperatures well below the rated temperature will minimize formation
of additional bonds and leave the brick relatively weak.
Resistance to Chemical Attack
Alumina-silica refractories are relatively inert although problems can
occur in some cases. Chemical attack usually results in refractory
failure of three kinds. In one kind, the presence of fluxing agents
can cause melting of the refractory surface. The rate of attack will
depend on how fast the melted material is removed and fresh refractory
surface exposed. Another kind of failure involves formation of solid
reaction products. This can occur if corrosive species diffuse to
cooler parts of the refractory and condense. Because the solid
reaction product will usually have a thermal expansion coefficient
markedly different from the bulk refractory, spalling can occur on
start-up or shut-down. A third kind of attack involves condensation
of water in the refractory during shut-down. Rapid heating on startup
can cause steam evolution and even explosion of the refractory.
145
-------
Accumulation of water in the refractory is accelerated by the presence
of hygroscopic products of high temperature attack.
Thermal Conductivity and Density
Thermal conductivities and density of brick and castable refractories
vary over a wide range. Generally, denser materials have higher thermal
conductivities. For example, a lightweight insulating castable could
have a density of 50-55 lb/ft3 and a thermal conductivity of about 0.16
Btu/hr ft°F at 1000°F. A dense castable (60% A1203) with a density of
about 140 lb/ft3 could have a thermal conductivity of about 0.43 at
1000°F. Similarly, densities and thermal conductivities of brick vary
from about 45 lb/ft3 for insulating firebrick (K = 0.2 at 1000°F) to
about 145 Ib/ft3 (K = 0.78 at 1000°F) for super-duty brick.
Dense refractories do have a decided advantage, however, over light-
weight materials: they have much better resistance to chemical attack
and usually more mechanical strenght.
Thermal Shock Resistance
Refractories are brittle and can crack when subjected to sharp tem-
perature fluctuations; hence, care is necessary when heating and cooling
refractory-lined vessels. Thermal shock resistance increases with
increasing thermal conductivity and decreasing density. Mullite (3A1203-
2SiC>2) is the best alumina-silica material for applications requiring
good thermal shock resistance.
Cost
The total cost of refractory includes the cost of materials, installa-
tion, and maintenance. Monolithic materials may cost more than the
equivalent brick for the same degree of insulation but this may not be
so for brick shapes other than simple configurations. Brick linings
are almost always installed in the field and skilled workmen are
required. Castable linings are much simpler to install and withstand
shipment rather well. Maintenance costs of various lining types
depend strongly on the type of operation, amount of supervision and
accessibility for repair.
4. Experience With Fluidized
Bed Combustion - Regeneration
Refractory materials which have been used in the batch and miniplant
fluidized bed combustor and regenerators are listed with their applica-
tion below:
• Grefco Litecast 75-28: refractory lining for combustor and
regenerator vessels.
• Grefco Bubbalite: refractory lining for pre-heat burner sections.
146
-------
• Aremco Ceramacast 511: supports for combustor cooling coils.
• Cast alumina and zirconia: batch unit regenerator fluidizing grids.
• Alumina and mullite tubes: liner in batch regenerator.
Performance of the cast ceramic fluidizing grids and the mullite and
alumina liners have already been discussed in a recent report (EPA-
650/2-74-001) and will not be repeated here.
Grefco Litecast 75-28 is the refractory lining for the batch equipment
and miniplant (excluding pre-heat burner sections). This material has
a service limit of 2800°F, a bulk density of 75 lb/ft3, and a thermal
conductivity of 0.32 Btu/hr ft °F at 1000°F. The manufacturer claims
that Litecast has excellent strength and good resistance to thermal
shock. This material has held up very well with use.
Grefco Bubbalite is a high temperature lightweight castable with a fused
bubble alumina aggregate and high-purity low-iron calcium aluminate
binder. It has a service limit of 3300°F and a thermal conductivity
at 1000°F of 0.5 Btu/hr ft °F. It was selected for use in the burner
sections of the combustors and regenerators because of its high tem-
perature rating. It has held up very well considering the high prob-
ability of flame impingement and rapid heating and cooling.
The combustor cooling coils were supported at flanges with Aremco
Ceramacast 511. This is a magnesium oxide/zircon refractory with a
temperature limit of 2800°F. No problems have occurred with this
material in this application.
5. Manufacturers of Refractories
In addition to General Refractories-Grefco and Aremco Products
(specialties)(Briarcliff Manor, New York) other manufacturers of
castable and brick refractories include Babcock and Wilcox Company,
Refractories Division (Augusta, Georgia), J. H. France Refractories
Company (Snow Shoe, Pennsylvania), and Harbison-Walker Refractories Co.
(Pittsburgh, Pennsylvania). Each produces a large line of competitive
refractories.
B. Metals
Successful application of metals in high-temperature process service
requires consideration of mechanical and metallurgical properties and
corrosion and oxidation resistance. These factors will be discussed
and specific references to high temperature alloys and corrosive atmos-
pheres will be included. Experience with materials used in the fluid-
ized bed combustors and regenerators will also be given.
147
-------
1. Mechanical and Metallurgical
Considerations at High Temperatures
At moderate temperatures, below about 700°F for steels, ordinary yield
strength is the mechanical property that limits the application of
engineering alloys. However, at very high temperatures, creep, rupture,
and short-time strengths are important. Creep is the gradual straining
that metals experience under relatively low stresses at high temper-
atures. It is usually expressed as the stress required to produce a
given creep rate, e.g., 1% in 10,000 hrs. Rupture strength is the
stress required to produce rupture in a definite time at a given tem-
perature. Short-time strength is simply the ordinary tensile strength
at a high temperature. Among alloys, as a rule, ferritic materials
are weaker than austenitic compositions, and molybdenum increases
strength in both groups.
In addition to changes in mechanical properties, many alloys are subject
to changes in metallurgical properties at elevated temperature. Specific
comments on various materials are given below:
Carbon Steel and 1/2% Mo Steel
Spheroidization of carbides and graphitization generally occur in carbon
steel and 1/2% Mo steel after prolonged exposure to temperatures in
the 800-1300°F range. Significant reduction in strength accompanies
both spheroidization and graphitization.
Low Alloy Steels (1% Cr, 1/2% Mo to 5% Cr, 1/2% Mo)
If heated above the transformation temperature ( 1300°F) and subsequently
cooled rapidly, these steels will have reduced ductility at lower tem-
peratures.
High Chromium Stainless Steels
Steels containing 17% or more Cr are susceptible to "885°F embrittle-
ment." This aging phenomena produces extensive decreases in toughness
and ductility after long term exposure in the 700-900°F range.
Chromium-Nickel Stainless Steel
(Types 304, 321, 347, 316, 309, 310, 330)
When chromium-nickel stainless steel are subjected to temperatures of
850-1200°F, "sensitization" can occur, which involves carbide precipita-
tion at the grain boundaries. Chromium carbides will be precipitated
leaving areas close to the grain boundaries low in chromium and thus
less resistant to intergranular corrosion. This would be particularly
detrimental under conditions where water is also present, such as during
"downtime." Sensitization can be prevented by using very low carbon
contents (about 0.03%), but generally the customary 0.08 percent carbon
is stabilized by the addition of titanium or columbium.
148
-------
Exposure of these steels to temperatures of 1000 to 1600°F can result
in formation of "sigma phase." Above 1000°F, this may increase tensile
strength without much effect on ductility; however, room temperature
ductility can be markedly reduced.
2. Oxidation and Corrosion Resistance
In many process applications, strength and mechanical properties become
of secondary importance compared with resistance to oxidation and cor-
rosion. All common alloys form oxides when exposed to hot oxidizing
environments. Whether or not the alloy is resistant depends upon
whether the oxide is stable and forms a. protective film. Mild steel
is seldom used above 950°F because of excessive scaling rates. Adding
chromium to steel improves oxidation resistance; thus 4-6% Cr steel is
acceptable to 1150°F, 9-12% Cr steel to 1300-1400°F, 14-18% Cr steel to
1500°F, and 27% Cr to 2000°F. Silicon (0.75-2%) also improves oxida-
tion resistance when added to low chromium steels.
There is a multitude of possible corrosion reactions and it is sometimes
advantageous to classify them into a few broad types, including direct
corrosion and electrochemical corrosion. Although practically all cor-
rosion is electrochemical (anodic and cathodic regions on the metal sur-
face are involved), electrochemical corrosion usually refers to the type
of corrosion that takes place at or near room temperature as a result
of reaction of metals with water or with aqueous solutions of salts,
acids, or bases. Direct corrosion is essentially ordinary chemical
attack. The corrosive agent usually attacks the surface uniformly and
at an almost constant rate that can be conveniently measured, e.g.,
in inches penetration per year. Of course, more specific descriptions
are widely used for certain types of industrially important corrosion.
Some examples of non-uniform attack are pitting and intergranular cor-
rosion. Other specific types of corrosion include stress-corrosion,
corrosion fatigue, and erosion corrosion.
Problems of special chemical environments that can be encountered in
fluidized bed combustion are briefly discussed below.
Sulfur Dioxide and Hydrogen Sulfide
The relative effect of S02 in air plus water vapor on corrosion of
steel at 1650°F is given in Table A-l.
149
-------
TABLE A-l. EFFECT OF S02 AND H20 ON
CORROSION OF CARBON STEEL AT 1650°F
Atmosphere Relative Corrosion
Pure Air 100
Air + 2% S02 118
Air + 5% H20 135
Air + 5% S02 + 5% H20 276
Data from Tomashov, N. D., Theory of Corrosion and
Protection of Metals, The MacMillan Company, New
York, 1966, p. 114.
Sulfides usually form when an alloy contacts S or H2S at high tempera-
tures. Attack is generally more severe than during oxidation because
sulfides have higher molar volumes than oxides. High volume ratios can
lead to the build-up of stresses resulting in the cracking of scale and
exposure of fresh surface to further attack. Metal sulfides usually
melt at lower temperatures than do oxides and metal-metal sulfide
eutectics frequently have low melting points. The melting point of the
Ni-Ni3S2 eutectic is particularly low, and this seems to account for
the fact that alloys containing more than 15-30% Ni are exceptionally
sensitive to sulfur-containing gases under non-oxidizing conditions.
The liquid sulfide destroys the scale locally and opens up new centers
for attack by the corroding gas.
Alloys rich in nickel should not be used in contact with sulfur con-
taining gases, particularly in the absence of oxygen, and often even
in the presence .of oxygen. Even when excess oxygen is present, as in
the combustion of fuel, local reducing conditions can exist in regions
where combustion is incomplete.
Chromium is the most important material in imparting resistance against
sulfide scales. For this reason, straight chromium stainless steels
are recommended in high H2S atmospheres. High chromium steels (>20%)
are also quite resistant to S02 atmospheres. Corrosion in moist S02
is stronger than in the dry gas at both high and low temperatures (See
Table A-l). It is interesting and important that alloys rich in nickel
are relatively resistant only to dry S02 while they are heavily attacked
in moist S02 at low and high temperatures. It is also interesting that
moist H2S attacks the various steels at about the same rate as dry H2S.
Sulfur Trioxide
Small amounts of 803 normally present in flue gas are strongly absorbed
in water droplets so that the corresponding acid condensates are quite
concentrated. With S03 present, the dewpoint may be as high as 340°F,
150
-------
and possibly higher. Severe corrosion can result to steels, stainless
steels, and many nickel-base alloys. Hence, serious problems can occur
when metal temperatures are below the dewpoint.
Compounds Containing Vanadium, Sodium,
Potassium, Sulfur, Molybdenum, or Lead
Compounds containing the above elements can accelerate the corrosion
of steels, stainless steels, and other alloys at elevated temperature.
While the concentration of these elements in coal may be low, deposits
on metal surfaces within a fluidized bed boiler may not be. Vanadium
and sodium, in particular, react to form a corrosive liquid (m.p. 980°F)
However, calcium oxide and some other compounds can combine with vana-
dium oxides and minimize corrosion.
Carbon Dioxide and Carbon Monoxide
Carbon dioxide and monoxide in flue gas probably does not accelerate
corrosion with iron based alloys in most cases. However, high C02 con-
centrations or sufficiently high temperatures can decarburize or
carburize steels and other alloys, depending on the temperature and
C0/C02 ratio.
151
-------
APPENDIX B
REPORTS OF METALLURGICAL EXAMINATIONS
152
-------
RESEARCH AND ENGINEERING COMPANY
PO BOX 101 FLORHAM PARK NEW JERSEY 07932
EXXON ENGINEERING TECHNOLOGY DEPARTMENT
Mechanical and Materials Engineering Division
K. G. FEEHAN
Manager
R. D. MERRICK
Senior Engineering Associate
59983
Cable ENGREXXON NY
July 23, 1974
Linden Fluidized Bed Coal
Combustion Miniplant Analysis of
Cooling Coil Samples
Mr. Melvyn S. Nutkis
Government Research Laboratory
Exxon Research and Engineering Co.
