U.S. Environmental Protection Agency Industrial Environmental Research EPA-600/7-78-069
Office of Research and Development Laboratory .. ^ft^o
Research Triangle Park. North Carolina 27711 Apfll 1978
MINIPLANT STUDIES
OF PRESSURIZED
FLUIDIZED-BED
COAL COMBUSTION:
Third Annual Report
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 nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine 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
8. "Special" Reports
9. Miscellaneous Reports
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 sys-
tems. The goal of the Program is to assure ihe rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses 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 environ-
mental issues.
EPA 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 recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-78-069
April 1978
MINIPLANT STUDIES OF PRESSURIZED
FLUIDIZED-BED COAL COMBUSTION:
Third Annual Report
by
R.C. Hoke, R.R. Bertrand, M.S. Nutkis,
L.A. Ruth, M.W. Gregory, E.M. Magee,
M.D. Loughnane, R.J. Madon, A.R. Garabrant, and M. Ernst
Exxon Research and Engineering Company
P.O. Box 8
Linden, New Jersey 07036
Contract No. 68-02-1312
Program Element No. EHE623A
EPA Project Officer: D. Bruce Henschel
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, N.C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
-------
ABSTRACT
The pressurized fluidized bed combustion of coal and regeneration of
spent sorbent were studied in the continuous 480 Ib coal/hr (220 kg/hr)
"miniplant" FBC unit. The effect of coal sulfur content and the use of pre-
calcined limestone sorbent on control of SC>2 emissions was studied. NOX
emissions were well within the current new source performance standard for
NOX. Particulate emissions in the flue gas after passage through cyclone
cleaners exceeded the new source performance standard.
The continuous operation of the combustor and regenerator sections was
demonstrated in a 125 hour run. SC>2 emissions in the combustor flue gas were
less than the new source performance standard at all times during the run.
Makeup sorbent fed to the combustor was about 25% of the rate which would
have been required in once through operation to maintain the S02 emissions
at the same overall average level.
A granular bed filter was installed in the miniplant flue gas stream
and initial shake down was completed in a 24 hour continuous run. The
minimum particulate concentration measured in the filter outlet gas was about
0.1 g/m^ (0.05 gr/SCF), somewhat higher than anticipated.
A program to provide a comprehensive analysis of all potentially harmful
emissions from a pressurized FBC unit began in the miniplant. Analysis of
materials including those present in trace concentrations was completed.
The smaller 28 Ib coal/hr (13 kg/hr) batch unit was modified to operate
in a continuous fashion.
This report is submitted in fulfillment of Contract Number 68-02-1312
by Exxon Research and Engineering Company under sponsorship of the
Environmental Protection Agency. Work was completed in August, 1977.
iii
-------
TABLE OF CONTENTS
Abstract
List of Figures v
List of Tables ix
Acknowledgements x
Sections
I Summary 1
II Introduction 11
III Combustion Studies 15
Equipment, Materials, Procedures 15
Experimental Results and Discussion 46
IV Regeneration Studies 65
Equipment, Materials, Procedure 65
Experimental Results and Discussion 70
V Granular Bed Filtration Studies 84
Equipment, Procedures 85
Experimental Results and Discussion 100
VI Modification of Batch Unit 115
Combustor Section 115
Regenerator Facilities 121
VII Comprehensive Analysis 126
VIII Analysis of Desulfurization Data 127
IX Continuing Studies 150
X References 153
XI List of Publications 155
XII Appendices 157
iv
-------
LIST OF FIGURES
Page
II-l Pressurized Fluidized Bed Coal Combustion System 12
II-2 Exxon Fluidized Bed Combustion Miniplant 14
III-l Exxon Fluidized Bed Combustion Miniplant (Schematic) 16
III-2 Coal and Limestone Feed System 17
III-3 Combustor Vessel 19
III-4 Cooling Coils After 600 Hour Exposure 20
III-5 Liquid Fuel System 21
III-6 Liquid Fuel Injector Nozzle 23
III-7 Combustor Fluidizing Grid 24
III-8 Combustor Corrosion Probe Location 25
III-9 Combustor Lower Corrosion Probe Section 26
111-10 Combustor Lower Corrosion Probe and Bed Sampling Probe 27
III-ll Combustor Upper Corrosion Probe Section 29
111-12 Combustor Corrosion Probe 30
111-13 Westinghouse Erosion Probe 32
111-14 Combustor Bed Sampling Probe 33
111-15 Sorbent Utilization Vs. Time 34
111-16 Miniplant Particulate Sampling System 36
111-17 Combustor Escentric Acurex Probe Section 38
111-18 Acurex Probe 39
111-19 Coal Particle Size Distribution 41
111-20 Sorbent Particle Size Distribution 44
111-21 S02 Retentionvwith Dolomite Sorbent 47
111-22 S02 Retention with Precalcined Lir.estone 49
-------
LIST OF FIGURES (Continued)
Page
111-23 Effect of Low Temperature Operation on S02 Retention 50
111-24 Correlation of NO Emissions 53
X
111-25 NO Vs CO in Flue Gas 55
X
111-26 Calculated Combustion Efficiency Vs Observed 62
Combustion Efficiency
111-27 Secondary Cyclone Efficiency Vs. Particle Size 64
IV-1 Miniplant Solids Transfer System (Schematic) 68
IV-2 Miniplant Solids Transfer System 69
IV-3 Combustor and Regenerator Bed Heights During 75
Demonstration Run
IV-4 Total Inventory of Sorbent in the Combustor and 76
Regenerator during Demonstration Run
IV-5 Combustor S0« Emissions During Demonstration Run 78
IV-6 S02 Emission Vs. Ca/S Ratio 79
V-l Single Bed Test Rig 86
V-2 Single Bed Test Rig Schematic 87
V-3 Schematic of Single Ducon Filter Bed 88
V-4 Granular Bed Filter Operation 89
V-5 Ducon Filter Element and Shroud 90
V-6 Original Exxon Filter Element and Shroud 91
V-7 Ducon Sonic Filter Schematic 93
V-8 Schematic of a Single Exxon Filter Bed 94
V-9 Modified-Exxon Filter Element 95
V-10 Modified Filter Bed 96
V-ll Granular Bed Filter Pressure Vessel and Structure 97
vi
-------
LIST OF FIGURES (Continued)
Page
V-12 Granular Bed Filter Pressure Vessel Interior 99
V-13 Test Rig Results - AP Vs. Time 101
V-14 Ducon Filter with Plugged Screens 103
V-15 Effect of Blow Back on Filter Pressure Drop 109
Run 55
V-16 Modified Exxon Filter with 111
Plugged Inlet Retaining Screens
V-17 Flue Gas Temperature Profile 113
V-18 Schematic of the Ejector Blow Back System 114
VI-1 Schematic of Modified Batch Unit 116
VI-2 Bench Unit Cyclone Assembly 118
VI-3 Bench Unit Coal Injector Vessel 120
VI-4 Bench Unit Regenerator Fluidizing Grid 122
VI-5 Bench Unit Regenerator Cyclone 125
VIII-1 Ln (1-SR) Vs. Gas Phase Residence Time 129
VIII-2 Sulfur Removal Efficiency Vs. Ca/S 130
VIII-3 Sulfur Removal Efficiency @ 2 Sec. Vs. Ca/S 131
VIII-4 Sulfur Removal Efficiency @ 2 Sec. Vs. Temperature 133
VIII-5 Effect of Temperature on Desulfurization Rate Constant 134
VIII-6 Comparison of Set Vs. Calculated Ca/S Effects 135
VIII-7 Sulfation Rate Constant Vs. Sorbent Utilization 141
VIII-8 SO Retention @ 2 Sec. Vs. Ca/S Set 142
VIII-9 Time for a Particle to Reach Calcium Utilization 143
Level in Atmosphere of 2500 ppm SO at 900°C
vii
-------
LIST OF FIGURES (Continued)
Page
VIII-10 Contribution of Particles in Bed for Various Time 145
Periods to Steady State Value of Given Parameters
VIII-11 SO Retention @ 2 Sec, 900°C Vs. Ca/S Molar Ratio 147
VIII-12 Time in Combustor Vs. Ca/S Ratio 148
viii
-------
LIST OF TABLES
Page
III-l Coal Composition 42
III-2 Sorbent Composition 45
III-3 Sorbent Requirements 52
III-4 Particulate Sampling Summary I (Before Modification 56
to Sampling System)
III-5 Particulate Sampling Summary II (After First 57
Modification to Sampling System
III-6 Comparison of Particulate Loadings Measured by 58
Total Filter (Exxon) and SASS Train (Battelle)
III-7 Sizing of Particles Obtained on Balston Filters 58
Before Modification
III-8 Individual Filter Results (Runs 50 to 59) 59
III-9 Sizing of Particles Obtained on Balston Filters 60
After Modification
IH-10 Comparison of Particle Size Distributions 60
Measured by Three Different Methods
III-ll Second Stage Cyclone Lock Hopper Particle 63
Size Distributions
IV-1 Operating Conditions During Demonstration Run 71
IV-2 Feed Rates of Makeup Limestone to Combustor During 74
Combustor-Regenerator Demonstration Run
IV-3 Analyses of Bed Material Discharged from Combustor 81
and Regenerator After Demonstration Run
IV-4 Sulfur Balance Combustion-Regeneration 82
Demonstration Run
V-l Granular Bed Filter Run Summary for Modified 106
Filter Elements
VIII-1 Comparison of Rate Constants 146
ix
-------
ACKNOWLEDGEMENTS
Many individuals at Exxon Research and Engineering Company played major
roles in the conduct of this program. The authors wish to acknowledge the
contributions made by them and to express their gratitude. We particularly
wish to acknowledge the efforts of H. R. Silakowski, the miniplant operations
supervisor. His contributions played a large part in the successful opera-
tion of the miniplant. We also wish to acknowledge the efforts of the
operating and mechanical crews, A. Altobelli, J. Bond, R. Burakiewicz, D.
Duffy, J. Fowlks, T. Gaydos, E. Hellwege, F. Huber, M. Moroski, T. Morrison,
S. Pampinto, J. Sansone, R. Schroeder, W. Spond, T. Sutowski, L. Tucker and
G. Walsh. We also wish to thank our math clerk, S. Walther, W. Dravis and
N. Bissoni of the Mechnical Division G. Milliman of the Analytical Division,
V. S. Engleman, and G. A. Gagliardo for their support. A special acknowledge-
ment goes to N. Malinowsky who typed this report.
The personnel of the Industrial Environmental Research Laboratory of the
EPA have been most helpful and deserve special thanks. We wish to express
our gratitude for the help of D. B. Henschel, the EPA Project Officer,
P. P. Turner and R. P. Hangebrauck.
-------
SECTION I
SUMMARY
The pressurized fluidized bed combustion of coal (PFBC) and regeneration
of sulfated SC>2 sorbent were studied in the continuous miniplant unit. In
the combustion program, the effect of operating conditions on S02, NOX, and
particulate emissions from the miniplant was studied. In this latest series
of combustion tests, the effect of burning a higher sulfur Illinois No. 6
coal was determined, the effect of using precalcined limestone sorbent and
operation at very low combustor "turndown" temperatures were studied. Sorbent
levels required to meet S02 emission standards were determined. Combustion
efficiency and cyclone collection efficiency were also measured.
The regenerator and combustor sections of the miniplant were operated
continuously for 125 hrs and the operability of the system was demonstrated.
The drop in activity of the regenerated sorbent with time, and the SC>2 con-
tent in the regenerator off gas were measured.
A granular bed filter was installed on the miniplant flue gas and initial
shakedown was completed. A comprehensive analysis of all potentially hazardous
emissions from the miniplant was completed. The smaller, semi-batch unit was
not operated during this period, but was modified to operate in a continuous
fashion.
COMBUSTION STUDIES
The miniplant combustor consists of a refractory lined vessel 10 m (33
ft) high with an inside diameter of 32 cm (12.5 in). A number of vertical
water-cooled tubes are mounted in the combustor to remove the heat of com-
bustion. Two new sections of the combustor were installed to hold air-cooled
tubes used for materials testing purposes. One of the sections holds tubes
immersed in the fluidized bed, the other holds tubes mounted in the freeboard
above the fluidized bed.
Premixed coal and sorbent are injected into the combustor a single point
28 cm (11 in) above the fluidized bed support grid. The combustor is capable
of operating at pressures up to 1000 kPa (10 atm), at temperatures up to the
ash agglomeration temperature of the coal (usually less than 980°C), at
superficial velocities of up to 3 m/s (10 ft/sec) and with expanded beds of
up to 6.1 m (20 ft). The maximum design coal feed rate is 200 kg/hr (480
Ib/hr). Flue gas leaving the combustor passes through two cyclones in series
to remove most of the particulate matter. A granular bed filter was recently
installed and piped into the flue gas system such that the flue gas leaving
the second stage cyclone can either be sent to or by pass the filter.
Particulates captured in the first cyclone are recycled to the combustor to
improve combustion efficiency. Particulates captured in the second stage
cyclone and the granular bed filter are rejected through lock hoppers.
Spent sorbent is also rejected from the combustor through a lock hopper
system to maintain a constant bed level in the combustor.
-------
Runs were made with an Eastern bituminous Pittsburgh seam coal (Champion)
containing 2% sulfur screened to a particle size distribution of 200 to
2400 microns, and an Illinois No. 6 coal containing 4.2% sulfur also screened
to a particle size distribution of 200 to 2400 microns. Two sorbents were
used: a Virginia limestone (Grove No. 1359) and an Ohio dolomite (Pfizer No.
1337). Both were screened to a size range of 840 to 2400 microns.
Operational performance of the combustor continued to be good. As of
July 1977, over 1500 hrs of coal combustion time was accumulated. Additional
modifications were made to the combustor which further improved performance.
The baffled vertical cooling coils performed very well, showing no sign
of erosion after about 600 hrs of running time. A new fluidization grid was
fabricated with better distribution of the cooling water flow through the
grid. The refractory lining in the second stage cyclone began to fail and
was replaced with a more erosion resistant material. A sampling system was
also developed which permits the extraction of solids from the combustor
during the run.
S02 retention results were obtained using 4.2% sulfur Illinois No. 6 coal
and Pfizer No. 1337 dolomite. S02 retention data plotted against the Ca/S
molar ratio fit reasonably well with the data obtained with 2% sulfur Eastern
coal. This suggests that the sulfation reaction is first order in 862 con-
centration and is consistent with results published by other laboratories.
S02 retention was also measured using precalcined Grove No. 1359 limestone
sorbent. Calcination was carried out in the combustor burning natural gas at
930 kPa (9 atm) and 870°C (1600°F). The calcined sorbent was then removed
and fed with the coal during a subsequent series of runs. The precalcined
limestone was found to be as active, at an equivalent Ca/S ratio, as dolomite
and was much more active than "raw" limestone. The activity was also main-
tained at temperatures as low as 865°C, where limestone would be expected
to be present as the carbonates. The precalcined limestone was also active
at a very low "turndown" temperature (760°C) where previously, "raw" limestone
was shown to be completely inactive. At 760°C, precalcined limestone is
slightly less active than at temperatures in excess of 865°C. It is, however,
still as active as dolomite, at an equivalent Ca/S ratio, even though the
low temperatures would strongly favor the formation of CaCOo rather than
CaO. The high activity is believed due to the formation of very large
pores during precalcination. The pores are apparently large enough that
carbonation of the stone does not prevent diffusion of S02 into the interior
of the particles. Sorbent requirements were then estimated for precalcined
limestone and compared to dolomite and limestone requirements. Precalcined
limestone is more effective, on a weight basis, than either dolomite or
limestone. The weight requirements, expressed on an uncalcined limestone
basis, are half the dolomite requirements and as little as 40% of the require-
ment for "raw" limestone.
NOX emissions were found to follow the same trend line developed in
earlier studies. The NOX emissions vary from 50 to 200 ppm or 0.04 to 0.17 g
(as N02)/MJ (0.1 to 0.4 Ib/M BTU). The primary variable affecting NOX emis-
sion is excess air (or 02 content in the flue gas). Temperature over the
range of 670 to 940°C (1250 to 1750°F) had a secondary effect. The emissions
are well below the EPA new source performance standard of 0.3 g/MJ (0.7 Ib/M
BTU) and have an average value of 0.09 g/MJ (0.2 Ib/M BTU) at 15% excess air,
-------
the level most likely to be used in commercial practice. Indications that
NOX could be reduced even further by two stage combustion were made during
an upset period of one run. In this case, oxygen content in the flue gas
dropped to very low levels, CO content increased to 1800 ppm and as it did,
the NOX content dropped to 10 to 20 ppm.
Particulate emissions in the flue gas after two stages of cyclone cleanup
averaged about 2.3 g/m^ (1 gr/SCF) with a mass mean particle size of 3 to 6
microns. The second stage cyclone was found to have an overall efficiency of
85% with a 50% cut point between 3 and 3.5 microns.
Carbon combustion efficiency results were satisfactorily correlated
using multiple regression. Temperature and excess air were shown to be the
most significant variables. Two other variables which had a significant
effect were the cube of the residence time and the cross product of tempera-
ture and excess air.
REGENERATION STUDIES
The regenerator consists of a refractory lined vessel with an inside
diameter of 22 cm (8.5 in) and an overall height of 6.7 m (22 ft). Gaseous
fuel is burned in a plenum below the fluidized bed to achieve the reaction
temperature. Additional fuel is injected directly into the fluidized bed
just above the fluidizing grid to create a reducing zone in which the CaSO^
reduction reaction occurs. Supplementary air is injected directly into the
bed at a higher elevation to create an oxidizing zone. The oxidizing environ-
ment at the top of the bed assures high selectivity to CaO, the desired pro-
duct of the regeneration reaction, by minimizing the formation of CaS, an
undesired by product.
The successful operation of the coupled regeneration/combustor system
was demonstrated in a 125 hr uninterrupted run. The miniplant was operated
with limestone sorbent continuously recirculating between the combustor and
regenerator until the run was voluntarily terminated. The major purpose
of the run was to demonstrate that the system could operate continuously for
100 hrs. Operating conditions were deliberately chosen to be conservative
in order to maximize the chance of reaching this goal. All conditions were
held constant except the feed rate of fresh limestone into the combustor.
This was adjusted to maintain constant bed levels in the combustor and
regenerator. Used sorbent was not planned to be removed from the system
during the run unless the S02 content in the combustor flue gas exceeded the
EPA new source performance standard of 1.2 Ib/M BTU. At that point, used
sorbent was planned to be removed and the fresh sorbent feed rate would be
increased. Pressure in the two vessels was controlled at a somewhat lower
level (about 760 kPa (7.5 atm)) during most of the run in order to increase
fluidization velocity in the regenerator to a more realistic level and to
avoid agglomeration problems. Regeneration temperature was also held con-
stant at a lower level, 1010°C (1850°F) to minimize the chances of bed
agglomeration. Combustion temperature was 900°C (1650°F), Eastern coal and
Grove No. 1359 limestone were used.
For the first 24 hours of the run, the regenerator was operated in
oxidizing conditions in order to establish base line operating conditions.
Sorbent recirculated between the combustor and regenerator during this
-------
period. Subsequently, reducing conditions were established by increasing
the flow of supplementary fuel to the regenerator. Operation of the system
continued, uninterrupted, for the next 100 hours. Operation of the com-
bustor-regenerator during the extended run was exceptionally smooth. A hot
spot did develop in the regenerator just above the grid in which the tem-
perature was 200 to 250°C (450°F) higher than the rest of the regenerator
The temperature profile was improved somewhat by decreasing the pressure to
760 kPa, thereby increasing the fluidization velocity. Some signs of
agglomeration were noted in the bed after completion of the run. This could
have caused the hot spot to form.
The limestone addition rate was initially set at a Ca/S ratio of 0.74.