P.O. Box 8
Linden, N.J. 07036
Dear Mr. Nutkis:
We have completed our examination of the cooling coil samples
from your Fluidized Bed Coal Combustion Miniplant. The samples accompanied
your letter of June 17, 1974. We found no evidence of corrosion or deteri-
oration of the coil. Apparently the cooling water maintained the coil at
a sufficiently low temperature so that neither sensitization nor attack from
flue gas occurred. However, the rod, which was not cooled, suffered both
sensitization and corrosive attack. We recommend that a thicker 316 stain-
less steel rod or a rod made of 310 stainless steel be used in this applica-
tion in the future.
The samples are from a damaged cooling coil removed from the
combustor after approximately 60 hours of exposure. Three samples were
received for metallurgical examination; sample #1 was a tube section near
the water inlet, sample #2 was a tube section near the water outlet, and
sample #3 was a length of stiffening rod which was tack welded to the coil.
The coil was fabricated from 3/4" OD type 316 and 316L stainless steel
tubing. The stiffening rod is 3/16" type 316 stainless steel rod.
In operation, the tubing was water cooled and did not experience
the reaction temperature of 1500-1850°F. The rod, however, was not cooled
and was exposed to the high temperature.
The microstructure of the rod (Figure 1) shows carbide preci-
pitation in the grain boundaries, which is typical of sensitized stainless
steel. Type 316 stainless steel is susceptible to sensitization in the tem-
perature range of 950 to 1450°F. Above 1450°F sensitization is not a problem.
However, stainless steels operated above this temperature may become
sensitized when cooled through the range. Thus, the rod apparently became
sensitized during cooling and not during the operation of the vessel.
Figure 1 also shows a layer of corroded metal at the outer surface of the
rod. The layer is approximately one mil thick and is probably due to high
temperature attack from flue gases containing S0£ and 803.
153
-------
- 2 -
The coil (Figure 2) appears to be less affected than the rod.
There is no heavy carbide precipitation at the grain boundaries. We feel
that the cooling water maintained the coil at a temperature below the sen-
sitization range, and that the microstructure of the coil is "normal". In
addition, there is no evidence of metal depletion on either the outside or
inside surfaces of the coil.
There was a layer of residue on both the outside and inside sur-
faces of the coil. A qualitative analysis of the outside layer indicated-
that the major constituents were calcium, sulfur, and silicon. Small
atomic weight atoms such as oxygen cannot be determined by this method.
The outside layer is probably a residue of the combustor reaction. The
inside layer analysis showed copper, sulfur and small amounts of aluminum
present. This residue was probably formed from particles in the cooling
water. We do not feel that this residue has had a detrimental effect on
the tubing.
In general, the coil was unaffected by the service; however,
since the system operated for only 60 hours it is possible that a signifi-
cant corrosion rate could go undetected. We know that oxidizing agents
such as SC>2 will corrode type 316 stainless steel above 1600°F. It appears,
however, that the cooling water has maintained the coils at a safe tempera-
ture. We feel that type 316 stainless steel can provide good service in
this application.
The rods, where not cooled, were subjected to the operating tem-
peratures of 1500-1850°F. Type 316 stainless steel is not recommended for
use above 1500°F. We feel, however, that 316 stainless steel can be used
in this application since the rods are not a critical component of the
system and are not subjected to large stresses. However, if 316 stainless
steel is specified, we would recommend that a thicker rod (at least 1/4"
in diameter) be used. Alternatively, a more corrosion resistant material
such as 310 stainless steel could be used.
Very truly yours,
/J.W. SLUSSER
JWS:km
Attach.
cc: K. Macnamara ( 000 )
154
-------
- 3 -
r • • ' f • ,-''^Y'vA'^%
' i ' * »"J ^ ' / \ ^^ "" ' ' * "^_-l- _>**X*^ " ' /
^fi|Sfei||i?
fc -Ar/. \ 'JOi-vV-., r^,-' -\ v"^CS*v/
. . •., . ._
:
Figure 1
Micrograph of the rod (sample #3) near the outer surface.
Mt //7903 Etch: oxalic acid
500X
r'»«*
' (\
* '
.. • /: \
\. ~.
Figure 2
Micrograph of the coil (sample #1) from the Fluidized Bed Coal Combustion
Miniplant - Linden
Mt #7901 Etch: oxalic acid 500X
155
-------
RESEARCH AND ENGINEERING COMPANY
P.O. BOX 101. FLORHAM PARK. NEW JERSEY 07932
EXXON ENGINEERING TECHNOLOGY DEPARTMENT
Mechanical and Materials Engineering Division
K. G. FEEHAN
Manager
J.E. Guthrie
Section Head
Cable: ENGREXXON. N.Y.
54882 June 30, 1975
Analysis of Cooling Coil Samples
- Linden Coal Combustion Miniplant
Mr. D.D. Kinzler
Government Research Laboratory
Exxon Research & Engineering Company
P.O. Box 8
Linden, N.J. 07036
Dear Mr. Kinzler:
Our analysis of the tubing from your coal combustion miniplant
indicates high temperature gas attack (probably oxidation/sulfidation)
has caused most of the corrosion. We also found some evidence of high tem-
perature slag attack most likely caused by the combination of Na, V, and S
in the coal combustion products. The worst corrosion and metal loss occurred
on the tube elbows with only minor pitting occurring on the straight sections.
This was due to continuous erosion of the corrosion product formed at the
elbows and the constant exposure of fresh surface. We believe the corrosion
can be reduced by insuring cooling water flow during operation and, thus,
lower tube metal temperature (TMT). If this is not practical, Type 310SS
should be used instead of Type 316SS. Type 310SS has better high temperature
corrosion resistance than Type 316SS.
Samples from cooling coils 2A, 4A, IB, and 2B were received for
examination. The tubes are 3/4" OD, 0.048" thick Type 316SS (straight
sections) and Type 316L SS (elbows) and were removed after 60 hours of
exposure in the combustor. The environment inthe combustor contains C02,
S02, and 02 gases at 1500 to 1850°F. Type 316 and 316L SS have similar high
temperature corrosion resistance in this type of environment and are usually
acceptable to about 1600°F TMT. Tubeside environment is cooling water which
is converted to saturated steam during operation.
We previously examined tubing from this plant (July 23, 1974, letter
No. 59938) and found it unattacked. One of the changes made since then was
to start the reaction without cooling water in the coils. Previously, cooling
water was turned on at the startup of the reaction; now it is turned on after
the reaction has started, and as a consequence, it is believed that the TMT
has exceeded 1600°F.
Table 1 gives wall thickness measurements taken randomly on the
tube samples and shows the attack was localized with considerable metal loss
156
-------
- 2 -
occurring on the outside of the elbows. The inside of the elbows and the
straight sections suffered little attack. We also found that there was no
scale on the OD surface where the highest metal loss had occurred. This
localized metal loss indicates erosive removal of scale.
Coil 2A in particular showed this localized metal loss at elbows.
Figure 1 is a micrograph of an elbow at a section in which the wall thickness
has been reduced to 16 mils. Notice the uniform metal loss at the OD surface.
A combination of high temperature gas corrosion and erosion of the resulting
scale by the bed particles would produce this type of metal removal. The
bed particles are quite hard, but the low velocity of the particles (6 fps)
discounts the possibility of erosion of the stainless steel. The particles,
though, may abrade the scale which forms during corrosion. When this scale
is removed, it exposes fresh metal to the environment and results in uniform
metal loss.
Examination of the coil 4A elbows again showed high metal loss at
the outside of the elbows; however, in this case, pitting attack was present.
Microprobe analysis showed sulfur was present in the pits. Figure 2, a micro-
graph of a straight section adjacent to an elbow, shows the undercutting type
of pitting typical of sulfidation/oxidation gas attack. The attack in this
case is similar but less severe than that found on coil 2A. The difference
may be due to an oxide scale on the surface of 4A which hindered attack.
However, this scale was permeable, and extensive pitting occurred beneath
the scale.
Pitting attack, although less severe than on the 4A elbows, also
occurred on a straight section of coil 2B as shown in Figure 3. This pitting
does not have the undercut appearance shown in Figure 2, and we believe it
was due to slag attack. At high temperatures, elements in the coal such as
Na, V, and S can combine on the tube surface to form liquid slags which are
corrosive.
The scale on the coils indicates the coils experienced temperatures
above 1600°F. Little or no scaling would be expected in only 60 hours of
exposure if the TMT had been maintained below 1600°F; however, above this
temperature, Type 316 and 316L SS scale rapidly. The microstructure of the
coils, especially the large grains and rounded grain intersections of Figure 2,
also reflect high temperature service.
The structure shown in Figure 1, while quite different from that
in Figure 2, is also typical of high temperature service. The difference
is that the structure shown in Figure 2 is an annealed Type 316 SS while that
shown in Figure 1 is a recrystallized Type 316L SS, typical of a metal which
is cold worked and then heated. The temperature required for recrystallization
is a function of the metal, the amount of cold work, and the time at tempera-
ture. Type 316L SS will not recrystallize below 1600°F. In this case, the
cold work resulted from the bending of the elbow, and the recrystallization
occurred during operation. We cannot determine the exact temperature reached,
but, regardless of the degree of cold work, recrystallization would not occur
below 1600°F. Also, the thin grain boundaries show this structure is not
sensitized because the metal is Type 316L SS which, unlike Type 316 SS, will
not sensitize at any temperature in only 60 hours.
157
-------
- 3 -
The structure shown in Figure 3 is that of a lightly cold worked
structure which has not recrystallized. The cold work is indicated by the
twinning shown in Figure 3 (note the absence of twinning in Figure 1) . Since
the metal is only lightly cold worked, the recrystallization temperature
would be well above 1600°F.
The scale found on the OD surface of all the tubes and the micro-
structure of the tubes indicate the coils have experienced temperatures above
1600°F. This high temperature has resulted in oxidation/sulfidation attack
which has been aggravated by erosion of the corrosion scale. We also found
indications of a slag attack. Lowering the TMT by insuring cooling water
flow will result in reduced attack. If continuous cooling water flow is
impractical, Type 310 SS should be used instead of Type 316 and 316L SS.
Type 310 SS has better high temperature corrosion resistance and is accep-
table to 1900°F.
Very truly yours ,
JOSEPH W. SLUSSER
JWS:km
Attach.
cc: K. Macnamara (5150)
158
-------
OD
[D
Figure 1
Micrograph of Coil 2A elbow (Type 316L S3) showing uniform metal loss. Sample
was over etched to show recrystallized structure. Note large grains indica-
tive of high temperature service.
Mt. 8412 200X
Grain Size
TO
Figure 2
Micrograph of Coil 4A straight section adjacent to elbow (Type 316 SS)
showing pitting on OD surface. Note large grains and rounded grain
intersections typical of high temperature service.
Mt. 8379 100X
159
-------
OD
Twin
Twin
Figure 3
Micrograph of Coil 2B straight section (Type 316 SS) showing pitting from
slag attack. Structure has twins indicating light amount of cold work.
Mt. 8413
200X
160
-------
TABLE I
Thickness Measurements (inches) of Samples from
Coils 2A, 4A,
(Randomly
2B, and
taken)
IB
Elbows
Thickness
Sample Coil Outside of Bend
1 2A 0 (failure)
.017
2 2A .020
.021
.019
3 2A .022
.025
.016
4 4A .023
.025
.027
5 IB .032
.030
.031
Straight
Sample Coil Thickness
3 2A .048
.045
.046
Loss
.048
.031
.028
.027
.029
.026
.023
.032
.025
.023
.021
.016
.018
.017
Sections
6 2B .046 scale on ID
Thickness
Inside of Bend*
.050
.047
.039
.048
.053
.048
.052
.054
.040
.037
.050
.037
.045
.035
removed for
Loss
-
.001
.009
-
-
-
-
-
.008
.011
-
.011
.003
.013
Loss
-
.003
.002
.002
measurement
* The tube scale has resulted in some wall
thicknesses greater than the original thickness.
JWS:km
161
-------
APPENDIX C
SAFETY CONSEQUENCES OF A STEAM-COIL BREAK
STEAM-CARBON EQUILIBRIUM
The consequences of a rupture in an internal heat transfer coil in a
fluidized bed coal combustor were examined because of the possibility of
hydrogen formation via the carbon-steam reaction should such an event
occur.
The equilibrium constants for the reactions of carbon with oxygen
(IQlO-lO20 at 1000K) are much higher than those for the reactions of
carbon with steam (about 1-4 at 1000 K) . Therefore, in the presence of
oxygen and at the temperatures involved in fluid bed combustion, the
amount of steam decomposition will be insignificant. The steam-carbon
reactions can only occur where the oxygen partial pressure is very small
as, for example, might occur if the fluidizing air were shut down after
steam was introduced. Hence, the system is fail-safe and potentially
dangerous hydrogen-oxygen mixtures cannot form. The recommended action
in the event of a break in a steam coil is to (1) shut down coal flow,
(2) leave fluidizing air ON, (3) shut down water supply to affected coil.
To support these conclusions, a computer program was run to determine
both adiabatic reaction temperature and equilibrium composition. It was
assumed that the miniplant was operating at maximum conditions when one
heat transfer coil ruptured. The material inputs were:
carbon - 480 Ib/hr. entering at 125°F (equiv- to about
630 Ib/hr of coal)
air - 1200 SCFM entering at 125°F
steam - 353 SCFM entering at 1000°F (equiv. to about
2 gal/min. liquid water)
pressure - 10 atm
These inputs represent nearly stoichiometric proportions and the tem-
perature produced will be a maximum. The computer results are given
below:
162
-------
adiabatic reaction temperature - 3200°F
gas composition -
N9 61.0 mole percent
H20 22.4
C02 16.0
CO 0.3
02 0.1
H2 0.09
NO, DH, H, 0 balance
It should be noted that the maximum temperature reached during a real
accident would be considerably lower than the adiabatic reaction tem-
perature. Since equilibrium for hydrogen production is less favorable
at lower temperatures, the concentration of hydrogen would also be less.
In the absence of air, steam and carbon could react but the reaction
temperature would drop rapidly due to the high endothermicity of the
steam-carbon reactions. For the same input of materials as before, but
without any air, equilibrium conditions are:
adiabatic reaction temperature - 680°F
gas composition -
H«0 54.6 mole percent
C02 22.1
CH. 21.0
4
H2 2.2
CO <0.1
The high methane concentration is the result of favorable equilibrium
for methane production at 680°F; however, reaction kinetics are poor at
low temperatures so that little methane should actually form. Moreover,
for the same reason, only a small amount of steam decomposition would be
expected to occur.
163
-------
APPENDIX D
MINIPLANT ALARM ANNUNCIATORS
Identification
Action
Remarks
Low U.V. Level Inside Plenum
Upon Initiation of Gas Feed
for Preheating
Low U.V. Level Inside Plenum
During Gas Feed for Preheating
Low Pressure Difference
Between Primary Injector and
Combustor
Low Flow in Primary Injector
Feed Line
High Coal Feed Line or Feed
Nozzle Temperature
Alarm; Gas Feed Stops if
No Ignition Occurs in 10
Seconds; Delay of 30
Seconds Before Ignition
Can Be Attempted
Alarm; Gas Feed Stops;
Delay of 30 Seconds
Before Reignition Can Be
Attempted
Alarm; Block Valve on
Coal Feed Probe Closes
Alarm
Alarm; Block Valve on
Coal Feed Nozzle Closes
Prevents Accumulation of
Uncombusted Natural Gas in
Combustor
Prevents Accumulation of
Uncombusted Natural Gas
in Combustor
Prevents Entry of Combustion
Gases into Primary Injector
Relative Flow Level Monitored
by Measuring Sound Intensity
in Feed Line with Scarpa Meter
Prevents Entry of Combustor
Gas and Solids into Atmosphere
if Feed Line Breaks
Low Pressure from Main Air Alarm
Compressor, Auxiliary Air
Compressor, or Site Air
Low Combustor Air Flow Shutdown
Prevents High Bed Temperature
Due to Poor Bed Mixing
High Exterior Surface
Combustor Temperature
High Burner Grid or
Fluidizing Grid
Temperature
High Fluidizing Grid
Pressure Drop
High Combustor Pressure
High Bed Temperature
Low Bed Temperature (for
Coal Ignition)
Alarm
Alarm
Alarm
Shutdown
Alarm @ 980°C;
Shutdown @ 1090°C
Alarm; Primary Injector
Feed Valve Closes
Indicates Blockage of Grid
Holes
Prevents Temperatures Causing
Bed Agglomeration
Alarm Setting is 650°C
164
-------
MINIPLANT ALARM ANNUNCIATORS (CONTINUED)
Identification
Low Bed Temperature (for
Liquid Fuel Combustion)
Action
Remarks
Alarm; Solenoid Valve on Alarm Setting is 425°C
Fuel Line Will Not
Activate
High Water Temperature From
Any Cooling Coil
Low Water Flow From Any
Cooling Coil
Low Pressure in Cooling Water
Tower
Alarm; Shutdown After 4 Prevents Cooling Coil Damage
Minutes if Uncorrected
Alarm; Shutdown After 4 Prevents Cooling Coil Damage
Minutes if Uncorrected
Alarm
High or Low Level in Cooling
Water Tower
Alarm
Low Pressure Drop Across Bed Alarm
Low Temperature in Dipleg Alarm
of Either Cyclone
High Temperature in Rejected Alarm
Solids Lock Hopper
High Downstream Flue Gas Alarm
Temperature
High or Low Pressure Shutdown
Difference from Expansion
Joint Compensation
Indicates Bed Level Too Low to
Reject Solids
Indicates Blockage of Solids
Flow Down Dipleg
Prevents Flue Gas Scrubber
Damage
Prevents System Damage Due to
Vertical Thermal Expansion of
Combustor
165
-------
APPENDIX E
MINIPLANT DATA LOGGER POINTS
Point No. Function
1 Constant 40 millivolt standard signal
COMBUSTOR
2 Weight of coal in primary injector
3 Cooling water flow, Coil 1A
4 IB
5 2A
6 2B
7 3A
8 3B
9 4A (future)
10 4B (future)
11 Air flow
12 Pressure
13 AP, coal feed vessel
14 AP, grid
15 AP, Bed Ports 4 to 31
16 AP, Bed Ports 4 to 11 or 15
17 Cooling water heat transfer loop
18 Spare
REGENERATOR
19 Burner air flow
20 Burner fuel flow
21 Supplemental air flow
22 Supplemental fuel flow
23 Pressure, AP to combustor
24 AP, grid
25 AP, bed Ports 29 to 34
26 AP, bed Ports 29 to 31
COMBUSTOR
27 Ratio coal to limestone
28 Air pressure to measuring orifice
29 Run identification no.
30 Burner grid metal temp.
^31 Cooling water temp., burner grid
32 Fluidizing grid metal temp.
33 Cooling water temp., fluidizing grid
34 Port #3, burner zone temp.
166
-------
Point No. Function
35 Temp, port #5, 6"
36 7, 18"
37 8, 27"
38 9, 45"
39 12, 63"
40 13, 81"
41 14, 87"
42 17, 117"
43 20, 135"
44 22, 162"
45 26, 216"
46 32, 341"
47 Coal injection probe temp., (spare)
48 Coal injection line temp., (spare)
49 Cooling water temp, to all coils
50 Cooling water temp, from Coil # la
51 IB
52 2A
53 2B
54 3A
55 3B
56 4A
57 4B
58 5A (spare)
59 5B (spare)
60 Cyclone gas discharge temp., 1st stage
61 Cyclone gas discharge temp., 2nd stage
62 Off gas temp, upstream of nozzle
63 Off gas temp, downstream of nozzle
64 Cyclone dipleg temp., 1st stage (mid)
65 2nd stage (btm)
66 Solids reject line temp, before pulse pot
67 Lockhopper temp., solids reject from bed (spare)
68 Lockhopper temp., flyash from cyclone #2
69 Surface temp. "A", lower deck, east side
70 Surface temp. "B", 1st deck, west side (spare)
71 Spare
72 Air temp, to measuring orifice
73 Ambient air
74 Fuel temp, to measuring orifice
REGENERATOR
75 Spare
76 Spare
77 Temp, port #3, 11"
78 5, 23"
79 7, 35"
167
-------
Point No. Function
80 Temp, port #8, 41"
81 9, 48"
82 10, 56"
83 12, 74"
84 14, 92"
85 21, 158"
86 26, 218"
87 Cyclone gas discharge
88 Spare
89 Spare
COMBUSTOR
90 Analyzer for SQ2
91 NO
92 02
93 C02
94 CO
REGENERATOR
95 Analyzer for S02
96 02
97 C02
98 CO
99 Clock
168
-------
APPENDIX F
ANALYTICAL TECHNIQUES
Analysis of Solids
Solids from combustion runs were analyzed for SO- ^-, C0o~^, Ca+2, Mg+2
carbon and total sulfur. The analytical techniques that were used are
described below.
-2
SO^ - The sample was treated with acidic BaCl2 solution.
The BaS04 precipitate was weighed.
_2
CO., - HC1 was added to an acidified sample. The solution
was stripped with N2 and the gas passed through
drierite, CuS04 and ascarite. C0-j~2 was determined
from the weight gain of the ascarite.
+2
Ca - The sample was digested by heating vigorously in a
Mg medium of perchloric acid/nitric acid. The
determination of Ca and Mg was made by atomic absorption.
Total sulfur - The sample was mixed with sodium peroxide and a catalyst.
The sulfur was converted to the sodium sulfate. The
sample was heated above the melting point and the melt
was extracted with water. The sulfur was converted to
barium sulfate, precipitated and weighed.
Carbon - Samples were combusted within a packed tube in an oxygen
atmosphere. Helium was used to sweep the combustion
gases into the analytical system. Carbon dioxide was
determined by differences in thermal conductivity.
Analysis of Flue Gas by Wet Chemical Methods
SO., - The amount absorbed by an 80% isopropanol solution was
determined titrimetrically using 0.01N barium perchlorate
as the titrant and thorin as the indicator.
S0» - The amount absorbed by a 3% hydrogen peroxide solution
was determined titrimetrically using 0.01N sodium
hydroxide as the titrant and methyl orange as the
indicator.
169
-------
APPENDIX G
MINIPLANT RUN SUMMARIES
Page
TABLE Gl. SUMMARY OF MINIPLANT RUN CONDITIONS 171
TABLE G2. MINIPLANT M> EMISSION DATA 187
X
TABLE G3. MINIPLANT SOLIDS ANALYSES 189
TABLE G4. MINIPLANT CALCIUM AND SULFUR BALANCES 197
170
-------
TABLE Gl. SUMMARY OF MINIPLANT RUN CONDITIONS
(See explanation of footnotes on pages 185 and 186)
Q
Run No. 2.1 31
Date , b 4/11/74 5/3/74
Run Length (hrs) 3.5 5 5
Cooling Coil Orientation0 H H
Area of Coils, Total (m2) 5.80 5.80
Limestone Feed No No
Bed Rejectiond No No
Bed Depth, Settled (m)
Start of Run 0.76 0.92
End of Run
Coal/Limestone Feed Ratio
Ca/S Molar Feed Ratio
Combustion Efficiency, Entire Run (%)e
Conditions ,
Time Interval 1 9:24-11:34 11:30-13:40 13:50-16:00
Pressure (kPa) 395 395 395
Super. Velocity (m/sec)
Average8 2.46 2.30 2.50
S.D. of 10 Min. Avg's 0.08 0.04 0.09
Minimum 10 Min. Avg.* 2.35 2.23 2.41
Maximum 10 Min. Avg.1 2.56 2.35 2.59
Temperature, Bed (°C)
Average 861 840 939
S.D. of 10 Min. Avg's 31 15 46
Minimum 10 Min. Avg. 796 827 887
Maximum 10 Min. Avg. n 896 865 989
Temperature, Flue Gas (°C)
Average 295 291 326
S.D. of 10 Min. Avg's 7 14 6
Minimum 10 Min. Avg. 280 268 314
Maximum 10 Min. Avg. 303 308 333
Coal Feed Rate (kg/hr)
Average 64.3 69.5 85.3
S.D. of 10 Min. Avg's 3.6 2.7 5.6
Initial 30 Min. Avg.J 66.2 70.1 84.4
Final 30 Min. Avg.3 65.3 68.1 81.0
Excess Air (%)°
Average 27.5 11.8 -8.9
S.D. of 10 Min. Avg's 8.9 4.2 5.5
Initial 30 Min. Avg. 25-° 10-7 ~7'8
Final 30 Min. Avg. 23-° ^^ ~5-4
171
-------
TABLE Gl. SUMMARY OF MINIPLANT RUN CONDITIONS
(See explanation of footnotes on pages 185 and 186)
Run No.3 4.1 4.3 5.1
Date b 5/10/74 5/20/74 6/20/74
Run Length (hrs) 3.5 4 4
Cooling Coil Orientation0 H H H
Area of Coils, Total (m2) 5.80 5.80 5.22
Limestone Feed No No No
Bed Rejection*1 No No No
Bed Depth, Settled (m)
Start of Run 0.92.