However, solids inventory in the combustor and regenerator began dropping
and the feed rate was increased to a Ca/S ratio of 1.3. Part of the drop in
bed levels was due to the inadvertent removal of some combustor solids early
in the run. After the bed levels increased, the Ca/S was dropped to 1.06,
then 0.46 and finally was decreased to 0 for the last 40 hours of the run.
S02 emissions from the combustor leveled out at 550 ppm (about 1.2 Ib/M
BTU) during the first 24 hours of the run while the regenerator was opera-
ting under oxidizing conditions. Emissions fell rapidly to less than 200 ppm
when reducing (regenerating) conditions were established in the regenerator.
S02 emissions from the combustor increased as the run progressed and as the
sorbent deactivated. At the end of the run, after 100 hours of regeneration,
the emissions were about 550 ppm. If the run had gone longer, removal of
used sorbent and an increase in the fresh Ca/S ratio would have been required
to maintain the S02 emissions at this level. During the 100 hours of regen-
eration, the sorbent underwent about 15 cycles of combustion and regeneration.
The average Ca/S ratio during the run was 0.55, although as mentioned above,
it was varied from 0 to 1.3. This represents about a four fold reduction in
the limestone feed rate which would have been required in a once through
system to control the S02 emissions to the average level measured in this run.
The concentration of S02 in the regenerator off gas was nearly steady
throughout the run and averaged 0.53 mole percent (dry basis). This is very
close to the concentration predicted by a sulfur mass balance based on the
feed rate and sulfur content of the coal entering the combustor. The cal-
culated equilibrium concentration at the operating conditions of the regen-
erator was 2.9 percent; hence, higher S02 levels would probably have been
achieved by burning in the combustor more coal of a high sulfur content.
The sulfur balance of the run was 104%, an acceptable level.
GRANULAR BED FILTRATION STUDIES
The objective of the granular bed filtration program was to determine
if such a filter could reduce particulate concentrations in the flue gas, at
high temperature and high pressure, to a level which would prevent damage to
the gas turbine as well as satisfy environmental requirements. Tentative
estimates of the particulate concentration which could be tolerated by the
turbine range from 45 to 1 mg/m3 (0.02 to 0.004 gr/SCF), and also depend on
the particulate size distribution. Current environmental standards requires
the particulate emission to be less than approximately 100 mg/m3 (0.05
gr/SCF). To meet these requirements, a particulate removal system consisting
of two cyclones followed by a high efficiency third stage device such as a
-------
granular bed filter is now envisioned. The granular bed filter test program
is intended to determine if the filter can achieve these particulate emission
targets, if the performance can be maintained, to uncover operating problems
and to measure the long term life of the filter hardware.
A granular bed filter was purchased from the Ducon Company and evaluated.
The Ducon-type granular bed filter consists of a number of small beds packed
with suitable granular filter media such as alumina, quartz, etc. A stack
of the filter beds form a single filter element. A number of filter elements
can be used depending on the volume of gas to be filtered. Dirty gas passes
through inlet screen sections down into the filter beds immediately below the
screen sections. Clean gas from the beds is collected in a manifold in the
interior of the element and then passes to the clean gas outlet system. As
the filtration step proceeds, the pressure drop across the element increases
and eventually the element must be cleaned by the reverse flow of clean gas.
This "blow back" occurs by flowing clean gas in reverse direction through the
outlet gas manifold, up through each filter bed and out through the screens.
The function of the screens is to retain the filter media during the blow
back step, keeping it inside the filter beds, while allowing the fine par-
ticulates removed from the filter media by the blow back gas to pass through.
The fine particulate then settles outside the filter elements and is collected
at the bottom of the vessel containing the filter elements.
A pressure vessel 2.4 m (8 ft) in diameter by 3.4 m (11 ft) high was
installed on the miniplant to house up to four filter elements. Each element
is contained within a shroud inside the pressure vessel. Inlet gas is piped
to each shroud, passing through orifices which measure the flow rate to each
filter element. Clean gas exits each element and fills the interior of the
pressure vessel. Particulates removed from the filter elements during blow
back impinge on the inside surface of the shrouds, fall to the bottom and
are collected in lock hoppers. The pressure vessel and filter elements are
heated prior to the start of a run with a gas-fired burner to a temperature
greater than the dew point of the flue gas.
A number of high temperature runs were then attempted but the pressure
drops across the filter were extremely high and' all attempts at blow back
were unsuccessful. Inspection of the filter elements after each of these
runs showed that a hard filter cake had formed on the inlet retaining
screens. The filter medium was usually particulate free indicating very
little penetration through the screens. The initial tests were made before
the preheat burner was installed and it was thought that the plugging was
caused by condensation of moisture during start up. However, runs made
after the preheat burner was in operation were also terminated by screen
plugging problems. As a result, this filter design was deemed unacceptable
for this application.
Discussions with the Ducon Company led to the design and fabrication
of modified filter elements. Ducon suggested removing the inlet screens
and designing the individual filter beds with more freeboard to prevent
entrainment of the filter media during blow back. It was also suggested
that a gas distribution plate be used at the bottom of each bed to assure
good distribution of the blow back air.
-------
Operability of the modified filter system was demonstrated. The end
of the initial shakedown phase was signified by the successful completion of
a 24 hour demonstration run. This run was preceded by a number of shorter
duration runs used to establish suitable operating conditions for the demon-
stration run. These runs were successful in that filtration, ability to blow
back, ability to maintain low pressure drops and collection of particulates
after blow back were demonstrated. Collection efficiencies of 90 to 95% were
measured for the first few hours of the runs based on outlet particulate con-
centrations of about 100 mg/m3 (0.05 gr/SCF). Operation for up to 24 hours
was also demonstrated with no significant increase in base line pressure
drop across the filter. Blow back was usually required every 10-20 minutes
during which time the filter pressure drop had increased 14 kPa (2 psi)
above its base line value. A range of blow back conditions were used to
restore the base line pressure drop. Blow back durations ranged between 2
and 30 seconds and superficial velocity between 0.15 and 0.75 m/s (0.5 and
2.5 ft/s). The quantity of blow back air used ranged from 1 to 5% of the
filtered gas rate. Filtration velocities generally ranged between 20 and 24
m/min (60 and 80 ft/min). Filter media consisting of 300 to 600 micron
quartz particles were tested.
The particulates passing the filter had a mass median size of about 3
microns with about 10% larger than 10 microns.
A number of problems were defined during the shakedown and operation of
the filter. Demonstrated particulate outlet concentrations are still higher
than the tentative gas turbine inlet requirements. However, firm turbine
requirements have not been set as yet and it may be too early to reach any
conclusions regarding the suitability of the filter to protect a gas turbine.
The lower outlet particulate loadings of 100 mg/m3 (0.05 gr/SCF) meet the
current EPA emission standards. However, in all runs, it was observed that
the outlet loadings increased with time. It has not as yet been demonstrated
that the EPA emission standard can be met for more than a few hours of
operation.
Another observation has been the retention of a significant portion of
the filtered particulates in the filter beds. This has amounted to as much
as 30% of the weight of the filter media. The retained particulates were
also found to be uniformly distributed through the filter beds instead of
forming a layer on the top of the beds as had been expected. Distribution of
particulates throughout the bed was probably responsible for much of the
particulate penetration through the bed, and hence the observed increase of
outlet loadings with time.
Loss of filter media during blow back was another reoccuring problem.
Since inlet retaining screens were found to be susceptible to plugging,
better control of the blow back air rate must be established to minimize the
losses.
Another potential problem with the current design is its vulnerability
to upsets. If upsets occur, such as bed plugging or loss of filter media,
the operating problems caused by such upsets usually require shut down of
the system. It is usually not possible to take corrective action which
restores good operation. Another problem which may be unique to the mini-
plant was the interaction of the granular bed filter with the rest of the
-------
FBC system during blow back cycle. An increase in system pressure was noted
during blow back resulting in problems with the coal feed system which is
controlled by the differential pressure between the coal feed vessel and
combustor. Modifications to the coal feed control system were required to
minimize the effects.
COMPREHENSIVE ANALYSIS OF EMISSIONS
A program was begun to carry out a comprehensive analysis of all poten-
tially hazardous materials in the effluents from the miniplant unit. A series
of runs was completed in which samples of the input streams as well as the
effluents were taken and analyzed. This work was done in cooperation with
Battelle Columbus Laboratory, the contractor coordinating all such activities
for the EPA. The results of the test will be summarized in a report to be
issued jointly by Battelle and Exxon Research and Engineering Company.
MODIFICATION OF THE BENCH SCALE UNIT
The bench scale unit which was originally designed to operate in a batch
or semi-batch fashion was modified to permit continuous operation. The
modifications will permit the continuous feeding of coal and sorbent to the
combustor and the continuous withdrawal of spent sorbent. Solids captured in
the first stage cyclone will be able to be recycled to the combustor if
desired. The regenerator will also be capable of continuous feeding of
sulfated sorbent and withdrawal of regenerated sorbent.
ANALYSIS OF DESULFURIZATION DATA
Additional analysis of flue gas desulfurization data measured in the
miniplant and batch units was carried out. S02 retention results have been
correlated as function of the Ca/S molar ratio calculated from a sulfur
balance and based on measurements of sulfur contents in the flue gas and the
spent sorbent. This was done to minimize the effect of mechanical dif-
ficulties with the coal/sorbent blending equipment and an incomplete approach
to steady state conditions. The desulfurization data were again analyzed
using the Ca/S ratio set on the solids blending equipment but corrected for
variations in combustor temperature and gas phase residence time. Residence
time corrections were based on the use of a first order rate expression which
had been developed previously and more recently verified. Temperature cor-
rections were based on a calculated activation energy of 13 kcal/mole.
Correcting for residence time and temperature variations gave a good correla-
tion of S02 retention with Ca/S ratio set on the blender. The correlation
agreed reasonably well with that based on the Ca/S ratio calculated from
analysis of gas and solids.
S02 retention results measured in the batch unit and the continuous
miniplant were also compared after correcting for the effect of solids
residence time distribution in the miniplant. Reaction rate constants were
calculated from the batch unit results and were used to calculate the S02
retention as a function of Ca/S ratio for a continuous unit. It was assumed
that the gas moved in a plug flow fashion through the miniplant combustor,
that the solids were well mixed and the residence time distribution of the
solids could be described by an exponential distribution function. The
-------
corrected batch results were then compared to results measured in the mini-
plant. The corrected batch data did not exactly correspond to the miniplant
data, but were still quite representative of desulfurization performance with
dolomite sorbents.
Residence time distribution of solids in the combustor can also be used
to gain some insight into the factors which control desulfurization perfor-
mance. It was estimated that 15% of the sorbent particles in the combustor
at typical operating conditions have a residence time between 0 and 1 hour but
because of the higher activity of fresh sorbent, these particles account
for 55% of the SC>2 retention achieved in the combustor. Also, 85% of the S02
removal occurs on particles which have been in the combustor for 3 hours or
less, even though these particles represent only 31% of the combustor
inventory. It is this high level of effectiveness of fresh sorbent particles
which results in a fairly rapid approach to a steady state SC>2 concentration
after a change in the Ca/S feed ratio.
CONTINUOUS STUDIES
The particulate removal program will continue with the objective of
optimizing the performance of the granular bed filter. Particulate removal
efficiency must be improved, loss of filter media during blow back decreased
and the blow back made more effective. The use of transparent models to
observe the mechanics of the filtration and blow back steps will begin in an
attempt to learn more of what is occuring in the filter. Long term testing
with the filter will begin in conjunction with the Department of Energy (DOE)
sponsored program to evaluate gas turbine materials. The DOE hot corrosion
testing on the miniplant will be conducted under a cooperative agreement
between EPA and DOE.
Following the granular bed filter tests, one or possibly two alternate
particulate control devices will be evaluated. The choice of the devices
will be made after consultation with the EPA. Currently, devices such as a
high temperature electrostatic precipitator or a high temperature bag filter
are being considered.
The comprehensive analysis program will be extended to include operations
with the regenerator. More quantitative tests (Level 2 and Level 3 tests)
will be carried out based on the results of the initial screening tests
(Level 1 tests). Two high temperature, high pressure particulate sampling
systems capable of sample particulates at temperatures up to 870°C will be
fabricated and installed.
The regeneration program will consist in a series of runs of two to five
days duration aimed at generating sufficient data to evaluate the economic
feasibility of regeneration. The tests will determine the make up sorbent
requirements and will indicate the level of S02 in the regenerator off gas
which can be obtained.
-------
Future work in the bench scale combustor unit will consist in evaluating
various coals and sorbents. NOX control studies will also be made. Regenera-
tion studies will determine the activity maintenance of sorbent recycled
between the combustor and regenerator while operating under various regenera-
tion conditions. The use of coal as fuel for the regenerator will be
included.
-------
10
-------
SECTION II
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 temperature can be maintained at low and uniform temperatures in
the vicinity of 800 to 950°C. The lower temperatures allow the use of lower
grade coals (since these temperatures are lower than ash slagging tempera-
tures) , 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. Operation at the higher pressure
also results in a further decrease in NOX emissions.
In the fluidized bed boiler, limestone or dolomite is calcined and
reacts with S02 and oxygen in the flue gas to form CaSO^ 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).
The goal is to produce S02 in the regenerator off gas at a sufficiently high
concentration to be recovered in a by-product sulfur plant.
CO C02
CaS04 + H2 -> CaO + S02 + H20 (2)
A diagram of the pressurized fluidized bed combustion and regeneration
process is shown in Figure II-l.
Exxon Research and Engineering Company, under contract to the EPA, has
built two pressurized fluidized bed combustion units to study the combustion
and regeneration processes. The smaller of the two units, the batch unit,
was built under contract CPA 70-19 and was described in previous reports
(1,8,9). Those reports also described regeneration and combustion studies
carried on in the batch unit. The batch unit is being converted to a
continuous bench scale unit and the modifications required for the conversion
are described in this report. No experimental work has been done in the
batch/bench unit during this current reporting period.
11
-------
FIGURE ll-l
PRESSURIZED FLUIDIZED BED
COAL COMBUSTION SYSTEM
Gas Turbine
Air Compressor
High Efficiency
Separator
Cyclone
Separators
Solids
Transfer
System
Coal and
Sorbent Makeup
To Sulfur
—*" Recovery
Separator
Combustor
Regenerator
Fuel
-------
The larger unit, called the miniplant, was designed under EPA Contract
CPA 70-19 and built under Contract 68-02-0617. Figure II-2 shows a photo-
graph of the miniplant. The shakedown and operation of the unit was funded
under Contract 68-02-1312. Previous reports (1,7,9) described design,
shakedown and operation of the unit. This report includes additional
results from the operation of the combustion section of the miniplant. The
effect of operating conditions on 862, NOX and particulate emissions, and com-
bustion efficiency was measured. Various coals and sorbents were tested,
including precalcined limestone sorbent. The first test in a series of tests
aimed at developing a comprehensive analysis of all emissions especially trace
emissions, from the miniplant pressurized FBC unit was completed. The design
and shakedown of a granular bed filter intended to remove particulates from
the flue gas to very low levels is also described in this report.
This report also describes the operation of the regenerator section of
the miniplant. The previous report (1) described the regenerator and the
shakedown tests. This report describes a 125 hour run in which the combustor
and regenerator were both operated continuously and the operability of the
system was demonstrated.
The period of performance discussed in this report is August 1, 1976 to
August 12, 1977.
13
-------
FIGURE I1-2
EXXON FLUIDIZED BED COMBUSTION MINIPLANT
14
-------
SECTION III
COMBUSTION STUDIES
Combustion studies have been carried out in the EPA/Exxon pressurized
fluidized bed combustor referred to as the miniplant. The miniplant has
provisions for continuous addition of coal and sorbent and continuous
withdrawal of sulfated sorbent. The miniplant is shown schematically in
Figure III-l. As of July 1977, the combustor has been operated for a total
of approximately 1500 hours in a series of individual runs up to 240 hours
duration. This section of the report describes the combustor equipment,
operating procedures, combustor performance and combustion results. A
discussion of the regeneration work is given in Section IV.
EQUIPMENT, MATERIALS, PROCEDURES
This section will focus on the major system components which include:
1) solids feeding system, 2) combustor with internal subcomponents, 3) com-
bustor cyclones, 4) pressure control and flue gas discharge system, 5) flue
gas sampling and analytical system, 6) process monitoring and data genera-
tion system, 7) combustor safety and alarm system, 8) coal and sorbent
properties, 9) operating procedures, and 10) analytical procedures. A
detailed description of each of these systems can be found in an earlier
report and only a brief discussion will be included here (1,9).
Solids Feeding System
Figure III-2 displays a schematic of the miniplant coal and sorbent
feeding system. Crushed and sized coal and limestone or dolomite are held
in separate storage bins (20 tonnes for coal and 3 tonnes for sorbent) under
atmospheric conditions. On demand, the solids from the bins are proportioned
to a specific coal/sorbent ratio. Inverters control the motor speeds of
separate coal and sorbent screw feeders and volumetrically control the coal/
sorbent ratio. A blending screw transports mixture into a solids feed ves-
sel. The coal/sorbent mixture is held in this vessel until refill of the
injector vessel is required.
The solids feeding system provides for continuous solids delivery
(coal and sorbent) from the injector vessel to the pressurized pressurized
combustor, while allowing intermittent refilling of the injector vessel
(193 kg operating capacity). Load cells located under the injector vessel
monitor the solids feed rate and actuate control signals for the refill
cycle. Prior to initiation of a refilling operation, the injector vessel,
feed vessel, and the pair of solids storage bins remain isolated from
each other. When the load cell under the injector vessel detects a solids
loading of less than 102 kg, 91 kg of solids are automatically transferred
pneumatically from the feed vessel to the pressurized injector vessel without
interrupting feed to the combustor. Refilling is usually completed in about
5 minutes. After refilling, the feed vessel is again isolated from the injector
vessel, vented, and filled with solids from the storage bins. The feed vessel
is then isolated and repressurized to await another cycle.
15
-------
FIGURE Ill-l
EXXON FLUIDIZED BED COMBUSTION MINIPLANT
Cooling
Water
To Scrubber
Air (Pressure Control)
Granular Bed Filter
Cooling Water
Out In
I f
To
Scrubber
Cyclone
Separator
.
Isolids
f Discharge
Solids
Reject
Vessels
Feed
Water
Reservoir
Coal and
Limestone
Feed Supply
Natural Gas
Compressor
Auxiliary
Air
Compressor
Main Air
Compressor
Liquid Fuel Storage
-------
FIGURE 111-2
COAL & LIMESTONE FEED SYSTEM
Limestone Bin
1/2 S.S. Pipe
-------
Solids in the injector vessel are continuously aerated by the pressuriz-
ing air stream, which is automatically controlled at a pressure level slightly
above that in the combustor. The coal/sorbent mixture is discharged from
the injector vessel through a 1.3 cm diameter orifice and pneumatically
conveyed by a stream of dried transport air through an s-shaped 1/2-inch
stainless steel pipe to the combustor. A short section of high pressure
rubber hose is used to connect the discharge orifice to the injection line
to avoid interference with the load cell operation.
Final entry of solids into the combustor is through a 1.3 cm I.D. nozzle
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 penetra-
tion of the solids feed into the fluidized bed and to protect the feed nozzle
from blockage with bed solids. The flow of solids into the combustor is
controlled to maintain constant temperature in the combustor.
Combustor
The combustor consists of a 61 cm I.D. steel shell refractory lined to
an inside diameter of 33 cm. The 9.75 m high unit is fabricated in flanged
sections to allow insertion and removal of the cooling coils. Various ports
are strategically located to allow for material entry and discharge. Numerous
taps are also provided for monitoring both pressure and temperature. A
schematic of the combustor is shown in Figure III-3.
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 m^ and consists of vertically-oriented loops constructed of 1/2-inch
Schedule 40 316 stainless steel pipe. The number of coils can vary from one
to ten depending on the combustor operating conditions and the amount of
cooling required. A high pressure pump is used to pump the cooling water
through a closed-loop arrangement consisting of a demineralized feed water
reservoir, cooling coils, and a heat exchanger. The flow rate and exit
temperature from each coil can be separately controlled and monitored.