End of Run
Coal/Limestone Feed Ratio
Ca/S Molar Feed Ratio
&
CombustionfEfficiency, Entire Run (%)
Conditions ,
Time Interval* 13:50-14:50 14:30-15:40 14:12-15:22
Pressure (kPa) 1033 922 912
Super. Velocity (ra/sec)
Average6 1.69 1.89 1.88
S.D. of 10 Min. Avg's 0.03 0.06 0.07
Minimum 10 Mia. Avg.i 1.65 1.80 1.80
Maximum 10 Mih. Avg.1 1.74 1.92 1.92
Temperature, Bed (°C)
Average 885 919 849
S.D. of 10 Min. Avg's 12 42 28
Minimum 10 Min. Avg. 868 857 802
Maximum 10 Min. Avg. 902 949 887
Temperature, Flue Gas (°C)n
Average 403 350 331
S.D. of 10 Min. Avg's 10 11 12
Minimum 10 Min. Avg. 397 334 361
Maximum 10 Min. Avg. 411 360 397
Coal Feed Rate (kg/hr)
Average 80.7 64.8 72.9
S.D. of 10 Min. Avg's 3.9 5^9 3^4
Initial 30 Min. Avg.J 79.5 ^\2 73^8
Final 30 Min. Avg.3 82.0 64.4 72*4
Excess Air (%)°
Average 78.9 115 98.9
S.D. of 10 Min. Avg's 8..1 19 12
Initial 30 Min. Avg. 80.4 116 95.9
Final 30 Min. Avg. 77.4 115 102
172
-------
TABLE Gl. SUMMARY OF MINIPLANT RUN CONDITIONS
(See explanation of footnotes on pages 185 and 186)
a
Run No. 5^2 g j_ 6 2
SateT ^ /t, ^ 6/24/74 7/17/74 7/22/74
Run Length (hrs) 4.5 fi 4>5
Cooling Coil Orientation H H H
Area of Coils, Total (m2) 5.22 4.62 4.62
Limestone Feed No ^0 ^0
Bed Rejectiond No No No
Bed Depth, Settled (m)
Start of Run 0.92
End of Run
Coal/Limestone Feed Ratio
Ca/S Molar Feed Ratio
Combustion Efficiency, Entire Run (%)
Conditions ,
Time Interval 1 12:10-12:50 11:20-15:40 11:52-13:32
Pressure (kPa)~ 896 902 912
Super. Velocity (m/sec)
Average8 , 1.98 1.96 1.91
S.D. of 10 Min. Avg's 0.07 0.17 0.04
Minimum 10 Min. Avg.1 1.89 1.77 1.86
Maximum 10 Min. Avg.1 2.07 2.01 1.98
Temperature, Bed (°C)
Average 889 949 938
S.D. of 10 Min. Avg's 39 29 21
Minimum 10 Min. Avg. 828 839 912
Maximum 10 Min. Avg. 944 980 979
Temperature, Flue Gas (°C)
Average 358 377 407
S.D. of 10 Min. Avg's 11 10 11
Minimum 10 Min. Avg. 342 353 396
Maximum 10 Min. Avg. 375 388 423
Coal Feed Rate (kg/hr)
Average 74.2 73.9 68.5
S.D. of 10 Min. Avg's 4.4 6.9 4.5
Initial 30 Min. Avg.J 74.3 80.8 72.5
Final 30 Min. Avg.J 77.6 69.9 65.1
Excess Air (%)°
Average 96.1 87.8 100
S.D. of 10 Min. Avg's 10 17 13
Initial 30 Min. Avg. 94.9 74.1 88.6
Final 30 Min. Avg. 88.2 97.1 109
173
-------
TABLE Gl. SUMMARY OF MINIPLANT RUN CONDITIONS
(See explanation of footnotes on pages 185 and 186)
Run No.3 7.1 7.2 7.3 P
Date , 7/25/75 8/1/74 8/20/74
Run Length (hrs) 4 72
Cooling Coil Orientation0 H H H
Area of Coils, Total (m2) 4.62 4.62 4.62
Limestone Feed No No' No
Bed Rejectiond No No No
Bed Depth, Settled (m)
Start of Run 1.53
End of Run
Coal/Limestone Feed Ratio
Ca/S Molar Feed Ratio
Combustion Efficiency, Entire Run (%)
Conditions
Time Interval , 14:19-15:49 11:10-15:00
Pressure (kPa) 912 912 698
Super. Velocity (m/sec) 2.1
Average8 . 1.84 1.89
S.D. of 10 Min. Avg's 0.11 0.08
Minimum 10 Min. Avg.1 1.65 1.71
Maximum 10 Min. Avg.1 1.95 2.04
Temperature, Bed (°C) 880
Average 876 884
S.D. of 10 Min. Avg's 65 35
Minimum 10 Min. Avg. 782 809
Maximum 10 Min. Avg. 953 944
Temperature, Flue Gas (°C)n
Average 440 444
S.D. of 10 Min. Avg's 31 15
Minimum 10 Min. Avg. 387 416
Maximum 10 Min. Avg. 476 476
Coal Feed Rate (kg/hr) 106
Average 114 105
S.D. of 10 Min. Avg's 9.1 11.3
Initial 30 Min. AvgJ 108 107
Final 30 Min. Avg.3 109 112
Excess Air (%)o
Average 23.2 36.7
S.D. of 10 Min. Avg's 10 16
Initial 30 Min. Avg. 29.9 36.0
Final 30 Min. Avg. 27.2 26.6
174
-------
TABLE Gl. SUMMARY OF MINIPLANT RUN CONDITIONS
(See explanation of footnotes on pages 185 and 186)
a
Run No. 7^4 g ,p „ .
Date , b 8/27/74 10/9/75 10/24/75
Run Length (hrs) 4.5 ^ 13 5
Cooling Coil Orientation0 H H H
Area of Coils, Total (m2) 4.62 4.62 4.62
Limestone Feed Yes Yes Yes
Bed Rejectiond No No No
Bed Depth, Settled (m)
Start of Run 0.58 0.61
End of Run 1.98 1.30 1.17
Coal/Limestone Feed Ratio 3.0 7.35 12.0
Ca/S Molar Feed Ratio 4.0 1.67 1.0
Combustion Efficiency, Entire Run (%)e 96 96.4
Conditions
Time Interval , 12:30-14:10 15:00 21:00 15:33-23:33
Pressure (kPa) 912 907 906 932
Super. Velocity (m/sec)m 1.83 1.77
Average8 , 1.73 1.76
S.D. of 10 Min. Avg's 0.06 0.03
Minimum 10 Miu. Avg.1 1.62 1.71
Maximum 10 Min. Avg.1 1.80 1.83
Temperature, Bed (°C) 908 877
Average 877 881
S.D. of 10 Min. Avg's 38 20
Minimum 10 Min. Avg. 807 846
Maximum 10 Min. Avg. 927 933
Temperature, Flue Gas (°C)n
Average 434 371
S.D. of 10 Min. Avg's 23 23
Minimum 10 Min. Avg. 407 343
Maximum 10 Min. Avg. 476 422
Coal Feed Rate (kg/hr) 75 112
Average 97.2 75.7
S.D. of 10 Min. Avg's 5.1 13.2
Initial 30 Min. AvgJ 93.5 52.5
Final 30 Min. Avg.3 97.8 96.9
Excess Air (%)° 67 15
Average 3*-3 *°'5
S.D. of 10 Min. Avg's 13 . 37
Initial 30 Min. Avg. 39'^
Final 30 Min. Avg. 32'2
175
-------
TABLE Gl. SUMMARY OF MINIPLANT RUN CONDITIONS
(See explanation of footnotes on pages 185 and 186)
. m
Run No.
Date ,
Run Length (hrs)
Cooling Coil Orientation
Area of Coils, Total (m2)
Limestone Feed
Bed Rejection*1
Bed Depth, Settled (m)
Start of Run
End of Run
Coal/Limestone Feed Ratio
Ca/S Molar Feed Ratio
Combustion Efficiency, Entire Run (%)'
Conditions
ir
Time Interval ,
Pressure (kPa)
Super. Velocity (m/sec)'
Average** ,
S.D. of 10 Min. Avg's
Minimum 10 Min. Avg.*
Maximum 10 Min. Avg.1
Temperature, Bed (°C)
Average
S.D. of 10 Min.
Minimum 10 Min.
Maximum 10 Min. ^
Temperature, Flue Gas (°C)n
Average
S.D. of 10 Min.
Minimum 10 Min.
Maximum 10 Min.
Coal Feed Rate (kg/hr)
Average
S.D. of 10 Min. Avg's
Initial 30 Min, Avg.J
Final 30 Min. Avg.J
Excess Air (%)
-------
TABLE Gl. SUMMARY OF MINIPLANT RUN CONDITIONS
(See explanation of footnotes on pages 185 and 186)
Run No.a na u>2 12>1
S T v n, ^ 12/17/74 12/23/74 1/8/75
Run Length (hrs) 9<5 1Q g 5
Cooling Coil Orientation H/V H/V H/V
Area of Coils, Total (m2) 2.75 2.75 2.75
Limestone Feed No No Yes
Bed Rejectiond No No No
Bed Depth, Settled (m)
Start of Run 0.61 0.84q 0.76
End of Run 0.48 0.61 1.27
Coal/Limestone Feed Ratio — — 10
Ca/S Molar Feed Ratio — 1,2
Combustion Efficiency, Entire Run (%) 96.9 96.6 97
Conditions
Time Interval 1 12:20-17:10 12:00-17:30 11:29-19:19
Pressure (kPa) 915 912 922
Super. Velocity (m/sec)
Average8 2.05 1.89 1.78
S.D. of 10 Min. Avg's 0.03 0.04 0.03
Minimum 10 Min. Avg.* 2..01 1.74 1.71
Maximum 10 Min. Avg.1 2.14 1.95 1.86
Temperature, Bed (°C)
Average 909 915 903
S.D. of 10 Min. Avg's 16 28 22
Minimum 10 Min. Avg. 876 820 849
Maximum 10 Min. Avg. 954 954 963
Temperature, Flue Gas (°C)n
Average 507 514 527
S.D. of 10 Min. Avg's 15 12 39
Minimum 10 Min. Avg. 475 468 482
Maximum 10 Min. Avg. 537 533 623
Coal Feed Rate (kg/hr)
Average 61.9 69.1 80.2
S.D. of 10 Min. Avg's 1.5 2.6 14.2
Initial 30 Min. Avg.J 61.8 66.7 67.1
Final 30 Min. Avg.3 62.0 69.1 110
Excess Air (%)°
Average 139 95.8 66.5
S.D. of 10 Min. Avg's 6.0 7.8 26
Initial 30 Min. Avg. i^1 104 93-5
Final 30 Min. Avg. 139 95-1 18-9
177
-------
TABLE Gl. SUMMARY OF MINIPLANT RUN CONDITIONS
(See explanation of footnotes on pages 185 and 186)
Run No.a 12.2 13.2
Date . 1/16/75 1/30/75
Run Length (hrs) 8 10
Cooling Coil Orientation0 H/V H/V
Area of Coils, Total (m2) 2.75 2.75
Limestone Feed Yes Yes
Bed Rejectiond No No
Bed Depth, Settled (m)
Start of Run i-27 0.89
End of Run i-80 1-60
Coal/Limestone Feed Ratio 10 10
Ca/S Molar Feed Ratio i-2 1'2
Combustion Efficiency, Entire Run (%)e 93 96-7
Conditions"
Time Interval , 14:22-20:22 10:20-19:30
Pressure (kPa) 912 912
m
Super. Velocity (m/sec)
Average8 , 1.83 1.86
S.D. of 10 Min. Avg's 0.04 0.03
Minimum 10 Min. Avg.* 1.77 1.77
Maximum 10 Min. Avg.1 1.92 1.92
Temperature, Bed (°C)
Average 896 910
S.D. of 10 Min. Avg's 11 13
Minimum 10 Min. Avg. 878 874
Maximum 10 Min. Avg. 923 949
Temperature, Flue Gas (°C)n
Average 647 607
S.D. of 10 Min. Avg's 70 51
Minimum 10 Min. Avg. 549 538
Maximum 10 Min. Avg. 764 719
Coal Feed Rate (kg/hr)
Average 107 91.9
S.D. of 10 Min. Avg's 16.7 12.2
Initial 30 Min. AvgJ 84.5 79.8
Final 30 Min. Avg.3 134 120
Excess Air (%)°
Average 27.3 48.7
S.D. of 10 Min. Avg's 18 18
Initial 30 Min. Avg. 54.6 69.4
Final 30 Min. Avg. 0 10.5
178
-------
TABLE Gl. SUMMARY OF MINIPLANT RUN CONDITIONS
(See explanation of footnotes on pages 185 and 186)
Run No.
Date ,
Run Length (hrs)
Cooling Coil Orientation0
Area of Coils, Total (m2)
Limestone Feed
Bed Rejection^
Bed Depth, Settled (m)
Start of Run
End of Run
Coal/Limestone Feed Ratio
Ca/S Molar Feed Ratio
Combustion Efficiency, Entire Run (%)
Conditions ,
.m
Time Interval -,
Pressure (kPa)
Super. Velocity (m/sec)'
Average" ,
S.D. of 10 Min. Avg's
Minimum 10 Mia. Avg.1
Maximum 10 Min. Avg.1
Temperature, Bed (°C)
Average
S.D. of 10 Min.
Minimum 10 Min.
Maximum 10 Min. „
Temperature, Flue Gas (°C)n
Average
S.D. of 10 Min.
Minimum 10 Min.
Maximum 10 Min. Avg.
Coal Feed Rate (kg/hr)
Average
S.D. of 10 Min. Avg's
Initial 30 Min. Avg.J
Final 30 Min. Avg.3
Excess Air (%)°
Average
S.D. of 10 Min. Avg's
Initial 30 Min. Avg.
Final 30 Min. Avg.
Avg's
Avg.
Avg.
Avg's
Avg.
16:40-21
902
1.84
0.02
1.80
1.86
909
13
882
934
589
18
556
642
70.3
2.4
67.9
72.9
86.8
7
95.8
79.6
14.1
2/27/75
24
V/H
1.98
Yes
No
0.84
1.33
10
1.2
96.5
:50 1:40-8:30
912
1.83
0.02
1.80
1.86
916
7
904
947
679
12
658
714
80.8
2.5
77.7
81.4
61.0
5
67.1
60.5
14.2
3/10/75
27.5
V/H
1.98
Intermittent
No
1.33
2.00
10
1.2
93.7
18:52-7:12
912
1.91
0.03
1.83
1.99
934
7
913
960
796
54
681
851
110
4.9
108
110
22.5
6
24.8
22.6
179
-------
TABLE Gl. SUMMARY OF MINIPLANT RUN CONDITIONS
(See explanation of footnotes on pages 185 and 186)
Run No.
Date ,
Run Length (hrs)
Cooling Coil Orientation
Area of Coils, Total (m")
Limestone Feed
Bed Rejection^
Bed Depth, Settled (m)
Start of Run
End of Run
Coal/Limestone Feed Ratio
Ca/S Molar Feed Ratio
Combustion Efficiency, Entire Run (%)
Conditions
rC
Time Interval -,
m
Pressure (kPa)
Super. Velocity (m/sec)
Average*' ,
S.D. of 10 Min. Avg's
Minimum 10 Min. Avg.^-
Maximum 10 Min. Avg.1
Temperature, Bed (°C)
Average
S.D. of 10 Min. Avg's
Minimum 10 Min. Avg.
Maximum 10 Min. Avg.
Temperature, Flue Gas (°C)n
Average
S.D. of 10 Min. Avg's
Minimum 10 Min. Avg.
Maximum 10 Min. Avg.
Coal Feed Rate (kg/hr)
Average
S.D. of 10 Min. Avg's
Initial 30 Min. Avg.J
Final 30 Min. Avg.J
Excess Air (%)°
Average
S.D. of 10 Min, Avg's
Initial 30 Min. Avg.
Final 30 Min. Avg.
15.1
3/20/75
24
V/H
1.98
Yes
Intermittent
1.04
1.93
10
1.2
96.5
14:10-19:50
907
1.92
0.03
1.86
1.98
910
6
889
924
710
46
641
794
96.8
15.4
77.0
119
44.2
20
78.9
14.2
4:00-9:40
902
2.09
0.03
2.04
2.17
902
7
891
927
857
13
827
877
121
8.7
125
123
24.0
8
20.4
22;5
15.2
3/26/75
11
V/H
1.98
Yes
Yes
1.92
1.80
10
1.2
96.8
1:10-8:20
912
1.83
0.01
1.80
1.86
916
8
902
948
678
13
653
714
80.3
3.1
74.0
80.1
62.2
6
76.2
63.3
180
-------
TABLE Gl. SUMMARY OF MINIPLANT RUN CONDITIONS
(See explanation of footnotes on pages 185 and 186)
Run No.
Date
Run Length (hrs)
Cooling Coil Orientation0
Area of Coils, Total (m2)
Limestone Feed
Bed Rejection^
Bed Depth, Settled (m)
Start of Run
End of Run
Coal/Limestone Feed Ratio
Ca/S Molar Feed Ratio
Combustion Efficiency, Entire Run (%)'
Conditions
Time Interval ^
Pressure (kPa)
Super. Velocity (m/sec)
Average*5
S.D. of 10 Min. Avg's
Minimum 10 Min. Avg.-j-
Maximum 10 Min. Avg.1
Temperature, Bed (°C)
Average
S.D. of 10 Min. Avg's
Minimum 10 Min. Avg.
Maximum 10 Min. Avg.
Temperature, Flue Gas (°C)n
Average
S.D. of 10 Min. Avg's
Minimum 10 Min. Avg.
Maximum 10 Min. Avg.
Coal Feed Rate (kg/hr)
Average
S.D. of 10 Min. Avg's
Initial 30 Min. Avg.3
Final 30 Min. Avg.J
Excess Air (%)°
Average
S.D. of 10 Min. Avg's
Initial 30 Min. Avg.
Final 30 Min. Avg.
15.3
3/31/75
16
V/H
1.98
Yes
Yes
1.80
2.29
10
1.2
91.9
10:40-16:00
921
1.87
0.06
1.76
2.00
889
13
863
914
761
13
732
786
108
4.6
106
107
27.1
5.
31.1
28.4
4:00-10:00
902
2.09
0.03
2.04
2.17
901
8
891
927
858
13
827
879
122
5.4
125
1.22
23.6
6
20.4
23.4
15.4
4/2/75
15
V/H
1.98
Yes
Yes
2.29
10
1.2
98.0
14:20-19:10
912
2.03
0.02
1.99
2.10
905
14
880
924
803
55
736
887
119
11.9
105
134
25.0
12
39.1
11.6
181
-------
TABLE Gl. SUMMARY OF MINIPLANT RUN CONDITIONS
(See explanation of footnotes on pages 185 and 186)
Run No.
Date b
Run Length (hrs)
Cooling Coil Orientation
Area of Coils, Total (m2)
Limestone Feed
Bed Rejection^
Bed Depth, Settled (m)
Start of Run
End of Run
Coal/Limestone Feed Ratio
Ca/S Molar Feed Ratio
Combustion Efficiency, Entire Run (%)
Conditions
m
Time Interval -,
Pressure (kPa)
Super. Velocity (m/sec)
Average8 ,
S.D. of 10 Min. Avg's
Minimum 10 Mia. Avg.1
Maximum 10 Min. Avg.1
Temperature, Bed (°C)
Average
S.D. of 10 Min. Avg's
Minimum 10 Min. Avg.
Maximum 10 Min. Avg.
Temperature, Flue Gas (°C)
Average
S.D. of 10 Min. Avg's
Minimum 10 Min. Avg.
Maximum 10 Min. Avg.
Coal Feed Rate (kg/hr)
Average
S.D. of 10 Min. Avg's
Initial 30 Min. Avg.J
Final 30 Min. Avg.3
Excess Air (%)°
Average
S.D. of 10 Min. Avg's
Initial 30 Min. Avg.
Final 30 Min. Avg.
16.1
5/5/75
16.5
V
2.20
Yes
No
0.78
1.75
10
1.45
92.4
18:05-0:45 3
821
1.83
0.02
1.80
1.89
882
5
870
891
859
74
714
953
112
9.6
94.5
122
8.9
10.3
29.8
0.1
: 05-6: 55
826
1.84
0.04
1.80
1.92
875
8
865
887
914
19
866
951
128
8.9
122
134
-3.1
7.4
3.4
-9.1
16.3
5/14/75
25
V
2.20
Yes
Yes
0.84
1.04
10
1.45
9:50-13:20
932
2.05
0.02
2.01
2.10
893
5
876
898
784
29
731
829
117
6.2
112
126
31.0
6.9
37.5
22.1
182
-------
TABLE Gl. SUMMARY OF MINIPLANT RUN CONDITIONS
(See explanation of footnotes on pages 185 and 186)
Run No.
Date
Run Length (hrs)
Cooling Coil Orientation0
Area of Coils, Total (m2)
Limestone Feed
Bed Rejectiond
Bed Depth, Settled (m)
Start of Run
End of Run
Coal/Limestone Feed Ratio
Ca/S Molar Feed Ratio
CombustionfEfficiency, Entire Run (%)'
Conditions
Time Interval
Pressure (kPa)
Super. Velocity (m/sec)
Average8 .
S.D. of 10 Min. Avg's
Minimum 10 Min. Avg.-J-
Maximum 10 Mih. Avg.1
Temperature, Bed (°C)
Average
S.D. of 10 Min. Avg's
Minimum 10 Min. Avg.
Maximum 10 Min. Avg.
Temperature, Flue Gas (°C)n
Average
S.D. of 10 Min. Avg's
Minimum 10 Min. Avg.
Maximum 10 Min. Avg.
Coal Feed Rate (kg/hr)
Average
S.D. of 10 Min. Avg's
Initial 30 Min. AvgJ
Final 30 Min. Avg.J
Excess Air (%)°
Average
S.D. of 10 Min. Avg's
Initial 30 Min. Avg.
Final 30 Min. Avg.
Yes
Yes
10
1.45
15:51-1:31
906
2.13
0.05
2.04
2.26
898
4
891
912
902
28
847
934
133
13.1
118
140
16.7
11.7
28.1
15.8
17.1
5/27/75
55
V
2.20
No
No
0.74
1.95
—
—
95.1
23:31-6:11
913
2.03
0.06
1.86
2.13
863
7
841
869
894
10
867
912
124
5.0
120
130
35.3
6.6
39.6
29.0
18.1
6/9/75
23
V
1.10
Yes
Intermittent
0.66
10
1.45
13:50-3:00
935
1.71
0.06
1.59
1.80
899
10
877
921
850
24
791
877
82.6
5.4
72.2
88.3
57
9
84.4
46.2
183
-------
TABLE 61. SUMMARY OF MINIPLANT RUN CONDITIONS
(See explanation of footnotes on pages 185 and 186)
Run No.a 18.3S 19.2 19.3
Date 6/23/75 7/31/75 8/4/75
Run Length (hrs) 100 10.5 6
Cooling Coil Orientation0 V V V
Area of Coils, Total (m2) 1.10 2.21 2.21
Limestone Feed Intermittent Yes Yes
Bed Rejection*1 Intermittent No First 3 hrs
Bed Depth, Settled (m)
Start of Run 0.76 0.71 1.45
End of Run 0.98 1.45 1.58
Coal/Limestone Feed Ratio 10 10 10
Ca/S Molar Feed Ratio 1.45 1.45 1.45
CombustionfEfficiency, Entire Run (%)e 96.2 95.8
Conditions
Time Interval , 12:56-23:13 (6/25/75)s 11:00-20:10 12:00-16:20
Pressure (kPa) 958 922 922
Super. Velocity (m/sec)
Average8 h 1.86 1.81 1.89
S.D. of 10 Min. Avg's 0.01 0.02 0.01
Minimum 10 Min. Avg.* 1.83 1.77 1.86
Maximum 10 Min. Avg.1 1.89 1.83 1.89
Temperature, Bed (°C)
Average 910 874 879
S.D. of 10 Min. Avg's 1 43
Minimum 10 Min. Avg. 908 866 873
Maximum 10 Min. Avg. 912 882 883
Temperature, Flue Gas (°C)n
Average 901 785 883
S.D. of 10 Min. Avg's 3 74 10
Minimum 10 Min. Avg. 894 664 855
Maximum 10 Min. Avg. 904 887 894
Coal Feed Rate (kg/hr)
Average 89.0 110 136
S.D. of 10 Min. Avg's !.5 15<8 2<2
Initial 30 Min. Avg.J 90.4 82.7 136
Final 30 Min. AvgJ 89.5 133 136
Excess Air (%) °
Average 61>5 30>4 5>7
S.D. of 10 Min. Avg's 3.0 19.5 1.6
Initial 30 Min. Avg. 59.9 70.0 6.0
Final 30 Min. Avg. 60.9 6.0 6.4
184
-------
Notes for Table Gl.