The baffled cooling coils were inspected after Run 65 and found to be
in excellent condition. These coils were installed prior to Run 29 and
logged more than 600 hours of running time. There was no sign of the erosion
seen on earlier unbaffled coils. Even the baffles were not eroded as seen
in Figure III-4.
The combustion air to the unit is provided by a main air compressor
having a capacity of 40 Sm3/min at 1030 kPa (1400 SCFM at 150 psig). Preheat
of the combustor during start-up is made possible by a natural gas burner
which is housed in the bottom plenum section of the combustor. Once the
fluidized bed temperature reaches approximately 430°C a liquid fuel system
is used to heat the bed further to the coal ignition temperature.
The liquid fuel used to heat up the combustor is "Varsol" or kerosene.
The fuel is pumped, under pressure slightly higher than combustor pressure,
from either of two 218 1 (55 gal) drums through a rotameter to an injector'
nozzle located at port #5, 15.2 cm above the combustor grid (Figure III-5).
18
-------
FIGURE III-3
COMBUSTOR VESSEL
-------
FIGURE II1-4
COOLING COILS AFTER 600 HOUR EXPOSURE
20
-------
FIGURE III-5
LIQUID FUEL SYSTEM
AtfX. A/ff J-VP
XT ffTJfOCAfg
COMBUSTOR FUJL PffOBE
~™" FOR DIFAIL RfFFR To DRAW If*.
I .."-«--D
1 ^- SHfo fitsr MIU
JLJL *—i'sasi smut
BAii
LtQUlO FUfL
race M/Nt PLANT
-------
The injector nozzle atomizes the fuel to insure complete combustion (Figure
III-6). The fuel flow rate must be manually set at the rotameter, however,
a solenoid valve is used for on/off control from the control panel. The
liquid fuel system is also activated during coal feed problems to maintain
combustor bed temperature above the coal ignition temperature.
Coal and sorbent are injected pneumatically into the combustor through
a single side-entering port 28 cm above the fluidizing grid. The maximum
design solids feed rate is approximately 227 kg/hr. The combustor is capable
of operating at pressures up to 1000 kPa and temperatures to 1100°C, super-
ficial velocities up to 3 m/s with expanded beds to 6.1 m. The expanded bed
height can be controlled at any level above 2.3 m by the continuous with-
drawal of bed solids through a port located 2.3 m above the fluidizing grid.
Solids flow by gravity through a refractory lined pipe into a "pulse pot"
from where they are pneumatically transported by controlled nitrogen pulses
to a pressurized lock hopper.
The combustor fluidizing grid consists of 332 3/32 inch holes and four
independent cooling water loops consisting of five channels each (see Figure
III-7). The previous grid, which was installed through Run No. 49, had
only one cooling water loop with twenty channels in the grid. This grid
suffered from uneven grid temperature distributions due to preferential water
flow. Some erosion was seen on the old grid, however, it did not fail. The
current grid cooling scheme results in even grid temperatures and should
result in longer grid life.
Combustion gases exit the combustor and go to a two stage cyclone system.
The primary intent of the first cyclone is to recirculate larger unburned
carbon particles back to the combustor to improve combustion efficiency.
Particulates collected in the second stage cyclone are dropped into a lock
hopper and disposed of on a batch basis. Due to refractory deterioration
and severe pitting, the second stage cyclone was recast with a more resistant
refractory (Resco RS-17-E). The barrel was also made smaller to correspond
to the smaller inlet and outlet pipes that were installed last year. The
particulate collection efficiency was not changed significantly due to this
modification.
Flue gas is sampled at a point about 7 m downstream of the second stage
cyclone. The system is designed to produce a solids-free, dry stream of flue
gas at approximately ambient temperature and atmospheric pressure whose com-
position, except for moisture, is essentially unaltered from that of the
original flue gas. The system was described in the previous report (1).
Particulates are also sampled near this point. The particulate sampling sys-
tem is discussed in detail on page 35.
New Combustor Sections
Two new sections fabricated for a fireside corrosion program sponsored
by the Department of Energy (DOE) and conducted in cooperation with Westinghouse
Research Laboratory have been installed in the combustor. The DDE-funded
fireside corrosion work will be conducted on the miniplant under a coopera-
tive agreement between EPA and DOE. The orientation of these sections is
shown schematically in Figure III-8. The lower corrosion test section
(Figures III-9 and 111-10) measures 0.91 m in height and is located 0.91 m
22
-------
FIGURE III-6
LIQUID FUEL INJECTOR NOZZLE
F6CC MINI - P
1647-2-Z-D
-------
FIGURE III-7
COMBUSTOR FLUIDIZING GRID
1U-
gflcc MUSI-PLANT
-------
FIGURE 111-8
COMBUSTOR CORROSION PROBE LOCATION
2m
1.5m
0.9m
0.9m
0.9m
0.9m
0.9m
0.9m
I
1
1
1
m
1
m
1
m
1
m
Upper
Corrosion
Probe
Section
(Detail 'A')
r
Lower Corrosion
n
1
Tl
Probe Section
(Detail 'B')
I
1
I
1
1
1
I
Combustor
Ports
Detail 'A1
Combustor
Ports
Detail 'B1
Fluidizing Grid
25
-------
FIGURE III-9
COMBUSTOR LOWER CORROSION PROBE SECTION
N>
ON
ggiVuir mbSSF.
-------
FIGURE 111-10
COMBUSTOR LOWER CORROSION PROBE AND BED SAMPLING PROBE
-------
above the fluidizing grid. The upper corrosion test section (Figure III-ll)
measures 1.5 m in height and is located 5.5 m above the fluidizing grid. Both
sections are refractory lined to an internal diameter of 0.32 m to conform
to the dimensions of the combustor.
These two sections will provide a test site and environment for exposure
of 21 specimens of heat exchanger materials. Twelve of the specimen probes
will be located in the lower section within the expanded bed while the
remaining nine will be located in the upper section in the freeboard region.
The probes (Figure 111-12) will be air cooled and temperature controlled to
requirements specified by Westinghouse.
Temperature Control
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 combustor temperature, pressure
differential between the primary injector and combustor, the injector pres-
sure, and the transport air flow rate (Figure 1II-2).
A thermocouple in the lower zone of the combustor, 46 cm above the
fluidizing grid, is used as the sensor for the control of the combustor
temperature. The coal feed rate to the combustor is regulated by the pressure
differential between the injector vessel and the combustor - the greater
the pressure difference, the greater the feed rate. Temperature control
is accomplished by a cascade type control loop using two controllers, one for
temperature control and another for AP control. A deviation of the desired
temperature from the actual combustor temperature causes a signal to be trans-
mitted by the temperature controller to the AP coal feed rate controller.
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 combustor
causes a change in the coal feed rate which will tend to return the bed tem-
perature to the desired value. Proper tuning of the controllers is 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.
Pressure Control
The FBC Miniplant Combustor has the capability of operating at pressure
levels of up to 10.5 atmospheres (140 psig). Pressure control is achieved
by restricting the discharge flow of gas from the combustor, so as to achieve
an increase in back pressure. This is done by use of a converging nozzle
inserted in the discharge line. Adjustment of the combustor pressure is
accomplished by metering high pressure air into the discharge line just
upstream of the flow nozzle. A 2 inch-ball valve equipped with a pneumatic
positioner and actuator regulates the amount of air added in response to a
signal from the pressure controller.
The converging nozzle is machined from a silicon carbide section which
is inserted into an 8 inch carbon steel blind flange. Four nozzle inserts
are currently on hand, with throat diameters of 2.41, 2.67, 2.79 and 2.91
28
-------
FIGURE III-ll
COMBUSTOR UPPER CORROSION PROBE SECTION
-------
FIGURE 111-12
COMBUSTOR CORROSION PROBE
30
-------
centimeters. By proper nozzle preselection, a wide range of combustor flow
conditions can be handled. Control of combustor pressure has always been
very good, with variations from the set pressure usually less than 0.1
atmospheres.
Turbine Specimen Erosion Section
A turbine specimen/erosion section was fabricated for EPA by Westinghouse
Research Laboratory and inserted in the flue gas line downstream of the
secondary cyclone. This includes a nozzle which enables high velocities to
be reached at the specimen location with relatively low combustor air flow
rates. The section measures 45.7 cm long and has an entrance diameter of
8.9 cm which converges to 2.9 cm at the specimen position. Some of the fac-
tors affecting erosion were discussed by Westinghouse in a report to the EPA
(3).
During Miniplant Run 52 an erosion test rod supplied by Westinghouse
was exposed to superficial velocities of 259 m/sec (850 ft/sec). The base
material of the rod was X-45 and was coated with Co-Cr-Al-Y. The temperature
at the section was 760°C and the grain loading was measured to be 2.4 gm/m3
(1.1 gr/SCF). Inspection of the rod following 32 hours of exposure showed
considerable erosion (Figure 111-13). The rod has been returned to Westing-
house for analysis.
Sampling and Analytical Systems
Bed Sampling Probe—
A solids sampling probe has been designed, constructed and installed to
enable acquisition of samples from the combustor bed while a run is in
progress (Figure 111-10). The need for such a sampling probe was evident
during the regenerator demonstration run. In that run, which lasted 125
hours, bed solids did not have to be rejected to maintain bed height, and
accordingly no information could be obtained on the composition of the com-
bustor bed. A cross-section of the probe is shown in Figure 111-14. The
probe consists of a closed-end tube which contains a hole on the sidewall
near the top of the tube. In normal operation the top end of the probe is
recessed into the refractory wall of the combustor vessel. In this position,
the bed particles form a seal around the top of the probe preventing par-
ticles from reaching the hole in the sidewall. When it is desired to take a
sample, a pneumatic piston pushes the probe forward into the combustor bed,
exposing the hole on the upper sidewall. Bed particles rush into this hole
in the sidewall and fill the probe cavity. The probe is pneumatically pulled
back into the recess in the combustor refractory wall and the captured sample
of bed particles is removed from the lower end of the probe.
In Run No. 51, bed samples were taken using the sampling probe at one-
hour intervals. The utilization of the bed solids determined from analysis
of the particles obtained from the probe samples is shown in Figure 111-15.
The sorbent utilization decreased with time. This is consistent with an
approach toward the expected steady state value of 50 percent sorbent
utilization expected for the calcium to sulfur molar feed ratio of 1.67 for
this run. The sorbent utilization determined from a bed solid reject sample
taken at 5.8 hours into the run is included for comparison in the Figure
111-15 and shows good agreement with the results obtained using the sampling
31
-------
FIGURE 111-13
WESTINGHOUSE EROSION PROBE
32
-------
FIGURE 111-14
COMBUSTOR BED SAMPLING PROBE
CO
U)
-------
FIGURE 111-15
SORBENT UTILIZATION VS. TIME
RUN No. 51
0.80
1
T
T
BED SAMPLING PROBE SAMPLES
BED REJECT LINE SAMPLE
0.70
u>
0.60
0.50
0.40
I
I
I
345
TIME INTO RUN (HOURS)
6
7
-------
probe. One of the values determined for the sorbent utilization during this
period was well outside the range of the other values. This points out the
need for taking multiple samples to avoid erroneous results due to non-
representative samples.
Gas Sampling System—
The flue gas sampling system is installed at a point downstream of the
secondary cyclone. It was described in the previous report (1). The system
is designed to produce a solids-free, dry stream of flue gas at ambient
temperature and low pressure whose composition except for moisture, is
essentially unaltered from that of the original flue gas. This system pro-
vides a gas sample which is analyzed by the continuous, on-line, gas analyzers.
Another flue gas sample can be extracted which has been filtered, cooled and
depressured, but not dried. This system is used to obtain batch samples of
flue gas for analysis by wet chemistry methods.
Particulate Sampling System—
The particulate sampling system initially installed on the miniplant
is shown in Figure 111-16. A 1.09 cm probe was used, and as this system
was initially intended to be used at about 275°C, a three stage water cooled
heat exchanger was provided. Total particulate sampling was done with
Balston Type 30/25 and 30/12 filters containing high temperature resistant
(~530°C) BH cartridges. The gas passing the filter was cooled to 200°C and
passed through another filter which removed condensed matter. Size distribu-
tions were to be made with a Brink impactor at 175°C.
This system was intended to obtain samples at temperatures high enough
to prevent condensation of moisture. Some operating problems with this system
became apparent and modifications were made. The section containing the
impactor was removed after run 50 due to problems of gas leakage from the
impactor. It was also decided to increase the operating temperature of the
filters to the maximum to avoid condensation of substances other than moisture,
which could affect the particulate loading, particle size distribution and
composition. The modifications made to the sampling system for the above
and other reasons are described below.
• Before Run 50, a second 30/25 (high temperature) filter was
placed immediately after the first filter. The second filter
was used as a back-up filter in case the cartridge material
of the first filter failed or particle by-pass occurred due
to excessive pressure drop across the first filter.
• In order to increase the operating temperature of the filter
for Run 52, the three stage water cooled heat exchanger was
replaced with a five foot length of 1.27 cm OD ss pipe.
• At the end of Run 54, the system was again modified.
The changes corresponded to (a) the use of a smaller probe
diameter (0.77 cm) to decrease the pressure drop across the
filters by reducing the flow rate at isokinetic velocities,
and (b) the ability to place the first and second filters
next to the high temperature valve, if required, so that
the filters may operate at a high temperature (maximum^
allowable temperature for the filter cartridges is 530°C).
35
-------
FLUE GAS FROM
SECONDARY CYCLONE
u>
1
e —
HIGH
TEMPERAT
VALVE
. [N^-l— -
-i ^ r
— X 4
V V
I !
PURGE
NITROGE
PRESSURE PRESSURE
GAUGE GAUGE
JRE Tc9 9 TC ^
~i r 1 i 1 , _ -, r 1
LlLji U p,L_^ U
i L i i ' n n\A/
H
V
xj
HEAT COLLECTION STEAM pi^Rm VAI VF
1EXCHANGER FILTER MEAT CONTROL VALVE
EXCHANGER
N
ROTAMETER
FIGURE III - 16
MINIPLANT PARTICULATE SAMPLING SYSTEM
-------
Future work on particulate sampling will entail the current equipment
discussed above, as modified after Run 54, and two other more sophisticated
systems. One of these systems will be placed before the granular bed filter
and will contain Southern Research Institute's specially designed cyclone
set; the other system will be placed after the granular bed filter and will
contain an impactor and/or a total filter. Both these systems have been
designed to perform at 870°C and 909 kPa, and automatic flow and temperature
control devices have been incorporated.
Samples of flue gas will be extracted under isokinetic or constant flow
conditions and the temperature of the sample will be maintained by line
heaters. The sampling device (cyclone or impactor) will be enclosed in a
pressurized furnace in which the temperature is held at the flue gas tempera-
ture (up to 870°C) . The pressure inside the furnace will be the same as the
pressure inside the sampling device so the device will not have to be built
to withstand a high pressure. This will minimize structural, sealing and
leakage problems. Each system will be self contained and can be moved to
different locations on the miniplant. Although the systems will be built to
operate up to 870°C, the effect of sampling temperature on particulate
loading, size and composition will be measured and the most suitable operating
temperature will be determined.
In this report, results on particulate sampling will be given for all
runs up to Run 59, in which sampling experiments were carried out.
Acurex Sampling Section—
Several miniplant discharge line sections were modified in December 1976
to allow insertion of a new high temperature-high pressure (HTHP) particulate
sampling probe developed by the Aerotherm Division of the Acurex Company.
The probe was developed under EPA contract and is described in a recent
report (17).
Exxon was requested to build a sampling section to allow a test of the
sampling system under high temperature, high pressure conditions. The probe
sampling section fabricated by 'ER&E has a 10 inch stainless steel liner
inside of a 16-inch refractory lined pipe section (Figure 111-17). Eccentric
reducers at each end of the sampling section were necessary to adapt to the
existing discharge line piping and also avoid extensive modification of the
miniplant structure.
The initial miniplant field test of the Acurex HTHP sampling probe was
made during the four day Comprehensive Analysis test series (run #50) March
29 to April 1, 1977. Additional tests with the Acurex HTHP probe were con-
ducted during the week of May 23, 1977. A photograph of the Acurex HTHP
sampling probe installation at the miniplant is shown in Figure 111-18.
This photograph also shows the location of the Westinghouse turbine test
section.
Process Monitoring and Data Generation System
Data characterizing the miniplant operating conditions are recorded on
5 multipoint recorders. Three recorders (24-channel Honeywell Electronik
112) monitor output from various measuring instruments - thermocouples, dif-
ferential pressure transmitters, etc. Two recorders continuously monitor the
37
-------
FIGURE 111-17
COMBUSTOR - ECCENTRIC ACUREX PROBE SECTION
U)
CO
sn g*f#- gtt-tst-c
JS'TOB"EccfNTKic /fcovcfR .(^)
5TANDARD WALL 57L.
fO TO V " tCCEHTRIC fttDVCfR (?)
30V3.S. SCC 0FT4/LS BtiOW
ft) a'f/fC J-50 LB. FiANGCS
73 MATCH UP WITH
SCHP.J' 10"
S.5. PIPE
JO MAKH t/P H/ITU
SCUD J V
s. s. pipe
I I
MINI-PLANT
£CCfHTKIC ACURfX PROBf
SfCTldN
EXXON HOCMCH *MO CHCIMCCmiW
MECHANICAL OIVIWM
I.INOKN, M J.
"Thrums
-------
FIGURE 111-18
ACUREX PROBE
to
-------
output signals from the gas analyzers. 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.
Approximately 100 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 F.
Signals from the data logger are scanned every minute and appear as
digital output on printed paper tape and are also stored on magnetic tape.
The magnetic tape, containing about 6000 items of data per hour of run time,
is fed to a computer which converts the logger output to flow rates, pres-
sures, 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
temperature, based on four thermocouple readings covering the 15-114 cm
interval above the fluidizing grid, superficial gas velocity, excess air, as
well as the important gas concentrations.
Combustor Safety and Alarm System
A process alarm system was designed to warn of impending operational
problems. Two general alarm categories exist. The first, dealing with less
critical situations, alerts the operator of the problem so that appropriate
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. An alarm condition is brought to the attention of the
operators by a flashing light above the control panel accompanied by a high
pitch sound. The sensitivity of the individual alarms is controlled by
potentiometers located beneath the control panel.
Appendix G gives a brief description of the alarm triggering
condition, the mechanical sensoring device, and the follow-up actions taken
by the system.
Coal and Sorbent Properties
Coal—
Coals used in the miniplant variables study were a high volatile
bituminous coal from the Consolidation Coal Company's "Champion" preparation
plant in Pennsylvania and Illinois No. 6 seam coal obtained from Carter Oil
Company's Monterey No. 1 mine. The Champion coal was partly classified to
remove fines smaller than 40 U.S. Mesh. The Illinois coal was screened to
6 X 25 U.S. Mesh to prevent plugging of the primary injector feed vessel.
Particulate size distribution and composition data for both coals are shown
in Figure 111-19 and Table III-l respectively. During and after Run 50, the
Champion coal sulfur content changed from the average of 2+0.1% to between
1 and 1.6% due to dilution with a 0.6% sulfur Kentucky coal at the preparation
plant. This was not detected until after the runs had begun.
Sorbent
Sorbent—
Grove limestone (BCR No. 1359) and Pfizer dolomite (BCR No. 1337) were
the primary sorbents used in the miniplant variables study. The composition
40
-------
CO
UJ
O
h-
o:
o_
s
<
CO
co
:r
CD
LU
UJ
>
O
lOOh
90
80
70
60
50
40
30
20
10
FIGURE 111-19
COAL PARTICLE SIZE DISTRIBUTION
Illinois Coal
Champion Coal
I
0 200 400 600 800 1200
Mill I I I I I
1600
1800
I
2400
2800
3200 Microns
200 50 40 30 20 18 16 14
12
10
8
6 Mesh
PARTICLE SIZE
-------
TABLE III-l. COAL COMPOSITION
Run No.