Runs summarized represent those which achieved a reasonable period of
continuous, well-controlled operation. For runs made early in the shake-
down program, this arbitrary period was selected to be at least one hour,
while for later runs, a 5-10 hour minimum was chosen. Also see comment h.
f\
Run N,°.' was assigned in a sequential manner with the number preceding
the decimal point specifying the run series whose first run was made
using an initial bed of fresh limestone. Subsequent runs within a
series, specified by the number after the decimal point, used the
final bed from the preceding run as the starting bed.
Run Length is defined as the period between the start and end of coal
feeding and does not include periods when feeding may have been
interrupted.
c
Cooling Coil Orientation signifies the orientation of the cooling
loops with H indicating horizontal, V indicating vertical, while
mixed orientation is indicated using both symbols with the dominant
orientation listed first.
Bed Rej ectioii indicates whether a portion of the fluidized bed was
purposely being removed during a run in order to limit growth of
solids inventory within the combustor when limestone was fed with
coal.
e
Comb us t i on E f f i cien cy for a run is based on fraction of feed carbon
found in final solids. Since some uncombusted carbon may have escaped
with flue gas, initially calculated efficiency was adjusted downward
using calculated cyclone efficiency and assuming carbon composition
in escaping solids equal to that in solids captured by second stage
cyclone.
Conditions listed (pressure, superficial velocity, etc.) apply to
time interval indicated. Data characterizing mini-plant operation are
automatically recorded at one-minute intervals. Volume of data is re-
duced by averaging data, generally over consecutive 10-minute periods.
S Average is the true average of all one-minute readings over the entire
time interval.
^ S.D. of 10 Min. Ayg's is the standard deviation of the 10-minute
averages over the interval.
1 Minimum (Maximum) 10 Min. Avg. represents the minimum (maximum) value
of the consecutive 10-minute averages over the time interval. It is
presented for those quantities (e.g., superficial velocity, bed tem-
perature) which were purposely controlled at approximately constant
values during a run.
185
-------
J Initial (Final) 30 Min. Avg. is the average of the one-minute readings
over the first (last) 30 minutes of the time interval listed. It is
presented for those quantities which do not necessarily remain constant
during a run but which change continuously in a regular fashion during
a run.
k
Time jnteryal represents a period of continuous, well-controlled
operation during a specific run. In general, especially for long
duration runs, more than one such interval existed. If conditions
were essentially equivalent for different intervals, only one is
presented. If substantial changes in operating conditions occurred,
these are reflected by listing more than one interval.
Pressure was effectively constant (variation of <2%) and, accordingly,
only one value is listed except where noted otherwise. It was measured
above the fluidizing bed.
Superficial Velocity was calculated using the total air flow (combus-
tion plus solids transport) to the combustor with temperature equal
to the bed temperature as defined below and pressure as noted in
table.
Bed Temperature is the average of four readings taken over a distance
of 1.14 meters above the fluidizing grid.
Excess Air is a calculated value based on the total air flow to the
combustor and the coal feed rate, and assumes complete combustion of
coal. Accordingly, values listed are somewhat biased on the high side,
especially at excess air levels below 10-15%.
Conditions shown are typical values. Data logger was inoperable.
Fresh limestone added to final bed from previous run before starting
run.
Pressure varied from 880 to 970 kPa.
s
Data logger was inoperable during much of the run. Data logger was
operating properly during time intervals shown. The combustor per-
formed steadily during most of the run.
No limestone was fed or bed rejected over time intervals selected.
186
-------
TABLE 62.
MINIPLANT NO EMISSION DATA
x
Run No.
10.3
11.1
11.2
12.1
13.2
16.1
18.3
19.2
Timeb
14:15
14:30
14:45
15:00
15:15
15 : 30
15:45
16:00
16:15
16:30
16:45
17:00
17:15
17:30
17:45
18:00
18:15
18:45
13:50-14:00
15:20-15:30
15:50-16:00
16:20-16:30
16:50-17:00
18:00-18:10
19:50-20:00
19:00
20:00
21:00
22:00
23:00
24:00
11:25-11:35
12:25-12:35
13:25-13:35
14:25-14:35
15:25-15:35
16:25-16:35
17:25-17:35
19:25-19:35
N0_ (ppm)c
Jt " ~
90
45
140-185
170
165
160
170
160
155
165
170
160
160
165
175
160
155
170
150
150
160
130
140
135
135
125
115
75
225
225
205
140
130
125
140
210
175
170
160
140
120
115
90
Ib. N02/106 BTUd
0.19
0.13
0.35-0.46
0.33
0.34
0.33
0.38
0.33
0.29
0.35
0.29
0.30
0.29
0.29
0.32
0.25
0.24
0.25
0.20
0.22
0.23
0.26
0.27
0.25
0.24
0.21
0.19
0.10
0.32
0.31
0.26
0.18
0.16
0.15
0.29
0.43
0.33
0.32
0.26
0.22
0.18
0.17
0.12
% Excess Air
^9.2 ^80
^12.1 ^140
•^10.7
^6.7
6.4
5.2
5.1
3.6
2.8
2.6
1.7
6.9
7.4
7.0
7.6
6.8
6.0
6.5
5.8
6.4
5.7
5.6
7.4
4.6
4.5
4.6
3.2
3.7
3.6
6.5
5.3
5.0
4.5
3.8
3.0
1.3
3.2
2.0
2.0
1.2
0.5
0.4
50
55
51
58
49
40
46
39
44
38
37
55
28
28
28
18
22
21
46
34
32
28
22
17
7
18
10
10
6
3
2
^48
^65
44
33
32
21
15
15
9
187
-------
TABLE G2. (CONTINUED) MINIPLANT NOX EMISSION DATA
Run No.
19.3
M
ii
M
Time
14:25-14:35
14:35-14:45
15:25-15:35
15:35-15:45
NO (ppm)
X
65
80
45
80
Ib. NO /10 BTU
2
0.08
0.10
0.06
0.10
% 0
" 2
1.7
1.0
2.1
1.4
% Excess Air
9
5
11
7
Refer to Table Gl, Summary of Miniplant Run Conditions for
operating conditions.
When time interval is indicated, data represent averaged values.
Data obtained by infrared (NO measured) or chemiluminescence (NO + N0_
measured) analysis. Actual NO content was less than 10-15% of total
NO . In practice, both techniques gave essentially equivalent results.
Equivalent NO- based on measured concentration of NO or
Based on flue gas 00 content.
NO
188
-------
TABLE G3. MINIPLANT SOLIDS ANALYSES
Run No. 11.1
Wgt (kg) %Ca %S %S04 %C %CO^
Coal Feed 540.2 - 2.6 - 76.5
Limestone Feed 0.0
Initial Bed (Fresh) 68.0 39.1 - - - 59.6
Final Bed 50.8 35.1 - 26.5 - 1.53
Bed Withdrawal 0.0
First Stage Cyclone3 12.7 17.8 2.9 7.3 16.4
Second Stage Cyclone 40.0 7.4 1.5 4.4 23.3
a Solids from dipleg in first stage cyclone.
Run No. 11.2
Wgt (kg) %Ca %S %SOf %C
Coal Feed 441.8 - 2.6 - 76.5
Limestone Feed 0.0 - - -
Initial Bedb 50.8 35.1 - 26.5 - 1.53
Final Bed 64.4 29.2 - 35.0 - 4.21
Bed Withdrawal 0.0
First Stage Cyclone 27.2 15.3 - 8.5 18.5
Second Stage Cyclone 24.1 2.9 - 3.5 26.6
Does not include 36.3 kg fresh limestone added to bed before
start of run.
189
-------
TABLE G3. MINIPLANT SOLIDS ANALYSES (CONTINUED)
Run No. 12.1
Wgt (kg) %Ca %S %SO^ %C
Coal Feed 619.2 - 2.6 - 76.5
Limestone Feed 52.6 - - - - -
Initial Bed (Fresh) 83.9 39.1 - - - 59.6
Final Bed 131.1 38.0 11.7 34.2 2.4 7.7
Bed Withdrawal 0.0 - -
First Stage Cyclone 10.9 29.4 4.5 12.6 8.3 12.1
Second Stage Cyclone 51.8 11.6 3.1 8.7 17.3 2.4
Run No. 12.2
Wgt (kg) %Ca %S_ %SO= %C %CC5
Coal Feed 793.4 - 2.6 - 76.5
Limestone Feed 78.9 - - -
Initial Bed 131.1 38.0 11.7 34.2 2.4 7.7
Final Bed 176.9 34.3 12.8 34.0 2.5 5.4
Bed Withdrawal 0.0
First Stage Cyclone 19.4 ~ 10.2 26.4 2.1 6.4
Second Stage Cyclone 120.1 11.6 5.2 10.9 30.8 1.6
190
-------
TABLE G3. MINIPLANT SOLIDS ANALYSES (CONTINUED)
Run No. 13.2
(kg) %Ca %£ %SO= %C %CO|
Coal Feed 887.7 - 2.4 - 76.1
Limestone Feed 88.9 39.1 - - - 59.6
Initial Bedc 102. 1 39^1 - - - 59^6
Final Bed 175.1 35.5 12.2 35.1 2.2 9.1
Bed Withdrawal 0.0
First Stage Cyclone 20.9 27.3 4.8 13.5 4.4 6.1
Second Stage Cyclone 59.0 5.5 2.7 6.2 30.7 1.0
Essentially fresh limestone
Run No. 14.1
Wgt (kg) %Ca %S. _%SO| %C %COf
Coal Feed 1548.6 - 2.4 - 76.1
Limestone Feed 129.7 39.1 - - - 59.6
Initial Bed (Fresh) 93.6 39.1 - - - 59.6
Final Bed 141.1 39.0 - 34.3 1.3 2.1
Bed Withdrawal 0.0 - -
First Stage Cyclone 34.0 23.9 5.1 13.5 2.8 4.9
Second Stage Cycloned 177.8 22.4 2.6 5.8 28.5 2.0
d Analyses based on material collected between 15 and 18 hour
point during run.
191
-------
TABLE G3. MINIPLANT SOLIDS ANALYSES (CONTINUED)
Run No. 14.2
Wgt (kg) %Ca %S_ %SO^ %£ %CO=
Coal Feed 2508.9 - 2.4 - 76.1
Limestone Feed 117.5 39.1 - - - 59.6
Initial Bed 141.1 39.0 - 34.3 1.3 2.1
Final Bed 229.5 - - 46.9 1.3 7.0
Bed Withdrawal 0.0 - - -
First Stage Cyclone 53.3 30.7 - - 3.0 2.0
Second Stage Cyclone 191.9 7.7 2.4 5.6 35.7 1.5
Run No. 15.1
Wgt (kg) %Ca %S_ %SO| %C_ %COJ
Coal Feed 2409.5 - 2.4 - 76.1
Limestone Feed 240.9 39.1 - - - 59.6
Initial Bed (Fresh) 116.1 39.1 - - - 59.6
Final Bed 217.7 _____
Bed Withdrawal 220.9 - 8.0 22.0 2.3
First Stage Cyclone - - - -
Second Stage Cyclone 116.0 9.7 - 7.0 24.4 2.3
192
-------
TABLE G3. MINIPLANT SOLIDS ANALYSES (CONTINUED)
Run No. 15.2
Wgt (kg) %Ca ^S %SO| %C %COJ
Coal Feed 985.7 - 2.4 - 76.1
Limestone Feed 98.4 39.1 - 59.6
Initial Bed 217.7 -----
Final Bed 205.9 29.8 9.0 25.6 2.5
Bed Withdrawal 108.0 _____
First Stage Cyclone 24.0 20.5 4.0 11.6 5.0 7.6
Second Stage Cyclone 48.5 10.1 3.1 7.4 27.8 1.3
Run No. 15.3
Wgt (kg) %Ca %S_ %SO° %C_ %CO|
Coal Feed 1262.4 - 2.4 - 76.1
Limestone Feed 126.1 39.1 - 59.6
Initial Bed 205.9 29.8 9.0 25.6 2.5
Final Bed - _ _ - - -
Bed Withdrawal 95.7 37.2 11.3 30.9 4.1
First Stage Cyclone 19.3 26.7 7.1 19.7 12.8
Second Stage Cyclone 67.5 19.4 3.1 7.5 32.7
193
-------
TABLE G3. MINIPLANT SOLIDS ANALYSES (CONTINUED)
Run No.
Coal Feed
Limestone Feed
Initial Bed
Final Bed
Bed Withdrawal
First Stage Cyclone
Second Stage Cyclone
15.4
Wgt (kg)
1778.1
177.8
%Ca
39.1
%S
2.4
%so;
84.1
21.8
125.7
34.5
21.9
10.2
76.1
10.9 31.8 3.3
9.8 - 12.2
3.4 12.5 12.1
13.7
1.0
Run No.
Coal Feed
Limestone Feed
Initial Bed (Fresh)
Final Bed
Bed Withdrawal
First Stage Cyclone
Second Stage Cyclone
16.1
Wgt (kg)
1451.5
145.6
122.5
204.1
11.1
15.6
90.9
%Ca
_
39.1
39.1
46.2
17.0
8.2
6.3
%S
2.2
—
_
5.3
-
3.0
3.1
%SO~
_
—
_
15.7
15.2
4.4
6.4
%C
76.4
—
—
-
-
53.1
32.1
_%co=
_
59.6
59.6
37.4
4.8
2.0
1.8
194
-------
TABLE G3. MINIPLANT SOLIDS ANALYSES (CONTINUED)
Run No. 17.1
Wgt (kg) %Ca %S_ %SO| %C_ %CO|
Coal Feed 6053.7 - 2.2 - 76.4
Limestone Feed - 39.1 - - - 59.6
Initial Bed (Fresh) 82.1 39.1 - 59^
Final Bed 196.8 29.2 9.7 28.6 5.1 24.7
Bed Withdrawal 95.7 33,7 5.6 17.6 7.6 32.9
First Stage Cyclone 17.2 20.1 5.8 11.1 0.7 2.7
Second Stage Cyclone 321.1 9.5 2.7 4.0 20.6 3.0
Run No. 18.3
Wgt (kg) %Ca
Coal Feed 6481.3 - 2.2 - 76.4
Limestone Feed 280.8 39.1 - - - 59.6
Initial Bed 102.1 _ - - -
Final Bed 140.6 18.8 14.2 37.7 0.4 1.0
Bed Withdrawal 86.4 25.5 8 27.0 1.7 7.3
First Stage Cyclone 4.5 - - ~ 0.6
Second Stage Cyclone 475.8 5.1 2.6 13.5 6.1 0.5
195
-------
TABLE G3. MINIPLANT SOLIDS ANALYSES (CONTINUED)
Run No. 19.2
Wgt (kg) %Ca
Coal Feed 1072.3 - 2.2 - 76.4
Limestone Feed 107.0 39.1 - 59.6
Initial Bed 117.9 - - -
Final Bed 195.0 24.Qe 4.9e 23.0 - 20.9
Bed Withdrawal 17.7 32.2 5.9 18.2 - 6.4
First Stage Cyclone - _____
Second Stage Cyclone 149.5 9.6 1.4 6.0 22.5 1.7
e Questionable results.
Run. No. 19.3
Wgt (kg) %Ca %S %S04 _%C %CO~
Coal Feed 714.0 - 2.2 - 76.5
Limestone Feed 70.8 39.1 - - - 59.6
Initial Bed 195.0 24.Oe 4.9 23.8 - 20.9
Final Bed 217.3 25.6^ 7.8 20.8 6.2 29.8
Bed Withdrawal 30.2 23.3 7.7 20.4 5.7 30.2
First Stage Cyclone ~ _ _ _ _ _
Second Stage Cyclone 83.9 4.3 2.9 8.4 27.9 0.7
196
-------
TABLE G4. MINIPLANT CALCIUM AND SULFUR BALANCES
Sulfur (%)
Run No.
11.1
11.2
12.1
12.2
13.2
14.1
16.1
19.3
Average
Calcium (%)
85
70
108
99
110
115
94
88
96 + 15 (IS)
47
57
118
98
93
112
90
88 + 27 (IS)
197
-------
APPENDIX H
BATCH UNIT RUN SUMMARIES
Page
TABLE HI. SUMMARY OF BATCH COMBUSTOR
OPERATING CONDITIONS 199
TABLE H2. SUMMARY OF BATCH COMBUSTOR EMISSIONS DATA 202
TABLE H3. BATCH COMBUSTOR BED AND OVERHEAD SOLIDS
ANALYSES ( RUNS 145C, 2C-5C) 205
TABLE H4. BATCH COMBUSTOR BED AND OVERHEAD
SOLIDS ANALYSIS 206
TABLE H5. SUMMARY OF PARTICIPATE LOADINGS FOR BATCH UNIT 207
TABLE H6. BATCH COMBUSTOR CYCLONE COLLECTION EFFICIENCIES 209
TABLE H7. ANALYSIS OF ARKWRIGHT COAL ASH 210
198
-------
TABLE HI. SUMMARY OF BATCH COMBUSTOR OPERATING CONDITIONS
VO
Run No.
6-2690-
112aC
112b
127
128
129
133
134
136
138
139
141
142
143
145
675-1C
2
3
4
5
6
8
9
11
14
15
17
18
Sorbent
1C
1C
2HC
2HC
2HC
1C
1R
1C
1C
1C
2HC
2HC
2HC
2HC
2HC
2HC
1R
2HC
2HC
1R
2R
1R
1R
1R
1R
1R
1R
Pressure,
kPa
810
810
540
540
870
660
660
610
660
880
660
660
660
660
660
640
630
630
630
840
700
800
810
800
800
810
850
Sup. Vel.,
m/s
0.94
1.02
1.30
1.40
0.92
1.70
1.66
1.46
1.65
1.19
1.95
1.67
1.50
1.50
1.17
1.41
1.47
1.17
1.34
1.43
1.56
1.14
1.17-1.37
0.58
1.06
1.49
1.36
Settled Bed
Depth, m
0.49
0.49
0.49
0.49
0.49
0.17
0.29
0.09
0.23
0.38
0.50
0.50
0.50
0.50
0.50
0.50
0.51
0.50
0.50
0.52
0.50
1.03
0.51
1.03
1.04
0.51
0.51
Bed Temperature °C
Avg.
805
905
805
850
845
870
>1095
740
975
950
1095
955
795
925
845
875
900
925
805
880
875
690
Vertical
850
350
620
920
860
Range
640-925
805-1035
705-920
760-1000
730-905
815-975
760-1150
675-815
805-1180
795-1075
680-1370
860-1045
595-830
610-1000
705-955
730-1000
790-970
760-1010
650-960
790-980
705-945
480-980
Excess
Air, %
43
23
104
153
130
43
8
97
77
27
45
26
110
88
56
61
27
58
43
43
44
8
Coal,
kg./hr.
6.04
7.04
3.90
3.27
3.90
8.40
9.08
5.68
6.04
8.31
7.99
8.76
5.36
5.40
5.45
5.99
7.67
7.35W
7.951
8.99
7.26S
11.30
Run
Length,
hr.
0.30
0.48
0.53
0.43
0.55
2.00
0.33
0.50
2.00
1.00
2.00
1.80
1.00
4.50
1.50
3.00
2.63
3.08
1.02
1.97
2.00
1.25
Coils Installed
827-916
200-800
570-710
880-950
820-880
15
0
0
19
1
10.90
10.90
13.94(a)
10.49
12.39
0.67
0.82
1.08
1.92
1.92
-------
TABLE HI. SUMMARY OF BATCH COMBUSTOR OPERATING CONDITIONS (Cont'd)
O
o
Run No.
20
21
22
23
25
26
28
29
31
32
34
36
37
38
46
50
51
53
54
55
56
Sorbent
1R
1R
Bed from 21C
2R
2R
Bed from 25
1R
Bed from 28
2R
1R
3
3
3
3
1R
1R
1R
Bed from 51C
1R
1R
1R
Pressure,
kPa
800
800
800
800
800
800
800
810
800
800
800
790
800
800
800
800
811
800
800
800
800
Sup. Vel.,
m/s
1.22
1.27
1.26
1.25
1.28
1.23
1.07
1.10
1.58
1.58
1.10
1.23
1.20
1.25
1.14
1.40
1.43
1.45
1.09
1.65
1.10
Settled Bed
Depth, m
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
-
0.
0.
0.
51
51
43 (est)
51
51
27
55
46
88
82
52
52
52
52
46
77
77
-
77
77
55
Bed Temperature °C
Avg.
827
855
850
920
910
975
840
870
895
885
870
863
893
860
911
887
850
857
751
867
877
Range
780-840
810-930
820-910
890-970
875-950
(b)
785-890
750-890
855-950
865-945
860-895
825-895
875-930
815-900
872-963
825-965
795-915
750-975
652-872
717-940
794-975
Excess
Air, %
24
13
15
30
42
138
29
46
5
0
16
27
24
34
32
31
14
24
25
17
22
Coal,
kg
8
9
9
7
7
4
7
6
12
14
8
8
8
8
7
9
11
10
8
12
7
./hr.
.85
.90
.72
.72
.57
.13
.35
.54
.71
.12
.17
.35
.04
.17
.25
.08
.26
.25
.44
.25
.72
Run
Length,
hr.
2.00
1.33
2.92
2.50
5.10
4.50
4.50
5.08
2.12
2.67
5.58
3,83
1.50
1.67
2.00
3.50
3.83
3.53
2.67
1.75
3.50
-------
TABLE HI. SUMMARY OF BATCH COMBUSTOR OPERATING CONDITIONS (Cont'd)
fO
o
NOTES:
Coal: Arkwright Mine, W. Vir., 2.6% S, -16 mesh used unless otherwise noted.
S = Arkwright coal screened to remove fines less than 70 mesh.
W = Wyoming coal, 0.7% S.
I = Illinois coal, 4.1% S.
Stone: (1) = Grove limestone, 8 x 25 mesh (BCR No. 1359).
(2) = Tymoehtee dolomite, 8 x 25 mesh.
(3) = Alundum
C = Calcined
HC = Half-calcined
R = Raw stone, not calcined.
(a) Coal probe positioned downward for 15C and all subsequent runs.
(b) Only one thermocouple in bed.
-------
TABLE H2. SUMMARY OF BATCH COMBUSTOR EMISSIONS DATA
NO (Average)
Run No.
Phase I
ppm
Ib N02
106 BTU
S02 (Average)
Ib S02
%
ppm 106 BTU Reduction
Sampling System
6-2690~112aC
112b
127
128
129
133
134
136
138
139
141
142
143
145
675-1C
2
3
4
5
6
8
9
Vertical Coils
11
14
15
17
18
310
290
340
450
410
340
n.m.
310
300
380
470
610
300
365
335
370
240
250
230
220
160
220
180
170
150
150
170
0.54
0.43
0.83
1.36
1.12
0.59
—
0.71
0.64
0.59
0.82
0.93
0.76
0.82
0.63
0.73
0.37
0.47
0.41
0.38
0.28
0.29
0.22
0.18
0.16
0.22
0.21
240
250
145
250
145
n.m.
1000
1100
860
240
730
1700
120
340
135
150
545
40
1800
200
24
18
170
320
n.m.
25
520
0.58
0.52
0.50
1.05
0.55
—
1.84
3.50
2.56
0.52
1.78
3.62
0.43
1.07
0.35
0.42
1.17
0.10
4.50
0.48
0.06
0.03
0.29
0.47
—
0.05
0.89
85
86
87
72
86
—
50
8
34
87
54
6
89
72
91
89
69
91
35
87
98
99
93
87
—
99
76
S02 (Final)
Ppm
400
550
n.m.
1060
400
1250
225
675
650
600
900
75
2110
640
880
250
1100
630
n.m.
75
790
Ib S02
106 BTU
1.37
2.32
—
3.15
0.86
3.05
0.78
2.12
1.71
1.63
1.93
0.17
4.33
1.54
2.13
0.46
1.85
0.92
—
0.15
1.36
Reduction
64
39
—
18
78
22
79
44
55
57
49
83
24
60
44
88
52
75
—
96
64
Sulfation
55
34
Combustion
Efficiency
(%)
97.5
20
39
82.8
86.6
76.0
59.7
86.3
87.9
% Sulfur
Balance
83
43
6
13
3
95.3
92.3
98.1
94.2
95.3
93
67
106
81
82
120
-------
TABLE H2. SUMMARY OF BATCH COMBUSTOR EMISSIONS DATA (Cont'd)
o
U)
NO (Average)
Ib N02
Run No.
Phase II
Sampling System
20
21
22
23
25
26
28
29
31
32
34
36
37
38
Phase III
Sampling System
46
50
51
53
54
55
56
ppm 106- BTU
180
170
205
330
n.m.
160
n.m.
n.m.
n.m.
n.m.