41.1
41.2
43.1-43.5
45.1-48
50.1A-50.3
50.4-52
54-59
Coal Type
Champion
Arkwrlght
Illinois
Champion
Champion
Champion
Champion
Ultimate Analysis
Moisture
2.83
0.88
3.01
2.83
1.67
1.32
2.19
Ash
8.03
7.42
9.95
8.03
13.10
11.90
12.50
Total
Carbon
70.38
77.11
67.71
70.38
74.00
73.40
71.35
H
5.01
5.09
4.77
5.01
5.00
5.00
4.74
S
1.96
2.47
4.20
1.96
1.85
1.66
1.40
N
1.36
1.07
1.17
1.36
1.50
1.40
1.54
Cl
0.09
0.14
0.05
0.09
—
—
0.07
°2
10.43
5.96
9.19
10.43
2.40
4.20
6.31
Heating
Value
BTU/lb
13,346
13,699
12,254
13,346
12,973
13,268
12,514
(1) Diluted with 0.6% sulfur Kentucky coal.
-------
of these stones is given in Table III-2. Most of the runs were made with the
stone screened to give the distribution of limestone and dolomite shown in
Figure 111-20.
Operating Procedures
Prior to initiating a run, a detailed checkout procedure is followed to
insure that the system is ready for operation. This includes various equip-
ment checks, alarm system checks, calibration of flue gas analyzers, activa-
tion of process monitoring and control systems, and the turning on of all
cooling water systems. All runs are begun with an initial bed of sorbent
in the combustor. This consists of either a fresh charge of uncalcined
limestone or the bed from the previous run.
The first operation of start-up involves heating the gently fluidized
sorbent bed by burning natural gas in the burner plenum followed by injection
of kerosene into the bed. Prior to ignition of natural gas, an air flow
of about 9.9 Sm^/min (350 SCFM) or about half that used at normal opera-
ting conditions is fed through the burner while combustor pressure is raised
to 280 kPa gauge. Once ignition begins, this procedure maximizes incoming
gas temperature under conditions which allow good natural gas combustion and
adequate bed fluidization. Water flow rates through the combustor cooling
coils are kept low to reduce heat loss to the coils. Ignition begins by
simultaneously feeding 0.57 Sm3/min (20 SCFM) of natural gas through the
burner while activating an ignition electrode.
Because of the limited capacity of the gas compressor, natural gas burned
in the plenum 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 tempera-
ture 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 1.7 Sm^/min (60 SCFM) 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 stopped. 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 constant by the automatic coal feed rate
control system.
43
-------
ISI
- —
"
LU
O
LJ <
> Q-
O
cn
c/i
LJ
100
90
80
70
60
50
40
30
20
10
0
I
FIGURE 111-20
SORBEKlT PARTICLE SIZE DISTRIBUTION
O Grove Limestone
Q Tymochtee Dolomite
Mesh
3^ 30 :
25 20
18
I
Microns
500
1000
PARTICLE SIZE
1500
2000
2400
-------
TABLE III-2. SORBENT COMPOSITION
Run
Number
41.1-41.2
43.1-43.5
45.1-47
48-51
52
56-59
Quarry
Grove
Pfizer
Grove
Pfizer
Pfizer
Pfizer
Sorbent
Type
Limestone
Dolomite
Limestone
Dolomite
Dolomite
Dolomite
Weight Percent
CaO
97.0
54.0
97.0
54.0
57.8
57.8
MgO
1.2
44.0
1.2
44.0
40.7
40.7
Si02
1.1
0.9
1.1
0.9
0.25
0.25
A1203
0.3
0.2
0.3
0.2
1.9
1.9
Fe2°3
0.2
0.3
0.2
0.3
0.133
0.133
-------
Analytical Procedures
Particulate Sampling
The equipment for particulate sampling has been described on page
35. The key to sampling consists of accurately weighing the filter ele-
ments and cartridges before and after the experiment and sustaining iso-
kinetic velocity throughout the experiment. Before starting, air is
admitted to the probe to prevent it from plugging. This air is shut off
after the filters are in place, and the high temperature valve is opened.
Pressure and temperature readings are taken at regular intervals, and the
rotameter is constantly and carefully checked and/or adjusted for isokinetic
sampling. At the end of the sampling period, the high temperature valve is
shut off, and air is admitted into the sampling system and into the probe.
After the total weight of the particulates is obtained, the solids are
wet sieved to obtain the + 45 ym fractions. The -45 jam fraction, which usually
consists of 80%+ of the total, is then analyzed by a Coulter Counter to
obtain the particle distribution.
EXPERIMENTAL RESULTS AND DISCUSSION
The combustion experimental program was directed primarily toward under-
standing the factors which control SC>2 emissions from the pressurized combus-
tor. However, information on NOX emissions, particulate emissions, and carbon
combustion efficiency were also obtained. In addition, the first test in a
series of tests was completed designed to make a comprehensive analysis of
all emissions, especially trace emissions of potentially harmful materials.
Other information such as temperature profiles in the combustor and cyclone
efficiency data was also generated in the test program.
This section reports and discusses the results of the combustion program.
S02 Retention
The objectives of the S02 retention program were to study the effect of
coal type and the use of precalcined limestone sorbent on -SOo retention.
The effect of other variables was studied and reported previously (1).
Effect of Coal Type—
The effect of coal type and sulfur content was determined in a series of
runs made with an Illinois No. 6 coal containing 4.2% sulfur and Pfizer No.
1337 dolomite sorbent. Coal and sorbent properties are given in Tables III-l,
III-2. Details of the runs (43.2 to 43.5) are given in Appendix H-l. The Ca/S
molar ratio, i.e., moles of calcium fed in the dolomite to moles of sulfur
fed in the coal, was varied from 0.7 to 1.6. S02 retention results using the
Illinois coal are shown in Figure 111-21 as a function of the Ca/S ratio
and are compared to data reported previously using a Pittsburgh Seam coal
(Champion) containing 2% sulfur (1). As seen in Figure 111-21, the reten-
tions measured with the Illinois coal fit reasonably well with the data and
the correlating line established for the Pittsburgh coal. This suggests
that the dolomite sulfation reaction is first order in S02 concentrations
and is consistent with results reported previously by others (2,3,4).
46
-------
FIGURE 111-21
S02 RETENTION WITH DOLOMITE SORBENT
100
CM
o
CO
• •••
• Pittsburgh Seam Coal
A Illinois Coal
Sorbent: Pfizer (No. 1377) Dolomite
1 2
Ca/S (MOLE/MOLE)
47
-------
Effect of Limestone Precalcination—
Previous pressurized FBC studies with limestone sorbent indicated that
the limestone desulfurization activity was determined in part, by the extent
of calcination occurring in the combustor (1). Higher combustor tempera-
tures promoted calcination and increased limestone activity. However, even
at the higher combustor temperatures, limestone was less active than dolomite
at an equivalent Ca/S molar ratio. It was also found that limestone desul-
furization activity approached zero under low temperature "turndown" operating
conditions due to the inability of the limestone to calcine at those condi-
tions. Since the level of calcination figured so strongly in determining the
activity of the limestone, a series of runs were made in which the limestone
was precalcined and fed to the combustor in the calcined form along with the
coal. The runs were made with Champion coal and Grove No. 1359 limestone.
Coal and limestone properties are given in Tables III-l and III-2. The
limestone was calcined in the miniplant combustor at a temperature of about
870°C at 940 kPa pressure. The natural gas fired preheat burner was used
during calcination to maintain the temperature at 870°C. A single cooling
coil of reduced size was used to remove excess heat and control the tempera-
ture at the proper level. The CC>2 partial pressure in the combustor during
calcination was about 109 kPa. The calcined limestone was removed from the
combustor, stored in steel drums and then added, by hand, to the coal/limestone
blender located in front of the coal/stone injector vessel. The detailed
run conditions and results (runs 46.1-46.4) are in Appendix H-l. The runs
were made under variable temperature conditions with temperatures ranging
from 760 to 920°C. Ca/S molar ratios varied from 0.9 to 2.5.
The results of the runs using precalcined limestone at 865 to 920°C are
given in Figure 111-22. Figure 111-22 also shows, for comparison, the correla-
ting lines for dolomite and raw limestone developed earlier and reported in
the previous report (1). As seen in Figure 111-22, the precalcined limestone
is as active, at an equivalent Ca/S molar ratio, as dolomite and is much
more active than raw limestone. Also, the activity is maintained at the
lower temperature (865°C) where limestone would not be expected to calcine
extensively in the combustor.
A run was also made with precalcined limestone at a very low "turndown"
temperature (760°C) where previously, raw limestone was shown to be com-
pletely inactive (1). The result is given in Figure 111-23 where it is
compared to earlier results obtained with dolomite. The solid line is the
correlating line developed for dolomite and precalcined limestone. The
individual data points are those obtained with dolomite and precalcined
limestone at the very low temperature "turndown" conditions (690-760°C).
The dashed line correlates the low temperature results obtained with dolomite
and shows that the dolomite is slightly less active at the lower temperatures.
Again, precalcined limestone is as active, at an equivalent Ca/S molar ratio,
as dolomite, even though the low temperatures would strongly favor the forma-
tion of CaC03 rather than CaO in the combustor.
The effect of precalcining limestone was studied previously in TGA
equipment by Westinghouse Research Laboratory (5). In that study it was
found that precalcination under a high C02 partial pressure resulted in a very
active sorbent. The results obtained in this study support the Westinghouse
results. In other studies carried out at Argonne National Laboratory (6)
it was found that the activity of a sorbent could be related to the surface
48
-------
100
ce:
CM
O
FIGURE 111-22
S02 RETENTION WITH PRECALCINED LIMESTONE
DOLOMITE (840-950°C
Combustor Temp.)
90
80
70
60
50
40
30
20
LIMESTONE
(Not Precalcined)
(925-950°C
Combustor Temp.)
LIMESTONE
(Not Precalcined)
(825-900°C
Combustor Temp.)
Precalcined Limestone
865°C Combustor Temp.
920°C Combustor Temp.
10
0
0
2 3
Ca/S MOLAR RATIO
49
-------
FIGURE 111-23
EFFECT OF LOW TEMPERATURE OPERATION ON S02 RETENTION
100
90
80
LU
QL
CM
O
CO
70
60
50
40
30
20
10
0
I
DOLOMITE & _
PRECALCINED
LIMESTONE
(840-950°C)
I
I
DOLOMITE & PRECALCINED
' LIMESTONE
(690-760°C)
Dolomite (690-760°C)
Precalcined Limestone (760°C)
Ca/S MOLAR RATIO
50
-------
area of the sorbent occurring in large diameter pores. Precalcining under
high C02 partial pressure conditions apparently produces such a favorable
pore structure. Also, the pores appear to be large enough that carbonation
of the limestone, which occurs under very low temperature conditions, does
not reduce the pore diameter enough to prevent diffusion of SC>2 into the
interior of the limestone particles.
Further evidence of the high activity of the precalcined limestone can
be seen from the level of calcium sulfation in the used sorbent. The sulfa-
tion levels of limestone varied from 20 to 39% in previous runs made in the
miniplant. In the series of runs made with precalcined limestone, the sulfa-
tion level ranged from 55 to 67%. This is comparable to sulfation levels
normally measured with dolomite sorbent.
As a result of the above and previous studies, the sorbent requirements
needed to satisfy the current EPA new source performance standards for SC>2
emissions from a coal fired boiler (1.2 Ib S02/M BTU coal fired) can be esti-
mated. The estimate is shown in Table III-3. The estimate was based on a
gas phase residence time of 2 s and a boiler temperature of 930°C. As seen
in Table III-3, dolomite and precalcined limestone are more effective than
limestone on a molar basis. However, on a weight basis, limestone is slightly
more effective than dolomite with a coal containing 2% sulfur. Limestone and
dolomite are equivalent for a 3% sulfur coal. For coals containing more than
3% sulfur, dolomite is more effective than limestone even on a weight basis.
However, precalcined limestone is more effective than dolomite for all sulfur
levels.
The advantage of precalcined limestone as seen in Table III-3 is signi-
ficant. The weight requirements, which are expressed on an uncalcined stone
basis, are half the dolomite requirements and as little as 40% of the lime-
stone requirement without precalcination.
NOX Emissions
NOX emissions measured in all miniplant runs, including those reported
previously, are plotted in Figure 111-24 as a function of percent excess air.
Data obtained in the runs made in this reporting period follow the same trend
line and are within the data scatter of the earlier runs. The NOX emissions
were found to vary from 50 to 200 ppm or 0.04 to 0.17 g (as N02)/MJ (0.1 to
0.4 Ib/M BTU). Though the operating conditions varied greatly, the only
statistically significant variables were the excess air (or the flue gas
oxygen concentration) and bed temperature. The NOX emissions increased 4
fold, from 0.04 to 0.17 g/MJ over the 5 to 110% range of excess air. The
temperature effect in the 670 to 940°C (1250 to 1750°F) range was secondary
and caused only a 25% increase in the emission level. The emissions are well
below the EPA new source performance standard of 0.3 g (as N02)/MJ (0.7 Ib/M
BTU) and have an average value of only 0.09 g/MJ (0.2 Ib/M BTU) at 15% excess
air, the level most likely to be used in a commercial size boiler.
Previous studies carried out by Argonne (4) and Exxon (7) indicated
that NOX emissions from an atmospheric FBC unit could be reduced by a two
stage combustion process. In this process, a substoichiometric quantity of
air would be used to fluidize the bed and partially burn the coal. The
balance of the combustion air would be added at a slightly higher elevation
51
-------
TABLE II1-3. SORBENT REQUIREMENTS
Ul
Ca/S Wt. Uncalcined Sorbent/100 Wt . Coal
S (%) Retention (%)
2
3
4
5
Residence Time
Temperature
59
73
79
84
2 s
>930°C
Precalcined
Limestone Limestone Dolomite Limestone
1.3 0.8 0.8 8.2
2.1 1.0 1.0 20
2.8 1.2 1.2 34
3.2 1.3 1.3 51
Precalcined
Limestone
5.0
9.4
15
20
Dolomite
10
20
29
40
S0? Emissions 1.2 lb/M BTU
-------
FIGURE 111-24
0.8
CORRELATION OF NO EMISSIONS
X
0.7 -
0.6
5 0.5
D
E 0.4
"•s.
^x 0.3
0.2
0.1
0
i
o
20
40
60 80
EXCESS AIR, %
100
120
140
-------
in the combustor. In the zone between the coal injection point and the point
where the secondary air was added, a reducing environment would be established
which would promote NO destruction reactions such as reduction with CO, H2 or
carbon. This concept was tested by Argonne and Exxon and was found to reduce
NOX emissions significantly. No extensive, planned experimental work has been
carried out to investigate the use of staged combustion in a pressurized FBC
system. However, some unplanned results were obtained in a miniplant run
which suggests that staged combustion may have merits in a pressurized system
as well. In one run (run 47), air compressor problems resulted in poor com-
bustion of coal and the formation of CO levels in the flue gas up to 1800 ppm.
As the CO concentration increased, the NOX concentration dropped to 10 to 20
ppm. This is shown in Figure 111-25. These very low NOX levels measured
under reducing conditions indicate that two stage combustion may be a pos-
sibility for pressurized FBC and could reduce NOX emissions to almost insigni-
ficant levels.
Other Gaseous Emissions
803 emissions in the flue gas were found to vary widely, usually over a
range of 0 to 30 ppm with some measurements as high as 50 to 150 ppm. No
correlation was found with operating conditions and it is postulated that
the 803 was most likely formed in the flue gas sampling system.
CO emissions were very low, generally in the range of 50 to 200 ppm at
bed temperatures above 825°C except in those cases, as described above, where
upsets occurred which produced poor coal combustion conditions. As the tem-
perature was reduced below 825°C, CO emissions increased sharply to 300 to
800 ppm at 700 to 750°C.
Particulate Emissions
Particulate emissions in the flue gas downstream of the secondary
cyclone were measured using the system described on page 35.
Tables III-4 and III-5 summarize particulate emission results. The
first table gives results before the sampling system was modified by place-
ment of a second Balston filter immediately after the first filter. The
second table gives the particulate test results after the modification. Data
are also shown after further modification (see p. 35) and after the granular
bed filter was operated. In Table III-4 some deviation in grain loading can
be observed. When the loading values were >4 g/m3, it was due to the plug-
ging and ineffectiveness of the secondary cyclone during those particular
runs. Run 43 showed consistently low loadings, and it may be related to
the fact that Illinois No. 6 coal was used in that run. In most other runs
the values fluctuate between about 1 to 2.3 g/m3 (0.4 to 1.0 gr/SCF) giving
a mean value of 1.5 g/m^ (0.65 gr/SCF) with a standard deviation of 52%.
After system modification, however, and before the granular bed filter was
operated, particulate loading value (Table III-5) were quite constant with
a mean value of 2.3 g/m3 (1.0 gr/SCF) and a standard deviation of 13%.
These results seem to indicate that when low loadings were observed before
system modification, they may have been due to particulates by-passing the
first filter either due to a leak developing in the cartridge or displacement
in the cartridge-housing assembly. Such problems may occur due to cartridge
54
-------
FIGURE 111-25
N0v VS. CO IN FLUE GAS
X
Q.
n.
LO
<
o
LJ
Z)
110
100
90
80
70
L_
\ 60
o
< 50
u
a
^
o
o
40
x 30
20
10
0
O
_L
O
_L
_L
_L
J_
_L
0 200 400 600 800 1000 1400 1800
CO CONCENTRATION IN FLUE GAS (ppm)
2000
55
-------
TABLE III-4. PARTICULATE SAMPLING SUMMARY I
(BEFORE MODIFICATION TO SAMPLING SYSTEM)
Ln
Run No.
26
28
28
28
29
30.1
31
32.2
32
33
34
36
37
39
39
43
43
.3
.1
.1
.2
.2
.3
43.4
43.5
45
47
48
Sampling Time
(h)
13.58
1
1
1
2,
2.
2.
2,
2,
5,
1,
2,
70
30
25
58
58
00
75
50
47
4.42
58
00
5
3.5
,50
.00
2.50
1.00
2.08
Total Solids
Collected
(g)
224.5
16.9
4.0
7.0
19.0
14.0
45.0
25.0
27.0
95.0
85.0
77.0
86.0
175.5
260.4
14.5
9.9
11.0
4.5
38
Deviation from
Isokinetic Sampling
+42
0
0
0
-16
-22
-7
+4
-12
+11
+11
+10
-48
+3
+3
-4
+10
+12
+12
0.80
0.59
0.37
0.30
0.53
0.43
2.88
0.92
1.65
1.42
1.58
2.20
2.36
4.81
4.81
0.73
0.37
0.32
0.29
1.15
5.03
-3
Particulate Loading
in Flue Gas
(K/m3) (Rr/SCF)
0.80
0.59
0.37
0.30
0.53
0.43
2.88
0.92
1.65
1.42
1.58
2.20
2.36
4.81
4.81
0.73
0.37
0.32
0.29
1.15
5.03
1.53
0.35
0.26
0.16
0.13
0.23
0.19
1.26
0.40
0.72
0.62
0.69
0.96
1.01
2.10
2.10
0.32
0.16
0.14
0.13
0.4
2.2
0.67
-------
TABLE III-5. PARTICULATE SAMPLING SUMMARY II
(AFTER FIRST MODIFICATION TO SAMPLING SYSTEM)
Total Solids
Collected
Deviation from
Particulate Loading
in Flue Gas
Run No.