66
150
160
123
201
176
153
166
217
160
198
0.21
1.23
0.29
0.52
—
0.45
—
—
—
—
0.10
0.23
0.25
0.20
0.33
0.28
0.21
0.25
0.33
0.23
0.30
S02 (Average)
Ib SO? %
ppm 10& BTU Reduction ppm
31
0
280
47
4
160
100
240
60
75
594
830
839
936
545
636
772
1021
573
749
264
0.07
0.00
0.55
0.10
0.07
0.65
0.23
0.58
0.10
0.12
1.20
1.79
1.80
2.10
1.21
1.41
1.50
2.15
1.21
1.48
0.55
98
100
86
97
99
84
94
84
97
97
68
53
53
44
66
63
60
44
68
61
85
130
0
850
340
130
660
265
322
150
150
1130
1140
970
1140
790
870
1130
1510
1170
1130
830
S02 (Final)
Ib S02
106 BTU
0.27
0.00
1.66
0.75
0.31
2.61
0.58
0.79
0.27
0.25
2.29
2.45
2.08
2.58
1.76
1.94
2.19
3.19
2.49
2.25
1.73
7
/a
Reduction
93
100
56
80
92
31
85
79
93
93
40
35
45
32
51
49
42
17
34
41
54
% Sulfation
27
35
63
27
61
30
48
15
26
42
54
20
20
Combustion
Efficiency % Sulfur
(%) Balance
88.5
89.8
86.6
94.3
96.8
97.6
93.5
94.3
94.0
93.0
87.0
71
43
31
86
123
94
158
83
n.m. = not measured because of problems with equipment
-------
TABLE H-2. SUMMARY OF BATCH COMBUSTOR EMISSIONS DATA (CON'T)
Run No. 3675- CO (ppm) Temperature (°C) Excess Air _(%)_
1C 182 845 56
2 450 875 61
3 320 900 27
4 1190 925 58
5 2040 805 43
6 830 880 43
8 950 875 44
9 1610 690 8
11 4740 850 55
14 1400 350 0
15 n.m. 620 0
17 570 920 19
18 1900 860 1
20 3100 827 24
21 610 855 13
22 1230 850 15
23 47 920 30
25 105 910 42
26 4 975 58
28 430 840 29
29 129 870 46
31 700 895 5
32 700 885 0
34 710 870 16
36 470 863 27
37 407 893 24
38 710 860 34
46 n.m. 911 32
50 123 887 31
51 325 850 14
53 397 857 24
54 3642 751 25
55 71 867 17
56 165 877 22
204
-------
TABLE H3. BATCH COMBUSTOR BED AND OVERHEAD
SOLIDS ANALYSIS (RUNS 145C, 2C-5C)
Run 14 5C
Run 2C
Run 3C
Run 4C
Run 5C
Bed
CaSO,
CaC03
CaO
MgO
MgC03
AJUO,
/ 3
Si02
Fe203
NiO
Na20
V-0,-
2 5
H2°
Overhead
CaS04
CaC03
CaO
MgO
MgC03
A100Q
2 3
Si°2
Fe 0
2 3
NiO
Na2°
\r n
25
C
H2°
41.13
3.30
12.01
23.21
2.96
7.51
3.51
3.01
7.49
14.51
12.57
16.60
7.65
16.11
4.31
11.00
2.30
37.70 9.13
2.60 63.26
18.80 21.43
28.20 <0.32
4.40 <0.41
7.70 0.42
4.00 0.16
<0.15
<0.12
<1.82
9.40
10.21
2.97
1.65
11.32
18.63
6.06
<0.15
0.68
<1.82
19.00 36.08
12.39
24.72
20.93
17.73
2.65
11.12
2.39
<0.15
<0.31
<1.82
11.85
16.82
16.16
15.43
7.24
16.47
3.19
<0.15
<0.31
<1.82
6.18
2.59
64.26
0.00
21.64
0.71
1.02
3.61
0.16
<0.15
<0.31
<1.82
8.58
13.91
0.00
4.72
5.99
7.34
19.05
7.82
<0.15
0.81
<1.82
21.78
All values are weight percent
205
-------
TABLE H4. BATCH COMBUSTOR BED AND OVERHEAD SOLIDS ANALYSIS
K>
O
Bed
Run No.
3675-14
15
17
18
20
21
22
23
25
26
28
29
31
32
34
36
37
38
46
50
51
53
54
55
56
Ca
23
29
27
20
34
28
26
35
51
40
20
18
38
46
++
.90
.50
.20
.00
.80
.40
.80
.50
.10
.10
.21
.81
.70
.30
SO
22
22
22
21
29
18
22
30
22
41
19
41
18
24
20
24
18
21
4~ COS" Total Ca
.10
.30
.40
.00
.00
.90
.60
.30
.90
.50
.20
.00
.48
.66
.56
.55
.78
.87
58.
64.
4.34 60.
48.
53.
53.
67.
2.93 29.
22.
12.
29.90 45.
29.
18.
46.
36.
46.
45.
51.
36.16 42.
52.
54.
31.03 50.
8.17 35.
14
33
76
62
50
81
03
80
62
00
76
32
43
68
90
95
51
41
02
22
15
08
01
Cyclone 1
Total S S04= CO-}= Total C3
1.
1.
1.
1.
2.
2.
1.
2.
1.
6.50 68.48
2.76 65.88
2.56 36.37
3.60
0.77
1.85
2.63
9.45 36.14
7.70 39.63
11.30
5.89
59 2.44
24 1.84
28 1.83
75 2.45
19 1.41
20
80
30 0.71 27.94
30
Cyclone 2
Total S S04= C03= Total C&
6.50
2.85 47
8.51 44
11
0.43 28
5.05 46
30
41
31
42
30
40
39
51
2.49
58
.19
.74
.66
.67
.29
.20
.75
.80
.31
.83
.56
.64
.95
.33
Filter
Total S S04= C03=
9.51
3.26
0.43
0.87
4.77
11.90
7.06
5.89
6.35
3.48 1.41
2.53 0.66
3.21
2.72
2.02
All values are weight percent
3. ss
Includes carbon present as COo
-------
TABLE H5.
SUMMARY OF PARTICULATE LOADINGS FOR BATCH UNIT
to
o
Run Number
675-1C
2
3
4
5
6
8
9
11
14
15
17
18
20
21
22
23
25
26
28
29
31
32
34
36
37
38
46
(T)
(T)
(G)
(T)
(T)
(T)
(G)
(G)
(G)
(G)
(G)
(G)
(G)
(T)
(T)
(T)
(G)
(G)
(T)
(G)
(A)
(A)
(A)
(A)
(G)
Bed Outlet Loading (gr/scf)
7.83
7.25
8.24
7.91
9.35
No data available
6.99
No data available
11.54
Solids caught by filter burned
36.94
9.23
11.32
7.83
8.19
8.23
7.43
5.48
4.31
5.75
5.74
12.48
12.26
4.85
5.22
5.73
5.40
6.40
Cyclone #1 Outlet Loading
(gr/scf)
Cyclone #2 Outlet Loading
(gr/scf)
Particulates from Cyclones 2 and 3 filter not weighed.
0.33 0.03
1.16 0.38 (Cyclone #2 plugged)
0.01 0.01
0.07 0.07
0.73
8.29
0.76
3.60
0.72
0.76
0.35
0.61
0.87
0.78
1.64
2.19
3.
4.
1,
.42
.93
.66
1.11
0.72
0.60
1.07
0.60
11.54 (Cyclone #1 plugged) 3.43
1.42
0.19
0.27
0.10
0.15
0.10
0.35
0.78
0.47
0.84
0.70
3.
4.
1,
22
62
49
0.61
0.14
0.12
0.95
(Cyclone #2 plugged)
(Cyclone #2 plugged)
(Cyclone #2 plugged)
-------
TABLE H5. (Continued) SUMMARY OF PARTICULATE LOADINGS FOR BATCH UNIT
Run Number
50
51
53
54
55
56C
(G)
(G)
(G)
(G)
(G)
(G)
8.30
5.73
5.59
8.85
7.92
6.47
Bed Outlet Loading (gr/scf)
Cyclone #1 Outlet Loading
(gr/scf)
Cyclone #2 Outlet Loading
(gr/scf)
3.58
1.88
2.21
2.57
1.26
0.63
3.12
1.51
0.55
1.53
0.09
0.42
O
00
Note: Letters in parenthesis indicate bed material -
(G) - Grove limestone
(T) - Tymochtee Dolomite
(A) - Alundum
-------
TABLE H6. BATCH COMBUSTOR CYCLONE COLLECTION EFFICIENCIES
Run No- Cyclone I Cyclone 2 Overall
3675-2C 0.95 0.90 1.00
8 0.90 0.18 0.91
15 0.78 0.83 0.97
17 0.92 0.75 0.98
18 0.68 0.93 0.98
20 0.91 0.86 0.99
21 0.91 0.80 0.98
22 0.96 0.71 0.99
23 0.92 0.43 0.95
25 0.84 0.10 0.86
26 0.82 0.40 0.89
28 0.72 0.48 0.85
29 0.62 0.68 0.87
31 0.73 0.10 0.74
32 0.58 0.10 0.61
34 0.66 0.10 0.69
36 0.79 0.45 0.88
37 0.88 0.80 0.98
38 0.89 0.80 0.98
46 0.83 0.11 0.85
50 0.57 0.12 0.62
51 0.67 0.20 0.74
53 0.60 0.75 0.90
54 0.71 0.41 0.83
55 0.84 0.93 0.99
56 0.90 0.33 0.94
„„„. . mass collected
Efficiency =
mass
209
-------
TABLE H7. ANALYSIS OF ARKWRIGHT COAL ASH
Component
CaO
Fe000
2 3
A1000
2 3
Si02
Na2°
Wt. %
2.7
19.0
32.1
47.1
1.5
Mole %
3.7
9.2
24.5
60.8
1.9
Other -2.4
Total ash in coal = 7.3 wt. percent,
210
-------
TECHNICAL REPORT DATA
(rlease read Instructions on the reverse before completing
EPA-600/7-76-Q11
3. RECIPIENT'S ACCESSION NO,
». TITLE ANDSUBTITLE
STUDIES OF THE PRESSURIZED
FLUIDIZED-BED COAL COMBUSTION PROCESS
5. REPORT DATE
September 1976
6. PERFORMING ORGANIZATION CODE
•AUTHOR(S)R.C.Hoke, R.R.Bertrand, M.S.Nutkis, D D
Kinzler, L.A.Ruth, and M. W. Gregory
8. PERFORMING ORGANIZATION REPORT NO.
GRU. 15GFGS.76
9, PERFORMING OROANIZATION NAME AND ADDRESS
Exxon Research and Engineering Company
P. O. Box 8
Linden, New Jersey 07036
10. PROGRAM ELEMENT NO.
EHE623A
11. CONTRACT/GRANT NO.
68-02-1312 and -1451
12, SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
13. TYPE OF REPORT AND I
Phase; 8/73-8/75
14. SPONSORING AGENCY CODE
EPA-ORD
is.SUPPLEMENTARY NOTES ffiRL-RTP project officer for this report is D. B. Henschel, Mail
Drop 61, 919/549-8411 Ext 2825. APTD 1116 and EPA-650/2-74-001 (NTIS PB 210-246
and 231-374) are earlier EPA reports relating to this work.
16. ABSTRACT
The report gives results of studies of the environmental aspects of the pres-
surized fluidized-bed coal combustion (FBCC) process, using two experimental facil-
ities: a new 218 kg coal/hr "miniplant" combustor (0. 63 MW equivalent), and a 13 kg
coal/hr "batch" combustion unit. Successful shakedown of the miniplant combustor
culminated in a continuous 100-hr run. The miniplant combustor was operated at:
coal rates up to 155 kg/hr (340 Ib/hr), pressures up to 1020 kPa (10 atm), superficial
velocities up to 3.2 m/s (10. 5 ft/sec), temperatures up to 980C (1800F), and combus-
tion intensities of 5 MW/cu m (480,000 Btu/hr-cu ft). Improvements in the coal feed-
ing system and in the steam coil design were required to achieve these performance
levels Operating results from both facilities indicate that SO2 emissions can be
controlled to meet current EPA New Source Performance Standards for coal-fired
utility boilers with either limestone or dolomite sorbents. NOx emissions from
pressurized FBCC can be controlled to 0. 2 to 0. 4 Ib (as NO2)/million Btu (0. 09-
0.17 g/million J), compared to the current EPA standard of 0. 7 Ib/million Btu (0. 30
g/million J). Particulate emissions cannot be controlled to the current EPA standard
with two stages of conventional cyclones.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Desulfurization
Flue Gases
Limestone
Dolomite (Rock)
Calcium Oxides
Fluidized-Bed
Processors
Combustion
Air Pollution Control
Stationary Sources
Fluidized-Bed Combus-
tion
Limestone-Based Desul-
furization Process
13B
07A,07D
2 IB
08G
07B
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
?1. NO. OF PAGES
211
20. SECURITY CLASS (Thispage)
Unclassified
EPA Form 2220-1 (9-73)
211
------- |