50.12
50.2
50.4
50.5
51 (Catch 1)
51 (Catch 2)
52
57
59 (Catch 1)
59 (Catch 2)
59 (Catch 3)
(h)
5.28
1.15
1.48
0.58
0.63
0.60
0.50
1.08
1.17
13.33
2.00
(8)
112.63
35.09
33.69
13.31
14.78
13.64
14.73
RUNS WITH GRANULAR
(AFTER SECOND MODIFICATION
(%)
-5.8
+1.8
-7.4
-7.4
-4.3
-5.8
-0.4
BED FILTER
TO SAMPLING SYSTEM)
-0.9
+1.4
-3.0
-0.4
(g/m3)
2.60
2.88
2.31
2.33
1.95
1.92
2.65
(gr/SCF)
1.05
1.26
1.01
1.02
0.85
0.84
1.07
0.08
0.08
0.28
0.54
-------
Balston Filter
2.60
2.88
2.31
2.33
SASS Train
1.81
1.92
2.86
2.79
overloading and/or a high pressure drop across the filter. Though a slight
fluctuation in grain loading may be expected for different runs, an average
value of 2.3 g/m3 (1 gr/SCF) may be used at present.
Operation with the granular bed filter resulted in lower particulate
levels. This is discussed further in Section V.
A comparison was made between the loadings measured using the Balston
total filter system and a SASS train (3 cyclones plus a total filter) used by
Battelle during runs 50.1 to 50.5 (18). As seen in Table III-6, the loadings
compare well, with an average deviation of 30%.
TABLE III-6. COMPARISON OF PARTICULATE LOADINGS
MEASURED BY TOTAL FILTER (EXXON) AND SASS TRAIN (BATTELLE)
3
Particulate Loading (g/m )
Run No.
50.1
50.2
50.4
50.5
Individual filter loadings and temperatures are given in Table III-8
for Runs 50 to 59. The required high temperatures were realized in Run 52.
However, when the granular bed filter was operated, the temperatures were
lower due to heat loss in the granular bed filter. Future runs will aim
towards measurement near the cartridge allowable temperature of 530°C for
the first two filters and 206°C for the third low temperature filter. A
significant amount of material was collected in the second filter at times.
Much of the material collected in filter 3 (low temperature) was often
greenish-black in color and was due to corrosion of the piping and the heat
exchanger tubes.
Finally, Tables III-7 and III-9 show particulate distributions. The
values in the two tables, corresponding to runs before and after system
modification, are slightly different. The later values, Table III-9, show
a mass median particle size of 3 to 6 ym compared to earlier values of 6
to 8 jam. Runs made with the granular bed filter (Runs 57 and 59) show a
smaller median particle size of 2 to 5 ym as expected.
TABLE III-7. SIZING OF PARTICLES OBTAINED
ON BALSTON FILTERS BEFORE MODIFICATION
Run No.
31
32.2
32.3
33
36.2
37
Particle Size (]jm)
10%
Less Than
1.8
2.0
2.3
3.3
2.5
2
25%
Less Than
3.1
3.1
3.3
4.8
4.1
3.3
50%
Less Than
8.0
5.8
5.9
7.6
7.5
6.5
75%
Less Than
23.5
12
13
13.5
14.3
14.5
90%
Less Than
~50
24
32.5
25.5
30
30
58
-------
TABLE II1-8. INDIVIDUAL FILTER RESULTS (RUNS 50 TO 59)
Filter 1
Filter 2
Filter 3
Run
50.1
50.2
50.4
50.5
51 (Catch 1)
51 (Catch 2)
52
57(2)
59 (Catch 1)
59 (Catch 2)
59 (Catch 3)
Op. Temp.,(l) °C Loading, g
55 to 330 91.4
90 to 270 35.0
50 to 360 15.2
80 to 290 12.7
125 to 390 12.79
170 to 390 12.94
370 to 500 12.63
230(3)
360
380
390
Op. Temp., °C Loading, g
50 to 290 19.4
70 to 220 0.3
40 to 310 17.9
70 to 240 0.01
80 to 310 0.3
95 to 310 0.32
255 to 440 2.21
170(3>
240
270
270
Op. Temp., °C Loading, g
70 to 170 1.9
80 to 130 0.04
70 to 150 0.5
75 to 140 0.6
100 to 180 1.69
100 to 180 0.38
200 to 210 0.09
(1) Operating temperature at the beginning and end of sampling time.
(2) Runs 57 and 59 made with granular bed filter
(3) At end of sampling time.
-------
TABLE III-9. SIZING OF PARTICLES OBTAINED
ON BALSTON FILTERS AFTER MODIFICATION
Particle Size (ym)
Run
50,
50,
50,
50,
51
51
52
57
59
59
59
1
2
4
5
(Catch 1)
(Catch 2)
(Catch 1)
(Catch 2)
(Catch 3)
10%
Less Than
1.0
0.9
1.0
1.0
1.1
0.8
1.1
1.8
1.8
1.0
1.0
30%
Less Than
2.3
1.6
1.6
2.2
3.7
1.9
2.7
3.0
2.1
1.5
1.7
50%
Less Than
4.5
3.0
2.7
3.8
7.8
3.0
5.6
5.4
2.7
2.0
2.6
70%
Less Than
8.8
7.8
5.2
6.4
22.5
6.0
13.0
10.0
3.9
2.7
5.6
A comparison was also made with the particle size distributions measured
by Battelle using the SASS train (18) and by Acurex using a cascade impactor
for run 50.5 (17). The result, shown in Table 111-10, show good agreement
with those measured by Exxon by using the Coulter Counter method on particles
captured by the Balston filter.
TABLE 111-10. COMPARISON OF PARTICLE SIZE
DISTRIBUTIONS MEASURED BY THREE DIFFERENT METHODS
Particle Size (yim)
Laboratory
Exxon
Battelle
Acurex
Method
Coulter Counter
of Filter Sample
SASS Train
Cascade Impactor
10%
Less Than
1.0
1.2
1.0
0.8
30%
Less Than
2.2
2.6
2.8
2.0
50%
Less Than
3.8
5.0
5.8
3.7
70%
Less Than
6.4
10
11
7.4
Attempts made to date to determine if the particulate loading, size
and composition are affected by condensation have yielded inconclusive
results. The sampling temperatures have been too low. The new, high tem-
perature system should help determine if condensation occurs at temperatures
less than 870°C and if this does influence the particulate measurements.
Combustion Efficiency
Carbon combustion efficiency results published in the previous report
(1) indicated that a regression analysis determined that temperature was
the only variable influencing combustion efficiency. However, a plot of
combustion efficiency against temperature showed three distinct curves with
a fair degree of data scatter. The reasons for the three curves and the
60
-------
scatter were not explained at that time. Combustion efficiency data were
again studied using multiple regression analysis. Data from 55 miniplant
runs were correlated using 29 possible variables including interactions
between variables.
The best fit correlation indicated that the combustion efficiency is
a strong function of temperature (T) and also the excess air level (EA), a
result which is consistent with data reported by other investigators (10,11)
and also found in previous batch unit results by Exxon (1).
The other variables that showed a significant effect on combustion
efficiency were the cross product of temperature and excess air level and
the cube of the residence time (t). The equation correlating these variables
to combustion efficiency was determined to be:
C.E.(%) - 161.3 = 1.62 X 10'1 (T) + 1.0 X 10~4 (T2) +
7.51 X 10~5 (T) (EA) - 4.62 X 10~4 (EA)2 + 1.73 X 10~2 (t)3
where T is in °C, EA in % and t in s.
The correlation coefficient corresponding to this fit is 82%. The
standard error of the correlation is 0.6 combustion efficiency units.
A plot of the calculated combustion efficiencies using this equation
versus the observed combustion efficiency is presented in Figure 111-26. One
explanation for the apparent data scatter in the graph is the difficulty in
correlating a large number of observations with such a small span of the
dependent variable. It should be noted that the miniplant uses a particulate
recycle system to increase combustion efficiency. This may have the effect
of reducing to insignificance variables which may be highly significant in
a once through system without particulate recycle. Additional work will be
required to compare the efficiencies measured in a once through system with
those measured in a recycle system. A preliminary estimate of the effect of
recycle was obtained when a single run (run 51) was made with no recycle from
the first stage cyclone to the combustor. The cyclone was deactivated for
this run by slipping a blind between the flange at the bottom of the cyclone
and the top of the return line. The combustion efficiency in this run made
at 875°C at an excess air level of 45% was 98.2%. Runs made under similar
conditions but with recycle, gave combustion efficiencies 1.0 to 1.5 per-
centage points higher.
Cyclone Efficiency
The collection efficiency of the second stage miniplant cyclone was
calculated using collection and particle size data obtained in recent runs
(Runs 48, 50.3 and 51). The overall secondary cyclone efficiency was as
follows:
Run Overall Efficiency, %
84
84
86
61
-------
o
LU
CO
ID
ca
o
o
Q
LJ
O
O
FIGURE 111-26
COMBUSTION EFFICIENCY CALCULATED VS. MEASURED
100
99
97
96
95
T
T
1
1
96 97 98 99
MEASURED COMBUSTION EFFICIENCY (%)
100
62
-------
The values given above indicate the cyclone efficiency to be the same within
experimental errors. It should be noted that during Run 51 the primary
cyclone was not used, and the secondary cyclone was used alone. Yet the
overall cyclone efficiency value from Run 51 corresponds closely to the
results from those runs in which the primary cyclone was also operating.
The efficiency curves for all 3 runs are given in Figure 111-27. With the
amount of data available at present, it is difficult to state whether differ-
ences in the efficiency curves are due to the operation of the second cyclone
by itself or not. However, these curves are comparable to those obtained
from earlier runs. The 50% cut point lies between 3 to 3.5 pm, indicating
a relatively high efficiency for a conventional cyclone.
Particle size distribution of the material collected in the second stage
cyclone is typified by results for Runs 50.1, 50.2 and 50.3 shown in Table
III-ll. These distributions are similar to those reported previously (1).
Additional size distribution data for other runs are given in Appendix H-2.
TABLE III-ll. SECOND STAGE CYCLONE
LOCK HOPPER PARTICLE SIZE DISTRIBUTIONS
Particle Size (ym)
10 Wt. % 25 Wt. % 50 Wt. % 75 Wt. % 90 Wt. %
Run No. Less Than Less Than Less Than Less Than Less Than
50.1 7 13 22 52 98
50.2 6 9 17 38 66
50.3 6 10 17 37 69
63
-------
FIGURE 111-27
SECONDARY CYCLONE EFFICIENCY VS. PARTICLE SIZE
o1^-
-
o
UJ
o
u_
1 1
u_
UJ
2;
o
1-
o
UJ
1
o
0
100
90
80
70
60
50
40
30
20
10
0
i i i i i i i i i i ii
^OJ=s=.g==— °^°^ "
0^8^^° * ^\S"
X^A / V
/A xD
^^ S
/ /v
Of
- A s —
/ /
//D
I/
8 RUN No.
/A 048
A D 51
^y A5Q.3
a>/O
nn^X^^ | I 1 1 1 1 1 1 1 1 II
2 3 4 5 6 7 8 10 20 30 40 5
PARTICLE SIZE,
-------
SECTION IV
REGENERATION STUDIES
Work is continuing to develop a process for continuously regenerating
sulfated limestone or dolomite sorbent (CaS04) in a fluidized bed by reaction
with a reducing gas at about 1100°C and elevated pressure. In the previous
annual report (1), experiments were described in which batches of sulfated
sorbent were regenerated in the miniplant regeneration vessel. In those
preliminary tests, up to 3.7 mole percent (dry basis) of S02 was produced,
reduction of CaSO^ to CaO was nearly complete, and agglomeration of the
fluidized bed was avoided by careful control of temperature. The regenerator
was coupled to the miniplant combustor so that sorbent could be continuously
recirculated between the two vessels and a 24 hour shakedown run was com-
pleted. Since the last report, the operability of this system was demonstrated
in a five day run in which both combustor and regenerator operated without
interruption. In this report, the equipment and procedures used to make
the run, and the experimental results are described.
EQUIPMENT, MATERIALS, PROCEDURES
Equipment
The equipment and materials were described in detail in the previous
report (1). In most cases, no major changes have occurred since and only
summaries are reported. When changes have occurred, for example in the fluid-
izing grid for the regenerator, a detailed description is given here. Also
described here are the design and operation of the system used to transfer
sorbent between the combustor and regenerator.
Air System
The two separate air systems are (1) burner air and (2) supplementary air.
All air is supplied by the main air compressor. Automatic control systems,
consisting of control valves, flow measuring orifices, and electronic con-
trollers, are used to regulate air flow. Burner air is supplied to the bur-
ner, located beneath the fluidizing grid, in sufficient quantity to completely
burn the fuel (natural gas). Supplementary air is added about halfway up
the bed in order to create an oxidizing zone in the upper portion of the bed.
Fuel System
The two fuel systems are (1) burner fuel and (2) supplementary fuel.
Automatic control systems, similar to those used for air flows, are used to
regulate the flow of natural gas. Burner fuel is supplied to the burner
where it is burned with an approximately stoichiometric amount of air.
Supplementary fuel is added directly to the regenerator column just above the
fluidizing grid in order to produce reducing gases (CO, H£) .
Off Gas Handling
Hot pressurized gases leaving the regenerator are cooled in a single
pass double pipe heat exchanger and expanded to nearly atmospheric pressure
across a control valve. Dust is removed from the gas upstream of the cooler
65
-------
by a cyclone and upstream of the pressure control valve by a sintered stain-
less steel bayonet-type filter. The gas is filtered in order to reduce
wear on the pressure reducing valve.
Off-gases from the regenerator are sent to a Research-Cottrell scrubber
for cleanup before venting. An alkaline solution (Na2C03) is used in the
scrubber to absorb S(>2.
Gas-Sampling System
A slipstream is taken downstream of the pressure reducing valve. The gas
is filtered (Balston Model 33 filter) and dried (Perma-Pure Model PD-1000-24S
self-regenerative membrane-type dryer) before entering the analyzers.
Fluidizing Grid
The fluidizing grid had failed due to overheating at the conclusion of
the 24-hour shakedown run of the combustor-regenerator. This grid had 88
holes of 3.6 mm (9/64 in) diameter for passage of the fluidizing gases (from
burner located beneath grid) and 14 water cooling channels of 4.8 (3/16 in)
diameter. Cooling water was supplied through two 7.7 mm (0.305 in) I.D. tubes
to a header and flowed in parallel through the cooling channels. Examination
of the grid showed that the metal surrounding the outer channels had been
overheated. It was concluded that the flow of water through the channels was
non-uniform, with the inner channels receiving most of the water.
In order to achieve more uniform flow of water throughout all the cooling
channels, the grid was modified so that water flow could be controlled inde-
pendently through six separate cooling zones. Hence, the 14 channels were
separated into groups of 3, 2, 2, 2, 2, and 3, each group or zone having its
own water supply. The new design apparently solved the problem: after the
125 hour demonstration run, the grid was examined and found to be in
excellent condition, with no signs of overheating.
Burner
No changes were made to the regenerator burner since the last report.
This unit is identical to that used in the miniplant combustor and is des-
cribed in the previous annual report (1).
Sorbent Transfer System
Combined operation of the combustor and regenerator required development
of a transfer system to circulate sorbent between the two vessels. A number
of alternative approaches were considered and tested (1). Basically, these
approaches could be divided into two types: high and low density solids
flow. The approach thought to be most suitable for the miniplant utilized
high bulk density (stick-slip) flow of sorbent in transfer lines. This
technique had the important advantages that (1) large amounts of carrier
gas would not be needed, as would be the case if pneumatic conveying were
used, and (2) short, straight transfer lines could be used. The small distance
between the combustor and regenerator would have made it difficult to accom-
modate the U-shaped transfer lines that would have been required in a pneum-
atic system.
66
-------
Figure IV-1 shows schematically the sorbent transfer system. Pressure
in the regenerator was maintained slightly higher than that in the combustor.
Solids in the regenerator-to-combustor transfer line moved into the combustor
when a pulse of nitrogen was applied to the lower end of the transfer line.
The flow rate of solids was controlled by adjusting the frequency, duration,
and intensity of the pulse. Two slide valves were used in the combustor-to-
regenerator transfer line in order to prevent back flow of gas from the
regenerator up the line. These automatic valves trapped solids in the piping
between them. Solids were discharged into the regenerator when the bottom
valve was opened. The two solids' take-off plugs shown in Figure IV-1 were
inserted into the ports during start-up to prevent solids from entering the
lines. Plugging could occur if the solids became wet due to water condensa-
tion during start-up. The manual slide valve in the regenerator-to-combustor
line was also closed during start-up and also during upsets. A photograph
may be seen in Figure IV-2.
The components of the sorbent transfer system (valves, expansion joints,
etc.) are described in detail in the previous annual report. The transfer
lines themselves were fabricated from 6 inch Schedule 40 carbon steel pipe
and refractory lined to an inside diameter of 7.6 cm. The sloping portions
of the lines were sleeved with 2-1/2 inch Schedule 10 316 stainless steel
pipe, which had an inside diameter of 6.7 cm.
Miscellaneous Modifications to Equipment
Made Prior to Demonstration Run
In addition to redesigning the regenerator fluidizing grid, a number of
other changes were made to the combustor-regenerator prior to the demonstra-
tion run in order to improve reliability. The more noteworthy changes were:
(a) Spare supplementary fuel and air inlet ports were added.
If a plug occurred at one inlet port, the second could
be used to continue the run.
(b) The double pipe off-gas cooler was shortened in order
to raise the temperature of gas exiting from the cooler
and reduce water condensation.
(c) A second pressure control valve was added in parallel
with the original valve. One valve was used at a time
but the second could be put in service without inter-
rupting the run if a problem occurred.
(d) A new thermocouple port was added just above the
fluidizing grid in order to detect high temperatures
in this critical region.
(e) The system to measure pressure drop across the fluidizing
grid and across different points in the bed was modified
to be more reliable and provide more accurate data.
67
-------
REGENERATOR
COMBUSTOR
SOLIDS TAKE
OFF PLUG
AUTO
SLIDE
VALVE
MANUAL
SLIDE
VALVE
NITROGEN
PULSE
FIGURE IV-I
MINIPLANT SOLIDS TRANSFER SYSTEM (SCHEMATIC)
68
-------
FIGURE IV-2
MINIPLANT SOLIDS TRANSFER SYSTEM
69
-------
Materials
No changes in materials of construction occurred during the past year.
The regenerator is lined with General Refractories Litecast 75-28, a
castable refractory with a service temperature limit of 1540°C (2800°F) and
a thermal conductivity (at 540°C) of 0.55 W/m2°C (3.8 BTU/hr ft2 °F/in).
The thermocouples used to measure fluidized bed temperature are Type K
(chromel-alumel) protected by a silicon carbide sheath. Silicon carbide
provides excellent resistance to high temperatures and chemical attack.
Because this material has very little strength in tension, an Inconel tube
is used inside the sheath to provide reinforcement.
The alloy used for the gas-contacted sections of the regenerator off gas
cooler and filter is type 316 stainless steel. This material offers good
resistance to dry and wet S02 although H2SO^ can cause serious corrosion even
at low concentrations. This has been a problem with the filter element which,
because of its high surface area, is more susceptible to corrosion. Corrosion
has been minimized by thoroughly washing and drying the filter element after
each run.
EXPERIMENTAL RESULTS AND DISCUSSIONS
Combustor-Regenerator Demonstration Run
A successful combustor-regenerator demonstration run (Run 45) was com-
pleted in October, 1976. The miniplant was operated with limestone sulfur
acceptor recirculating between the combustor and regenerator until the run
was voluntarily terminated after 125 hours of continuous operation. No inter-
ruptions in coal feed to the combustor or fuel feed to the regenerator
occurred during the extended run.
Two unsuccessful attempts preceeded the successful run. The first
was shut down after only four hours when a temperature runaway in the com-
bustor resulted in agglomerated bed. This problem was caused by a malfunction
in the load cell amplifier for the coal feed system. Electrical problems
also plagued the second attempt and the run was terminated when a malfunction
of the AP transmitter in the control loop for regenerator fluidizing air
caused the air flow to rapidly increase, thereby blowing a portion of the bed
out of the regenerator.
Operating Plan and Conditions
The major purpose of the run was to demonstrate that the combustor-
regenerator could be operated continuously for 100 hours. Thus, operating
conditions were deliberately conservative in order to maximize the chance of
reaching this goal. It was intended to keep all operating conditions steady
throughout the demonstration run, with the exception of the feed rate of
fresh (makeup) limestone into the combustor (Ca/S ratio). Operating condi-
tions are summarized in Table IV-1.
70
-------
TABLE IV-1. OPERATING CONDITIONS
DURING DEMONSTRATION RUN
Pressure, kPa
Bed Temperature, Average, °C
Bed Height, Expanded, Avg., m
Superficial Gas Velocity, m/sec
Solids Recirculation Rate, kg/hr
Makeup Acceptor Addition Rate,
Equiv. Ca/S, Average
Range
Combustor Coal Feed Rate, kg/hr
Coal Type
Stone Type
Combustor
760
900
3.4
1.5
79
45
0.55
0-1.3
Regenerator
770
1010
2.3
0.6
Champion (Pittsburgh Seam), 2.0% S
Grove Limestone No. 1359
71
-------
One of the most pressing questions the operator of a fluidized bed com-
bustion system could ask is what savings in sorbent would be realized if a
regenerator were added to his system. In order to make a first attempt at
answering this question, it was decided to feed the smallest quantity of makeup
sorbent into the combustor that would allow operating the system with (1) SC^
emissions from the combustor below the EPA emission standard, and (2) constant
fluidized bed levels in both the combustor and regenerator. In other words,
the amount of fresh limestone required would be determined either by the
activity or the attrition rate of the sorbent. As it turned out, the limiting
factor during this run was maintaining bed levels. Over the 100 hour period
of regeneration, S02 emissions from the combustor were always below the EPA
standard of 1.2 Ibs S02/106 BTU, but makeup limestone had to be added because
bed levels declined during the first half of the run. However, it is quite
likely that had the run been continued, makeup limestone would have been
needed to increase sorbent activity and control S02 emissions.
There was an alternative manner in which the system could have been
operated. Since it was not known in advance exactly what attrition rates
would prevail, a makeup rate could have been chosen which was sure to be
higher than the maximum attrition rate expected, say 2.0/1 or 3.0/1 Ca/S.
The limestone makeup rate would then have been constant throughout the run
and bed levels could have been maintained by varying the rate at which stone
was rejected from the combustor. However, this would have been a very
unrealistic way of operating a real plant, since far more fresh sorbent would
be used than was needed. Furthermore, operating the system in this manner
would not have provided any information on the degree of reduction in sorbent
requirements possible in a regenerative system.
For the first 24 hours of the run, the regenerator was operated under
oxidizing conditions in order to establish baseline operating conditions.
Sorbent was recirculated between the combustor and regenerator during this
period. Subsequently, reducing conditions were established by increasing
the flow of supplementary fuel to the regenerator. Operation of the system
continued, uninterrupted, for the next 100 hours.
Operating Performance
Operation of the combustor-regenerator during the extended run was
exceptionally smooth and no interruptions in coal feed to the combustor, fuel
feed to the regenerator, or in the recirculation of solids occurred during
the 100 hour period when the regenerator was in reducing conditions. The
only potentially serious problem was a hot spot just above the regenerator
fluidizing grid. The hot spot developed during startup as the regenerator
bed (fresh limestone) was being heated under oxidizing conditions. When the
bed temperature reached about 1000°C a sudden shift in the temperature pro-
file occurred and the temperatures at the bottom of the bed quickly climbed
to about 1200°C. Temperature was reduced by decreasing the air and fuel
inputs. The regenerator was maintained at constant oxidizing conditions for
24 hours in order to establish baseline operation in the combustor. Transfer
of solids occurred continuously throughout this period.
72
-------
When reducing conditions were established in the regenerator, temperature
of the hot spot increased to about 1260°C. Fuel and air flows were decreased
and pressure was reduced from the initial pressure of 920 to the ultimate
operating pressure 760 kPa. The reduction in pressure caused an increase in
superficial gas velocity, which improved mixing of solids and flattened the
temperature profile somewhat. Superficial gas velocity was 0.6 m/s, which
is just slightly above the minimum velocity to achieve fluidization.
There were no changes in operating conditions in the regenerator for
the duration of the extended run. However, the hot spot, which was initially
located at a position about 13 cm above the fluidizing grid, increased in
size so that towards the end of the run, high temperatures were present as high
as 43 cm. above the grid. By comparison, the reducing zone of the regenerator
extended to 74 cm above the grid and the average total expanded bed height was
230 cm.
After the run, the plenum of the regenerator was dismantled and the
fluidizing grid and bed examined. The grid was in excellent condition and no
signs of overheating were present. There was a "crust" of bed material which
covered the grid to a depth of about 8 cm, this material was loosely packed
and porous and would not have interfered with flow of air through the grid.
Above the crust was about 30 cm of hard, fused bed, much of which must have
formed during shutdown. Upon shutdown, air was blown through the bed and
temperatures at the bottom of the bed increased over 1000°C. The temperature
excursion was probably caused by oxidation of calcium sulfide which was pre-
sent in the reducing zone of the bed. Normally, nitrogen is used during
shutdown to prevent this type of occurrence. However, in this case, the
compressed nitrogen supply was inadequate and compressed air was used instead.
Makeup Limestone Addition Rate.
Fluidized Bed Levels
Makeup limestone was added to the combustor in order to maintain reason-
ably constant fluidized bed levels in the combustor and regenerator. The feed
rates of makeup limestone are summarized in Table IV-2. The makeup rate
from the start of the run was 3.6 kg/hr, which is equivalent to a Ca/S molar
ratio (Ca in stone to S in coal) of 0.74. This makeup rate was chosen because
bed levels during the 24-hour shakedown run made in July 1976 were kept con-
stant using this rate. Bed height in both combustor and regenerator is
plotted against hours into the run in Figure IV-3. It can be seen that bed
levels in the combustor began to drop after about 40 hours. When levels
dropped to the height of the takeoff port to the combustor-to-regenerator
solids transfer line, the feed rate of makeup stone was increased to 1.30
Ca/S. Failure to increase the makeup rate would have resulted in the level
of solids in the combustor falling below the takeoff port. This would have
caused the transfer of solids to stop.
The total mass of solids in the combustor and regenerator is plotted vs.
time into the run in Figure IV-4. As can be seen, inventory was dropping
fairly sharply after about 40 hours and feed rates of makeup limestone had
to be increased.
73
-------
TABLE IV-2. FEED RATES OF MAKEUP LIMESTONE TO
COMBUSTOR DURING COMBUSTOR-REGENERATOR DEMONSTRATION RUN
Hours Feed Rate of Makeup Feed Rate of Makeup
Into Run Limestone, kg/hr Limestone, Equivalent Ca/S
0-5 0 0
5-24 3.6 0.74
24 Reducing Conditions Established in Regenerator
24 - 56 3.6 0.74
56 - 65 6.4 1.30
65 - 80 5.2 1.06
80 - 86 2.3 0.46
86 - 124 0 0
74
-------
500
400
300
200
100
FIGURE IV-3
COMBUSTOR AND REGENERATOR BED HEIGHTS
E
o
LU
CO
0
0.74 Ca/S
i i i
<-1.30-»f— 1.06-^*0.46
0
RUN 45
o COMBUSTOR
o REGENERATOR
H
25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100105110115120125
HOURS INTO RUM
-------
FIGURE IV-4
TOTAL INVENTORY OF SORBENT IN THE COMBUSTOR AND REGENERATOR
340
Ca/S INCREASED
I
25 30 35 40 45 50 55 60 65 70 75 80 85 90 95100105110115120125
HOURS INTO RUN
-------
It was intended not to reject any sorbent from the combustor during the
run. However, after the run, it was discovered that 180 kg of sorbent had
been inadvertently removed from the combustor early in the run. This dis-
covery explains why high limestone addition rates were required early in the
run*
Combustor S02 Emissions
S02 emissions from the combustor are given in Figure IV-5. For the first
24 hours the regenerator was operated under oxidizing conditions with sorbent
recirculating between the combustor and regenerator. Emissions gradually
increased to about 550 ppm (1.1 Ibs S02/10& BTU). Since the S02 emissions
would have been about 1330 ppm at zero retention, 550 ppm corresponds to 59
percent retention. The makeup rate of fresh limestone feed into the combustor
during this period was 3.6 kg/hr (7.9 Ibs/hr), which is equivalent to a Ca/S
feed ratio of 0.74 (see Table IV-2).
It is interesting to compare the S02 emissions from the combustor during
the first day (oxidizing conditions in regenerator) to the emissions that
would have been expected without the regenerator. Figure IV-6 gives S02
emissions from once-through operation of the combustor as a function of Ca/S
feed ratio using limestone at carbonating and calcining conditions. The com-
bustor was operated at 900°C (1650°F) (carbonating conditions) so at a Ca/S
of 0.74, the expected S02 emission level would have been about 1100 ppm.
However, the measured S02 concentrated started to level out at about 550 ppm.
The lower than expected emission level can be explained because even though no
regeneration was occurring in the regenerator, calcination of the limestone
was occurring. The calcining conditions in the regenerator were the same as
those used in the limestone precalcination tests (p. 48). The S02 retention
(59%) measured during this period was also comparable to that measured in run
46.4 made with precalcined limestone at similar operating conditions. There-
fore, the regenerator was operating as a precalciner and the S02 emissions
from the combustor were indicative of those expected from the use of a pre-
calcined limestone, rather than from uncalcined limestone used under carbona-
ting combustor conditions.
Within about four hours after reducing conditions were established in the
regenerator by increasing the flow of supplementary fuel, S02 emissions from
the combustor fell from 550 ppm to below 200 ppm. After about 50 hours into
the run (25 hours under reducing conditions) S02 levels gradually increased and
reached about 550 ppm when the run was terminated after 100 hours of regenera-
tion. The increase in emissions is, for the present, assumed to have been
caused by a gradual decline in the activity of the regenerated sorbent. About
fifteen cycles of sulfur sorption and regeneration should have occurred during
the 100 hour period during which the regenerator was in reducing conditions.
The average feed rate of makeup sorbent was equivalent to a Ca/S ratio
of 0.55; however, the makeup rate was varied over the Ca/S range of 0-1.3 in
order to maintain constant bed levels. It should be emphasized that this
variation in makeup sorbent rate was very small compared to the rate at which
77
-------
oo
600
500
400
a.
a.
csj
300
200
100
FIGURE IV-5
COMBUSTOR S02 EMISSIONS
REDUCING CONDITIONS ESTABLISHED IN REGENERATOR
I I
_L I I I
0 10 20 30 '10 50 60 70 80 90100110120130140
HOURS INTO RUN
-------
FIGURE IV-6
S02 EMISSIONS VS. Ca/S RATIO
LIMESTONE No. 1359 IN ONCE-THROUGH OPERATION (REF. 1)
1600
1400
400
310
200
CARBONATING CONDITIONS
0
CALCINING
CONDITIONS
0 0.55
Ca/S RATIO
(Moles Ca In Feed Sorbent/Moles S In Coal)
79
-------
recirculated sorbent entered the combustor, which was equivalent to a Ca/S
ratio of about 17. There was no apparent immediate effect of varying the
makeup rate on the emissions of SO^ from the combustor.
Levels of S02 in Regenerator Off-Gas
The concentration of SC^ in the regenerator off-gas was nearly steady
throughout the run and averaged 0.53 mole percent (dry basis). This is very
close to the concentration predicted by a sulfur mass balance based on the
feed rate and sulfur content of the coal entering the combustor. The cal-
culated equilibrium concentration at the operating conditions of the regen-
erator was 2.9 percent; hence, higher S02 levels would probably have been
achieved by burning more coal, of a higher sulfur content in the combustor.
A commercial-sized plant would not be expected to be mass-balance limited and
this would be expected to yield higher S02 concentrations in the off gas.
Attrition Rate
During the 125 hour demonstration run, 336 kg (740 Ibs) of fresh lime-
stone was added to the combustor to maintain constant bed levels; however, as
noted earlier, 180 kg (396 Ibs) of sorbent was inadvertently dumped from the
combustor early in the run. The approximate amount of sorbent lost due to
attrition and entrainment was 336-180 = 156 kg (344 Ibs). The inventory of
solids in the system was about 280 kg (620 Ibs). Hence, the average entrain-
ment rate was 156/125 (280) = 0.0045 kg/hr kg inventory or 156/336 =0.46
when expressed as a fraction of the limestone makeup rate. This rate is
higher than expected based on once through (non regenerative) entrainment
measurements. More work is needed to explain this anomoly.
Analysis of Bed Solids
After the run, samples of bed were taken from the combustor and regen-
erator for analysis. Three samples of regenerator bed were analyzed because
of the large variability among samples. Only one complete analysis of com-
bustor bed was made because sulfate analysis of three samples showed very
close agreement. The large variability in samples from the regenerator was
probably a result of sampling problems caused by the regenerator bed being
partly agglomerated. Agglomeration occurred during shutdown when air was
blown through the hot bed, causing an exothermic oxidation of CaS and a tem-
perature rise to over 1100°C (2000°F). Table IV-3 gives the analytical
results. Note that no CaS was found.
80
-------
Table IV-3
Analyses of Bed Material Discharged from
Combustor and Regenerator After Demonstration Run
Regenerator^
_ _ _ Mole Percent ___
Component Sample #1 Sample #2 Sample #T Average
Ca° 87.0 73.2 81.2 80.5
CaCO 2.7 1.4 2.7 2 3
CaS04 10-3 25.4 16.1 17;3
Combustor
CaO 35.8
CaCO 18.5
45.7
Sulfur Mass Balance
Table IV-4 is a sulfur mass balance for the demonstration run
(entire 125 hour period). Recovery of sulfur was 103.5 percent.
The sulfur accumulated section of Table IV-4 comes about because
the system was charged with fresh limestone prior to the run, but sulfated
limestone remained at the conclusion. Also, the sulfur balance is very
sensitive to the sulfur content of the coal. A sulfur level of 2.0 percent
was the most recent analysis of Champion coal available. The actual sulfur
level would have had to be only 2.07 percent to get a calculated sulfur
recovery of 100 percent.
Conclusions
Operation of the combustor-regenerator during the demonstration run was
encouraging. In particular, the sorbent transfer system operated for 125
hours without a single problem. The chances are excellent that this system
can be used successfully in future runs.
Several areas leave room for improvement. The superficial gas velocity
in the regenerator was only 0.6 m/sec, which is only slightly above the
minimum fluidizing velocity. As a result, mixing of the regenerator bed was
slow. This factor surely contributed to the poor temperature profile and
localized agglomeration. The velocity will be increased slightly in sub-
sequent runs by using higher bed temperatures; however, it may be necessary
to operate the regenerator at lower pressures or even reduce its size in order
to obtain suitable gas velocities.
Another area for improvement is the concentration of S02 in the regen-
erator off-gas. During the demonstration run, the average concentration was
only 0.5 mole percent. This level was determined by a sulfur mass balance
for the combustor-regenerator system and not by thermodynamics, which would
have permitted 2.9 percent S02. Hence, the S02 level could have been higher
81
-------
TABLE IV-4. SULFUR BALANCE
COMBUSTION-REGENERATION DEMONSTRATION RUN
kg % of Sulfur Entering
• Sulfur Entering System
a. Coal 194.8 100
Total 194.8 100
* Sulfur Leaving System
a. Regenerator off gas 91.8 47.1
b. Combustor flue gas 39.5 20.3
oo c. Combustor bed reject 17.5 9.0
N>
d. Combustor overhead solids (flyash) 22.6 11.6
e. Regenerator overhead solids 2.0 1.0
Total 173.4 89.0
• Sulfur Accumulated (A Inventory)
a. Regenerator bed 3.1 1.6
b. Combustor Bed 25.1 12.9
Total 28.2 14.5
• % S Recovery 103.5
-------
if coal of a higher sulfur content was burned, and if the coal feed rate into
the combustor was higher. Runs planned for the future will use operating
conditions which should result in higher regenerator S02 levels.
Continuous operation of the miniplant provided a realistic way of mea-
suring the potential benefits of a regenerative system compared to a once-
through system. An important question is what reduction in sorbent require-
ments can be realized in a regenerative system. Since only one continuous
run has been made thus far, and at non-optimum conditions, this question
cannot be answered with a high degree of certainty. However, the average
SC>2 emissions from the combustor were 310 ppm (0.63 Ibs S02/106 BTU) at an
average Ca/S ratio of 0.55. Figure IV-6, based on data from a once-through
combustion system, shows that at least four times this makeup rate would have
been required in a once-through system, had the combustor been under calcin-
ing conditions. Actually, carbonating conditions prevailed in the combustor
and Figure IV-6 shows that the observed emission level could not have been
reached in a once-through system at any Ca/S makeup rate. Another way to
show the effect of regeneration on sorbent requirements is illustrated in
Figure IV-6. At the average Ca/S ratio of 0.55 used in this run, the
expected S02 emission from the combustor in a once-through system would have
been about 1150 ppm compared to 310 ppm measured.
Additional combined combustor-regenerator runs will soon be underway
to study the effects on performance of the important variables, including
sorbent recirculation rate, regenerator temperature, and type of sorbent.
It is hoped to determine which combustor and regenerator operating conditions
provide, simultaneously, low emissions of SQX and other pollutants from the
combustor, high concentrations of SC^ in the off-gas from the regenerator, low
makeup rates of fresh sorbent, and moderate recirculation rates between com-
bustor and regenerator.
In the future, the use of coal as a fuel for the regenerator will be
studied. The batch combustor and regenerator are being converted to con-
tinuous units to study this and other areas. One problem is recovering
sulfur from the fairly low levels of S02 present in the regenerator off-gas.
Other problems are sorbent deactivation and attrition. These problems need
to be understood and solved before regeneration can become a commercial
endeavor.
83
-------
SECTION V
GRANULAR BED FILTRATION STUDIES
INTRODUCTION
The successful development of the pressurized fluidized bed coal com-
bustion process is dependent on the ability of particulate control devices
to remove particulates from the hot combustor flue gas to very low levels.
This must be done to assure that the expansion of the flue gas through the
gas turbine does not cause damage to the turbine by erosion, corrosion, or
deposition of solids on the turbine blades. At the present time, the turbine
requirements are not well defined. Current estimates of the allowable parti-
culate concentration in the flue gas entering the turbine range from 45 to 1
mg/m3 (0.02 to 0.0004 grains/SCF). In addition to the gas turbine inlet
requirements, the U.S. Environmental Protection Agency has imposed limits on
the emission of particulates from coal fired installations of 0.043 g/MJ
(0.1 Ib/M BTU). For a typical coal, this standard translates to a parti-
culate concentration in the flue gas of approximately 100 mg/m^ (0.05 gr/SCF).
Therefore, at the present time, removal efficiencies are dictated by the
turbine requirements. To meet these estimated requirements, the flue gas
leaving a pressurized combustor must first be precleaned in a two stage
cyclone system and then be sent to a third stage cleanup device. To meet
the current environmental standard the third stage device will be required
to have an efficiency of approximately 67%. To meet the estimated turbine
requirements, efficiencies must be in the range of 95 to 99.7%.
The objective of the flue gas particulate removal program is to
evaluate two removal devices which have the potential for reducing parti-
culate loadings to the required levels. The devices will be installed on the
miniplant which has a maximum flue gas flow rate of about 20 Sm3/min (700
SCFM). The first device is a granular bed filter of a design developed by the
Ducon Company. The results of the initial testing to evaluate this device
will be covered in this section of the report. Further testing of the
granular bed filter is planned, and will be reported in a subsequent report.
The choice of the second device has not been made at this point.
The objectives of the granular bed filter test program were to measure
the outlet loading from the filter, determine if the removal efficiency was
maintained with use, measure operational stability of the filter (e.g., can
a low pressure drop across the filter be maintained, does the filter plug,
is the amount of blow back gas needed to maintain steady operation within
reason, etc.), and finally, to measure the long term life of the filter
hardware. The primary operating parameters were: the filtered gas flow
rate, usually measured as the gas velocity entering each filter bed, the
reverse flow clean gas ("blow back") velocity, the duration and frequency
of the blow back step, the type of filter media used (i.e., particle size,
shape, density), and media bed depth.
84
-------
EQUIPMENT, PROCEDURES
Equipment
Single Bed Test Rig—
To provide support information for the evaluation of the granular bed
filter, a single bed test rig was designed and tested. A photograph of the
test rig is shown in Figure V-l. A schematic of the test rig is shown in
Figure V-2. The test rig was designed to filter an isokinetic sample which
was extracted through a 1.6 cm (0.5 in) probe which was located in the
miniplant ducting downstream of the second stage cyclone. The test rig
consisted of one of the Exxon designed filter beds described in a subsequent
section enclosed in a refractory lined vessel to permit filtration at system
temperature and pressure. The gas exiting the test rig was filtered by a
glass microfiber filter to allow a measurement of particle loading in the gas
leaving the rig. (The particle loading of the gas entering the rig was
estimated based upon past measurements on the miniplant without the granular
bed filter.) The pressure drop across the filter bed was measured by a AP
cell. When blow back was required, as determined by an unacceptable pressure
drop, the sample flow was stopped and the filter vessel was vented to the
atmosphere. The blow back cycle was then initiated using air rates up to
0.14 Sm3/m (5 SCFM).
Original Ducon Filters—
Three filter elements were purchased from the Ducon Company for evalua-
tion. The Ducon filter element consists of a number of small beds packed
with a suitable filter media such as alumina, quartz, etc. and supported on
a bottom retaining screen. A stack of the filter beds form a single filter
element. A number of filter elements can be used depending on the volume
of gas to be filtered. A schematic of a single filter bed is shown in Figure
V-3. The operation of the filter is shown schematically in Figure V-4.
During the filtration cycle, dirty gas passes through the 50 X 50 mesh inlet
screens down into the filter beds. Clean gas from the beds is collected in
a manifold in the interior of the element and then passes to the clean gas
outlet system. As the filtration step proceeds, the pressure drop across
the element increases and eventually the element must be cleaned by the
reverse flow of clean gas. This "blow back" occurs by flowing clean gas in
reverse direction through the outlet gas manifold, up through each filter
bed and out through the screens. The function of the screens is to retain
the filter media during the blow back step, keeping it inside the filter
bed, while allowing the fine particulates removed from the filter media by
the blow back gas to pass through. The fine particulate then settles outside
the filter elements and is collected at the bottom of the shroud containing
the filter elements.
The nominal flow capacity of each of the three elements purchased
from the Ducon Company was 8.5 Sm^/min (300 SCFM). Each element was 20 cm
(8 in) in diameter by 1.8 m (6 ft) long and contained twelve beds having a
total filtration area of 0.24 m2 (2.6 ft^). A photograph of one of the
filter elements and the shroud in which it is contained when placed in the
pressure vessel is shown in Figure V-5. One of the Ducon elements was
designed to be blown back by short sonic pulses of high pressure air as
shown in Figure V-6. The pulse duration was approximately 0.5 sec. This
blow back method was not tested. The other two Ducon elements were blown
85
-------
FIGURE V-l
SINGLE BED TEST RIG
86
-------
FIGURE V-2
SINGLE BED TEST RIG ASSEMBLY
Off-Gas from
Secondary Cycle
00
Cooled Ball
Valve
Filter
Vent
Pressure
Gauge
(/\Pressure Gauge
HT NTyiFlow Control Valve
v Rotameter
•low
Control
Valve
Balston
Filter
Pressure
Gauge
"| Pressure
J Regulator
Air From
Auxiliary
Compressor
-------
INLET
RETAINING
SCREEN
CO
CO
BOTTOM
RETAINING
SCREEN
FIGURE V-3
SCHEMATIC OF A SINGLE DUCON FILTER BED
20.3 cm
2.7 cm -
H-* 3.8 cm
FILTER
MEDIUM
CLEAN GAS
o
CNJ
LO
4.4 cm
-------
FIGURE V-4
DUCON FILTER SCHEMATIC
Filtration Cycle
Blowback Cycle
Filter
Media"
ZZ2
Z22
22Z
(Lu.
222
\
LLLL
Clean
Gas Exit
LLLL
Dirty Gas
-Retaining Screen
7
Fly ash
1
Lock
Hopper
i
Lock
Hopper
Dirty Gas
Fluidized
Filter Media
89
-------
FIGURE V-5
DUCON FILTER ELEMENT AND SHROUD
90
-------
FIGURE V-6
EXXON FILTER SCHEMATIC
Filtration Cycle
Dirty Gas
Filter
Media
Retaining
Screens
Blowback Cycle
te
/t>
Clean Air
91
-------
back with a larger volume of sub-sonic air for longer durations. The intent
was to fluidized the filter beds rather than shock them with a short pulse
of high pressure air. This method is described in Figure V-4.
Original Exxon Filter—
A fourth filter element designed by Exxon was also fabricated and
tested. The Exxon filter element consisted of ten beds (instead of twelve)
and was designed to permit easy disassembly and removal of the 50 X 50 mesh
retaining screens. A schematic of the filter element and the arrangement
within its shroud is shown in Figure V-7. Figure V-8 is a sketch of an
individual bed. The blow back method was also different. The Exxon filter
element used a "positive blow back" technique. It was equipped with shut
off valves to allow it to be completely isolated from the feed and product
streams during the blow back. The element was then depressurized and blown
back with a larger volume of low pressure air.
Current System—
Discussion with the Ducon Company were held after operating problems
developed with the original filter design. The problems are discussed in
a subsequent section. These discussions led to the design and fabrication
of a third filter system. The current system consists of two filter ele-
ments, each having five filter beds as shown in Figure V-9. The filter
beds incorporate the modifications recommended by Ducon which included
removing the inlet retaining screens to prevent plugging and increasing
the freeboard of each bed to prevent entrainment of the filter media during
blow back. A schematic of one of the modified filter beds is shown in
Figure V-10. An annular fluidizing grid containing 56, 0.5 cm diameter holes
was also installed at the bottom of each bed beneath the bottom retaining
screen, as shown in a photograph of the clean gas side of the grid (Appendix
J-2) . The intent of the grid was to assure good distribution of the blow
back air. Dirty gas enters the bed through the opening below the top
flange and passes downward through the filter bed and out into the clean
gas outlet tube in the center of the element. During blow back, the blow
back air passes up through the fluidizing grids supporting each bed, fluidizes
the beds and blows the fine particulates out through the inlet slot. A 18 cm
(7 in) freeboard above the filter beds acts as a disengaging section for the
filter media and prevents its entrainment through the outlet slot.
Pressure Vessel, Internal Piping—
The granular bed filter (GBF) was installed on the miniplant between the
second stage cyclone and pressure control nozzle. The piping from the
existing miniplant off gas line to the GBF is 30.5 cm (12 in) carbon steel
pipe. It is refractory lined with Grefco 75-28 to an internal diameter of
10.2 cm (4 in) and is lined inside with a 4 inch Schedule 5 stainless steel
pipe. The return line from the GBF outlet is 20.3 cm (8 in) diameter carbon
steel pipe also refractory and steel lined to a 10.2 cm (4 inch) diameter.
The pressure vessel housing the filter elements consists of a refractory
lined vessel approximately 2.4 m (8 ft) in diameter by 3.4 m (11 ft) high.
Figure V-ll is a photograph of the filter vessel and shows the size and
structural relationship of the GBF to the rest of the miniplant. The pressure
vessel is supported on a structure which is 5.6 m (18.5 ft) by 4.6 m (15 ft)
and 5.9 m (19.25 ft) high. Appendix J-3 and J-4 are drawings showing the
vessel dimensions and the locations of the various access ports. Appendix
J-5 is a closeup photograph of the pressure vessel. The piping arrangement
is shown schematically in Appendix J-6.
92
-------
FIGURE V-7
DUCOIM SONIC FILTER SCHEMATIC
Filtration Cycle
Filter
Media
X
Clean
A
(/
j(
•24.
\^M
• i _j
M
T2lt.
'/'}_
rWV
LLlj.
\Z22j.
I
\
/
\
k
OLJ.
LlM.
LLL^
LtM.
"ZLU
f>i~i
u&
///7
7/77
\77j A
bas txi
^ Dirty Gas
Retai
•"""" Scree
Blowback Cycle
Expanded
Filter Media
93
-------
FIGURE V-8
SCHEMATIC OF A SINGLE EXXON FILTER BED
21.3cm
•* 6. 6cm H-*
Dirty
Gas
Top Retaining
Screen
Filter
Medium
Clean Gas
Bottom Retaining
Screen
-------
FIGURE V-9
MODIFIED EXXON FILTER ELEMENT
95
-------
Dirty
Gas
Inlet
I I
FIGURE V-10
MODIFIED FILTER BED
Clean Gas
Outlet
I I 1,6cm
14.3cm
Fluidizing Grid
(56-0.36cm dia.
I | holes)
6.4cm
21.3cm
96
-------
FIGURE V-ll
GBF PRESSURE VESSEL AND STRUCTURE
Blowback Plungers
ntrol Valves
97
-------
Access to the interior can be made through a 70 cm (27 in) manhole The
vessel can hold up to four filter elements installed through four flanges at
the top of the vessel as shown in Figure V-ll. Each filter element is con-
tained within a shroud in the inside of the pressure vessel. Inlet gas is
piped to each shroud, passing through a measuring orifice to determine the
flow rate to each filter element. This is shown in Figure V-12. Clean
gas exits from each shroud through openings at the top (Figure V-12) and
fills the interior of the pressure shell. Blow back air enters each filter
element through the top flanges of the pressure vessel (Figure V-ll) and
flows in reverse direction through each filter element. Particulates removed
from the filter element during blow back impinge on the inside surface of the
shroud, fall to the bottom and are collected in lock hoppers. The blow back
gas leaving a filter element flows in reverse direction through the inlet gas
system into the other filter elements which are in the filtration mode. Each
element is blown back separately.
Startup Burner, Natural Gas Injection —
A natural gas burner was installed to preheat the interior of the pres-
sure vessel to a temperature above the dew point of the combustor flue gas
before starting a filtration test. The burner is a self-contained, direct-
fired air heater capable of burning natural gas to give heat input rates of
up to 146,500 W (500,000 BTU/hr) . A schematic of the burner flow systems is
shown in Appendix J-7. Burner operations are automatically controlled as
are all the required startup sequencing and emergency shutdown procedures.
The burner fires into the vessel through a side port for an 8 to 12 hour
period prior to the start of a run. The filter vessel is at atmospheric
pressure during this period.
Natural gas injection into the flue gas ducting between the miniplant
second stage cyclone and the granular bed filter was found necessary to
maintain the flue gas temperature above 843°C (1550°F) at the filter inlet.
Gas was injected at four points through injection probes located 2.3 m
(7.5 ft), 6.7 m (22 ft), 9.8 m (32 ft) and 14.6 m (48 ft) downstream of the
second stage cyclone. Gas flows of approximately 0.03 s m3/min (1 SCFM)
through each probe was found effective in maintaining the gas temperature
at the desired level.
Operating Procedures
To prevent condensation during startup, the pressure vessel housing the
filter elements vessel was preheated for a period of 8 to 12 hours. After
the preheat period, startup activities included calibration of all AP trans-
mitters, turning on purge air flows for the pressure taps, and turning on
the blow back air compressor. The proper blow back air pressure and flow
were then set. A schematic of the GBF blow back air supply is shown in
Figure J-9. At this point, the miniplant combustor was started up and flue
gas was sent to the granular bed filter. During the filtration cycle, the
pressure drop across the filter vessel and the flow to each element were
continuously monitored. When the pressure drop increased to an unacceptable
level (generally, 14 kPa above the baseline pressure drop), the blow back
cycle was initiated. Blow back was accomplished by stopping the flow to
one filter element by engaging a blow back nozzle and seal P^te and blowing
back with air at a pressure slightly above the ^ration P'^re' e
other elements would then pick up the additional fxltration load.
98
-------
FIGURE V-12
GBF PRESSURE VESSEL INTERIOR
99
-------
filter elements were blown back consecutively and all operations were con-
trolled from the miniplant panel board. After a number of filtration/blow
back cycles, the particulate collection lock hoppers were emptied.
EXPERIMENTAL RESULTS AND DISCUSSIONS
Test Rig Preliminary Results
The granular bed filter test rig was successfully operated during mini-
plant runs 46.1 and 46.2. The primary intent of these tests were to estab-
lish operating procedures and to determine, at least qualitatively, how readily
the fly ash could be cleaned from the filter bed. A 6.4 m (2.5 in) deep
bed of granular quartz was used during the test with a particle size ranging
between 300 and 600 microns with 50 weight percent being finer than 400
microns. This particle size distribution was selected as a compromise
between higher collection efficiency at a finer particle size and support
screen plugging problems associated with too fine a particle size. The
filtration velocity during these tests was approximately 6.1 m/min (20
f t/min) . The blow back duration ranged between 3 and 10 min in these pre-
liminary tests and the blow back air superficial velocity ranged between
0.18 and 0.34 m/sec (0.6 and 1.1 ft/sec). Some of the more important con-
clusions of the test are as follows:
1. The granular bed was capable of filtering fly ash.
2. As shown in Figure V-13, the bed exhibited a linear increase
in AP with time. An analysis of the data indicates that the
AP across the bed increased at a rate of 0.4 kPa/min
(1.6 in H20/min).
3. Although blow back conditions were more favorable than can
be afforded in a practical cycle, the AP across the bed was
always able to be brought back close to its baseline value
of about 1 kPa. The differences in the filter pressure
drop after each of the blow backs was caused by differences
in blow back conditions.
4. Extrapolation of the test results indicated that to maintain
the AP across the bed below the design value of 14 kPa (2 psi),
a cycle time of approximately 20-35 minutes was required
between blow backs.
5. Although the filter collection efficiency was not measured,
an inspection of the backup filter indicated that most of
the fly ash was being trapped by the granular filter.
6. Inspection of the filter bed showed no signs of interraction,
sticking or agglomeration of the fly ash and the filter media.
The fly ash was easily removed from the quartz particles.
Operation of the filter was also attempted during ^J1^ ™ *6'3
and 46?4 but was not successful due to condensation of moisture in the
filter vessel. The initial pressure drop across the bed was well above
100
-------
FIGURE V-13
7.5
5.0
ro
Q-
2.5
0
TEST RIG RESULTS - AP VS. TIME
SAND - GRANULAR QUARTZ
PARTICLE SIZE RANGE - -30+50 MESH
BED DEPTH -6.4 cm
i i i » i i
' '
J L
'
/
^
0 10 20
i
30 40 50 60 70 80 90 100 110 120
TIME (MIN)
-------
the design value of 14 kPa (2 psi) and the increase with time was much
greater than during the initial tests. Inspection showed plugging of screens
with caked carryover material, possibly precipitated by presence of Listure
Based on these observations, every precaution must be taken to assure that
the filter is not operated below the dew point of the flue gas.
Testing of Original Ducon and Exxon Filters
After installation of the filters, shakedown began with ambient tem-
perature testing of the system.
The objectives of these preliminary tests were to (1) check combustor
pressure control with the filter on line, (2) pressure test the system, (3)
check the alignment of the blow back nozzles, (4) check out the operation
of the blow back flow system, and (5) measure the distribution of flow to
each one of the filter elements. A number of mechanical problems were dis-
covered (leaks, misalignments, etc.) and had to be corrected before further
testing could be resumed.
Two high temperature runs were then attempted but the pressure drops
across the filter became extremely high and all attempts at blow back were
unsuccessful. Inspection of the filter elements after each of these runs
showed that a hard filter cake had formed on the inlet retaining screens.
This is shown in Figure V-14. Note in Figure V-14 that the particles are
adhering to all surfaces, not only to the inlet screens, although the cake
is thicker over the screens. The filter medium was usually particulate
free indicating very little penetration through the screens. Since these
initial runs were made before the preheat burner was installed it was orig-
inally thought that the plugging occurred during startup when moisture could
condense before the filter vessel had come up to temperature. The preheat
burner was later installed and a run (58) was made to re-evaluate the Ducon
filter. The same screen plugging problems occurred before even one blow
back could be successfully completed, and the original Ducon filter was
deemed to be unacceptable for our application.
Runs using the Exxon designed filter also proved unsuccessful. Some
screen plugging was in evidence but it was the inability to seal the blow
back nozzles which caused the most problems. Proper engagement and sealing
of the blow back nozzles were necessary in order to isolate the filter from
the system so that depressurization and blow back could be initiated.
Evaluation of the Exxon filter was discontinued at this point.
Modifications
At this point, a meeting was held with personnel from Ducon to discuss
the problem of screen plugging. The discussions led to the design and
fabrication of a third filter system. Ducon indicated that they had
encountered the same screen plugging problem with fly ash and prevented it
by removing the screens and designing the individual beds with more free-
board to prevent entrapment of the filter media during blow back. It was
also recommended that a fluidizing grid be used at the bottom of the beds
to assure good distribution of the blow back air. The use of an ejector to
replace the sometimes troublesome plunger type blow back nozzles was also
suggested. A detailed description of the modified filter was given in an
102
-------
FIGURE V-14
DUCON FILTER WITH PLUGGED SCREENS
103
-------
earlier section Filter elements incorporating these suggestions were fab-
ricated and shakedown continued using two of these elements each of which
contained five filter beds. An ejector system for blow back was also designed,
but was not tested during the period covered by this report.
Testing of Modified System
Operability of the modified filter system (not including the ejector)
was demonstrated. The end of the initial shakedown phase was signified by
the completion of a 24 hour run (Run 59) . The ability to filter and blow
back, the ability to maintain low pressure drops, and the ability to collect
particulates after blow back were demonstrated. Collection efficiencies of
over 90% were calculated during the initial portion of the run based on
measured outlet particulate concentrations of about 70 rag/m3 (0.03 gr/SCF).
The particulate loading in the outlet from the secondary cyclone had been
measured a number of times prior to the granular bed filter runs and had
averaged about 2300 mg/m3 (1 gr/SCF). It was assumed that this was the inlet
loading to the filter during these tests. In the future, provisions will be
added to sample particulates at the inlet and outlet of the filter simulta-
neously. The particulate escaping the granular bed filter had a weight
median particle size of 4 microns. A complete size distribution is shown
in Table III-8. Stable operation for up to 24 hours was also demonstrated
with no significant increase in baseline pressure drop across the filter.
Blow back was usually required every 10-20 minutes during which time the
filter pressure drop had increased by 14 kPa (2 psi) above its baseline
value. A range of blow back conditions were used to restore the baseline
pressure drop. Blow back durations ranged between 2 and 30 seconds and
superficial velocity between 0.15 and 0.75 m/s (0.5 and 2.5 f t/s) . Filtra-
tion velocities generally ranged between 18.3 and 24.4 m/min (50 and 80
f t/min) . Filter media consisting of 300 to 600 micron quartz particles were
tested. The quantity of blow back air used ranged from 1 to 5% of the
filtered gas rate.
A number of problem areas were defined during the shakedown portion of
the program. Demonstrated particulate outlet concentrations were still
higher than the tentative turbine inlet requirements, although the -lowest
levels measured to date are only slightly above the upper limit of the ten-
tative target range. Firm turbine requirements have not been set and it is
too early to judge the suitability of the filter to protect gas turbines.
However, at times, the filtration efficiency was very poor and the outlet
particulate concentrations were as high as 700 to 1200 mg/m3 (0.3 to 0.5 gr/
SCF) . It was also observed that the efficiency decreased with time in some
of the longer runs, including the 24-hour run, dropping from 90% initially
to about 50% later in the run. Loss of filter media during blow back was
another reoccurring problem during shakedown. Since inlet retaining screens
were found to be susceptible to plugging, screens probably cannot be used
and better control of the blow back air supply must be established to minimize
these losses. A significant buildup of particulates in the filter beds was
also observed amounting to about 30% of the weight of the filter media. A
noted during any of the shakedown runs.
104
-------
It was also observed that the particulates were not only building up
in the beds, but were uniformly mixed with the filter media. It is possible
that the buildup and mixing of particulates in the bed could be responsible
for the increase in the particulate concentration in the outlet gas with
time. The cleaning of the filter media by blow back must be improved.
Another potential problem with the current design is its vulnerability
to upsets. If upsets occur, such as bed plugging or loss of filter media,
the operating problems caused by such upsets usually require shutdown of the
system. It is usually not possible to take corrective action which restores
good operation. Another problem which may be unique to the Miniplant was
the interaction of the granular bed filter with the rest of the FBC system
during the blow back cycle. An increase in system pressure was noted during
blow back resulting in problems with the coal feed system which is controlled
by the differential pressure between the coal feed vessel and combustor.
This required modifications to the coal feed control system to minimize the
effects.
Test Details—
A detailed description of the runs made to date follows. Table V-l
summarizes the filter operating conditions for the five high temperature runs
made to evaluate the performance of the modified filter system. Prior to
these runs, a series of ambient temperature runs was made using fly ash and
talc particulates injected into a compressed air stream. Some indication
of the effect of blow back conditions on restoring the bed pressure drop was
obtained. An 8 s blow back at a velocity of 0.5 m/s seemed adequate.
However, loss of filter media during blow back and outlet screen plugging
occurred during the runs and the significance of the results was uncertain.
Miniplant run 54 was the first high temperature run using the modified
filter elements. The combustor was operated at a temperature of 910°C
(1670°F) and at a pressure of 595 kPa (86 psia). The combustor air flow
rate was set to give a filter face velocity of 25 m/min (83 ft/min). Coal
was burned for a total of five hours and the pressure drops across the two
filter elements were approximately 14 kPa (2 psi) just prior to the start
of coal combustion and were allowed to build up to as high as 41 kPa (6 psi)
during the first hour of coal combustion. The blow back cycle was relatively
successful and was able to reduce the pressure drops to approximately 21 kPa
(3 psi). The filter elements were blown back at superficial velocities of
0.34-0.49 m/sec (1.1-1.6 ft/sec) for durations ranging between 8 sec and
30 sec. However, in the latter part of the run, the pressure drops across
the filter elements just after blow back began to increase indicating a
decline in blow back efficiency. When the run was ended, the pressure
drops across the filter elements were approximately 69 and 110 kPa (10
and 16 psi).
During the latter part of the run a flue gas particulate sample was
taken downstream of the GBF. A grain loading of 1570 mg/m3 (0.69 gr/scf) was
measured, giving an overall filtration efficiency of approximately 20-30%.
Moisture condensed in the filter during heat up, since the preheat burner
had not been installed before this test was run, causing some of the par-
ticulates to agglomerate in the filter. This probably influenced the
effectiveness of the blow back and the filtration efficiency.
105
-------
TABLE V-l. GRANULAR BED FILTER RUN SUMMARY
FOR MODIFIED FILTER ELEMENTS
Filter Description
Run 54 (6/29/77)
Run 55 (7/14/77)
Run 56 (7/26/77)
Number of Elements
Number of Beds/Elements
Inlet Retaining Screens
Filter Medium
Bed Depth (cm)
Filter Medium Part. Size (ym)
Operating Conditions
Filter Vessel Preheat Temperature (°C)
Filter Vessel Inlet Temperature (°C)
Filter Vessel Outlet Temperature (°C)
Pressure (kPa)
Filtration Velocity (m/min)
Baseline AP (kPa)
AP Before Blow Back (kPa)
Run Length (hrs)
Blow Back Conditions
Superficial Velocity (m/s)
Duration (s)
Interval Between Blow Backs (min)
Particulate Emissions
Measured Part. Concentration (mg/nH)
Estimated Removal Efficiency (%)
2
5
None
Quartz
3.8
250-600
No Preheat
815
540
590
25
14 (70-100 at end)
40
5.5
0.34-0.49
8-30
5-10
1570
30
2
5
None
Quartz
3.8
250-600
450
740
510
570
21
20 (35-40 at end)
50
1.5
0.55-0.73
15-30
5
Not Measured
2
5
None
Quartz
3.8
250-600
450
760
510
800
18
20
50
2.5
0.76
8
5-10
Not Measured
-------
TABLE V-l (Continued). GRANULAR BED FILTER RUN SUMMARY
FOR MODIFIED FILTER ELEMENTS
Filter Description
Run 57 (8/2/77)
Run 58 (8/5/77)
Run 59 (8/11/77)
Number of Elements
Number of Beds/Element
Inlet Retaining Screens
Filter Medium
Bed Depth (cm)
Filter Media Part. Size (ym)
Operating Conditions
Filter Vessel Preheat Temperature (°C)
Filter Vessel Inlet Temperature (°C)
Filter Vessel Outlet Temperature (°C)
Pressure (kPa)
Filtration Velocity (m/min)
Baseline AP (kPa)
AP Before Blow Back (kPa)
Run Length (hrs)
Blow Back Conditions
Superficial Velocity (m/s)
Duration (s)
Interval Between Blow Backs (min)
Particulate Emissions
Measured Part. Concentration (mg/m3)
Estimated Removal Efficiency (%)
2
5
None
Quartz
6.4
250-600
650
860
700
815
18
28
55
6
0.46
8
10
40-180
98-92
2
12
50 X 50 mesh
Quartz
3.2
250-600
Original Ducon
system used. Run
unsuccessful because
of screen plugging.
2
5
50 X 50 mesh
Quartz
6.4
250-600
650
860
740
790
26
28
55
23
0.46
8
10, 60-90
180,640,1230
92, 72, 47
-------
Miniplant run 55 was also made at a low pressure because of blow
back air supply limitations. The natural gas preheat burner was used for
the first time to preheat the granular bed filter pressure vessel above the
§aS h^H n^r'rftSO^rrr W°rked WeU ^ the ^"t-e within the vessel
reached 454 C (850 F) before particulate laden gas was passed through the
filter elements. Baseline pressure drops of 21 kPa (3 psi) were measured
across the filter elements during the combustor preheat. During the combustor
preheat period, in which kerosene was burned in the combustor, the pressure
drop across the filters increased 24-28 kPa (3.5 - 4 psi) every 30 minutes.
Blow backs at 0.73 m/sec (2.4 ft/sec) superficial velocity were successful in
restoring the filter AP to the original 21 kPa (3 psi). The collected par-
ticulate had a blackish coloration indicative of collection of oil soot and
attrited sorbent.
During a 1-1/2 hour period burning coal, pressure build up was more
rapid necessitating blow back approximately every 5 minutes (Figure V-15).
The filtration velocity during this period was 27.6 m/min (90.6 ft/min).
Equal flow distribution was not achieved during the coal burning phase, one
of the elements constantly passing more gas than the other. The collected
particulates had the characteristic red-brown coloration of a mixture of
coal, fly ash and attrited sorbent. The quantity of particulates collected
in the lock hoppers amounted to about 50% of the anticipated particulate
loading to the GBF.
The filter elements were inspected after the run and considerable
build-up of particulates in the individual beds was evident. The particulate
sampling system was not in operation during the run and the filtration
efficiency was not measured. However, the clean gas outlet tubes and the
interior of the GBF pressure vessel showed little evidence of particulate
pass through. The outlet retaining screens were quite clean and showed no
evidence of plugging.
Miniplant run number 56 was an attempt to operate the granular bed
filter system at the full operating pressure of 928 kPa (135 psi). This run
was attempted after modifications were made on the blow back system to
increase the capacity of the system in terms of pressure and flow. Coal was
burned for almost 2-1/2 hours and the pressure drop across the filter was
able to be controlled between 21 and 48 kPa (3 and 7 psi) with 8 second
duration blow backs at a superficial velocity of 0.76 m/s (2.5 ft/s) every
5-10 minutes. The filter operations appeared to be successful and the run
was only terminated when a fire developed in the coal injection vessel.
However, upon inspection after the run, it was found that most of the filter-
ing media had been lost from the filter beds during the blow back cycle
probably because of a excessively high blow back air flow. Again, no parti-
culate measurements were made and the filtration efficiency was not deter-
mined .
Miniplant run number 57 was a repeat of the previous run with emphasis
on maintaing control of the blow back velocity since bed losses during blow
back were appreciable during run 56. Each filter bed was 6.4 cm (2-1/2 in)
deep and consisted of 250-600 micron diameter quartz particles. An approach
velocity of the gas to the filter elements of 18.3 m/min (60 ft/mm) was
measured.
108
-------
FIGURE V-15
EFFECT OF BLOWBACK ON FILTER PRESSURE DROP
RUN 55
g
co
0.
Element 1
Element 2
TIME (MIN.)
-------
The blow backs were made at 10 minute intervals using a pulse duration
of 8 seconds at a superficial velocity of 0 46 m/Q M •? f?/«, ? n • !u
filtration period between blow backs/the p^ssu^ £ 2 ^iheTiSr^
beds increased from the baseline level of 28 to 55 kPa (4 to 8 pal). The
profile of the bed pressure drop was very reproducible. During the 6 hours
of the run, collection of the particulates captured by the filters was
achieved in the lock hoppers. Although filter medium loss did occur, it was
not as appreciable as in run 56. Retention of a significant amount of
captured particulates in the filter beds was also apparent. A particulate
sample of the off gas from the GBF was obtained during this run. The parti-
culate loading was measured to be 187 mg/m3 (0.08 gr/SCF), which represents
a collection efficiency between 90 and 95%.
Miniplant run 59 was made after 50 X 50 mesh inlet retaining screens were
installed on the modified filter elements to prevent loss of filter media.
The objectives of the run were to: (1) demonstrated GBF operations over a
24 hour period, (2) observe any changes in baseline AP over a number of blow
back cycles, (3) determine again if inlet retaining screens could be used
to prevent loss of the filter medium during blow back without plugging, and
(4) measure temperature profiles from the combustor bed through the filter
vessel.
The GBF was continuously operated for a total of 28-1/2 hours during
which coal was burned for 23 hours. The baseline pressure drop across the
filter elements was measured to be 28 kPa (4 psi) and was allowed to build up
an additional 14 kPa (2 psi) before blowing back. Filtration was done at
an approach velocity of 25.6 m/min (84 ft/min) . Blow backs at a superficial
velocity of 0.46 m/sec (1.5 ft/sec) for an 8 second duration were successful
in restoring the AP to the baseline value.
The run appeared to have two distinct segments as far as particulate
removal was concerned. During the first 8 hours of operation on coal, the
time between blow backs was 8-10 minutes. During a one hour particulate
sampling period, a particulate concentration of the gas leaving the GBF of
180 mg/m3 (0.077 gr/SCF) was measured. This loading was comparable with that
measured during run 57. At this point in the run, the coal feed was inter-
rupted for a 4 hour period during which kerosene was burned to maintain the
combustor bed temperature. No blow backs were made during this period.
During the last 14 hours on coal, the time between blow backs increased to
between 1 and 1-1/2 hours. In an 9 hour particulate sampling period, a
loading of 653 mg/m3 (0.28 gr/SCF) was measured. During the last 2 hours of
the run, another particulate sample gave a loading of 1260 mg/mJ (0.54 gr/SCF),
On inspection of the filter elements after the run, it was seen that the
inlet retaining screens were partially plugged. Figure V-16 is a photo-
graph of one of the filter beds after the run which clearly illustrates the
nature of the problem. It appeared that the gas was entering the bed through
a narrow slit between the screen and the flange of the next bed A consider-
able amount of particulate was retained in the filter beds in the form of a
hard filter cake. The fact that rat holing had occurred was apparent. The
filter cake probably formed during the period when kerosene was burned and
the subsequent low measured collection efficiencies resulted because of the
rat holing.
110
-------
FIGURE V-16
MODIFIED EXXON FILTER WITH PLUGGED INLET RETAINING SCREENS
111
-------
Pressure control in the entire FBC system was susceptible to upsets
during blow back pulses. The differential pressure between the coal injec-
tion vessel and the combustor was easily upset. Three fires occurred in the
injection vessel as a result of pressure upsets which caused hot bed solids
to flow back into the injection vessel. Alternate blow back procedures
or modifications to the coal feed system must be developed in order to
minimize these upsets.
Natural Gas Injection Tests—
Natural gas injection into the flue gas line between the second stage
cyclone and the GBF was investigated as a means of maintaining the flue gas
temperature above 925°C (1700°F) before entering the GBF. During an initial
test, natural gas was injected at a single point 2.3 m (7.5 ft) downstream
of the second stage cyclone and the effect on the GBF inlet line temperature
profile was observed. As seen in Figure V-17, the GBF inlet was able to
be increased from 749°C (1380°F) to 829°C (1525°F) but a sharp temperature
rise near the injection nozzle caused by the instantaneous combustion of
natural gas was also observed. To minimize this temperature rise, natural
gas was later injected at three additional points but keeping the total
natural gas flow the same as in the single injection point test. This
technique was successful in maintaining the GBF inlet at 925°C (1700°F)
without any sharp temperature rises. A temperature profile is shown in
Figure V-17.
Blow Back Gas Ejector Design—
A blow back gas ejector was designed to replace the blow back gas inlet
plunger and seal plate assembly. Figure V-18 is a sketch of the ejector
system. The motive gas is air compressed to 2550 kPa. The ejector system
was not tested during this reporting period but is scheduled to be tested in
the future.
112
-------
o
o
L±J
Cd
LJ
O.
910 -
880 -
850 r
820 -
FIGURE V-17
FLUE GAS TEMPERATURE PROFILE
INJECTION
PT. 1
700
- 1650
- 1600
- 1550
m
m
- 1500 -j
- 1450 -n
- 1400
- 1350
1300
INJECTION
PT. 2
INJECTION
PT. 3
DISTANCE DOWNSTREAM OF
SECOND STAGE CYCLONE (m)
INJECTION
PT. 4 GBF
INLET
-------
FIGURE V-18
SCHEMATIC OF THE EJECTOR BLOW BACK SYSTEM
High Pressure
Ejector Motive Gas
Clean Gas
Outlet
Ejector I
Secondary Gas ^*- ,
Ejector
Filter
Shroud
114
-------
SECTION VI
MODIFICATION OF THE BATCH UNIT
The combustor and regenerator sections of the batch unit were modified
to permit continuous operation. Prior to these modifications they could only
be operated in a batch or semi-batch manner. Operation of the unit in the
continuous mode will increase its flexibility and enable the unit to provide
the type of data which can only be obtained from continuous units such as
the miniplant.
The modifications will permit the continuous feeding of both coal and sor-
bent, and the continuous removal of solids from the combustor. In the past the
facilities permitted only the continuous charging of coal to the combustor.
Sorbent was changed at the beginning of the run, with no provision for with-
drawal or makeup. The regenerator can now be operated continuously. Pre-
viously, the regenerator was operated only in a batchwise manner. In addition
there were no provisions for recycling the primary cyclone's solids to the
combustor. This has an adverse effect upon combustion efficiency, as the
first cyclone's solids contain an appreciable amount of unburned coal "fines."
The modified system now has provisions both to recycle or not and will permit
the measurement of combustion efficiency with either method of operation.
In the discussion that follows the modified facilities will be described.
Facilities retained from the batch unit were discussed in previous reports
(1,9) and will not be described in detail.
COMBUSTOR SECTION
Combustor Vessel
A schematic flow diagram of the modified combustor and the off gas and
solids flow are shown in Figure VI-1.
The combustor was constructed from four sections of 25 cm (10 inch) dia-
meter standard wall carbon steel pipe, lined with Grefco 75-28 refractory to
an inside diameter of 11.4 cm (4.5 inches). The height of the vessel above
the fluidizing grid is about 4.9 m (16 ft). Below the grid is a 61 cm (24
inch) burner section, lined with Grefco Bubblite Refractory. A 4 inch ID
hand hole is provided directly above the grid to facilitate the complete
removal of solids after a run. The preheat burner was described previously
(1).
A water cooled fluidizing grid is inserted between the burner section
and the first combustor section. The grid supports the static bed and pro-
vides a uniform flow of fluidizing air to the bed during operation. The
fluidizing air enters through 80-0.16 cm holes in the grid.
115
-------
FIGURE Vl-l
SCHEMATIC OF MODIFIED BATCH UNIT
Combustor
Shell
Sorbent
Feed
Hopper
\y
Refrac-
tory —
Lining
MOVs
1st
M
Solids
Overflows
.^ Grid
Cooling
Water
Burner
^, Off Gas Precooler
*--
- Cooling Water In
*•
^^ Off Gas Cooler
Back Pressure
Regulator
off X,
Gas
Filter
X
To Scrubber
To Analytical
Train
Solids
Overflow
Hopper
-------
Three sets of vertical cooling coils are located within the combustor
which control combustor temperature. The coils are made of 6.4 mm 316 SS
tubing with a surface area of 0.06 m2 per coil. The cooling coils are supplied
with dimineralized water, and the flow rate is controlled to produce a steam/
water outlet mixture. High flow rates can be used to prevent steam formation
when heat transfer coefficients are measured.
Two flanged inlets are welded into the side walls of the first section
above the grid at an angle of 60° to the horizontal (well above the angle
of repose for sorbents and fly ash). These serve as a sorbent charging line
and as a return line for solids from the primary cyclone. The sorbent
charging line is 1-1/2 inch IPS Sched. 40 and the primary cyclone return line
is 4 inch IPS, Sched. 40, CS refractory lined to a 5 cm (2 inch) ID. Both
are projected through the refractory and terminate about 0.013 cm (1/2 inch)
above the grid. Introduction of the solids immediately above the grid
improves mixing of the solids.
The second and third sections above the grid have outlets 0.46 cm (18
inches) long and inclined at 60° to the horizontal welded into three side
walls. The outlets are 4 inch IPS Sched. 40 CS, refractory lined to a 5 cm
(2 inch) ID. Both are projected through the refractory to the combustion
zone. These serve as solids overflow lines. The projection of the lower
outlet terminates 1.09 m (43 inches) above the grid, and provides a bed
volume of 0.0111 m3. The projection of the upper outlet terminates 1.85 m
(73 inches) above the grid, and provides a bed volume of 0.0189 m3. The
overflow lines discharge to a solids overflow hopper.
Combustor Off Gas System
The off gas system may be classified into three major components; they
are:
• Primary and secondary cyclones
• Off gas coolers
• Off gas filter and back pressure regulator
Cyclones—
Both cyclones have been rebuilt in order to increase their efficiency.
They were cast with Resco RS17-E refractory with the aid of wood mandrels,
which were "burned out" with a torch after the refractory had hardened. The
refractory is highly abrasion resistant. The gas discharge lines, at the top
of the cyclones, are made of 316 SS in order to resist corrosion. Details
may be found in Figure VI-2.
The primary cyclone will recover unburned coal and some fly ash and
sorbent from the off gas and return it to the combustor or to a receiving
hopper. The solids return line is fabricated with a length of 2 inch flexible
hose that may be connected either to the combustor or to a receiving hopper.
117
-------
FIGURE VI-2
BENCH UNIT CYCLONE ASSEMBLY
UIQ1T
EXXON RESEARCH AND ENGINEERING COMPANY
MECHANICAL DIVISION
LINDEN. N. J.
1633-I&-C
-------
The secondary cyclone will remove the remaining fly ash and sorbent from
the off gas in order to minimize the load on the off gas filter. The small
volume of removed fly ash will be collected in a standpipe and discarded
periodically.
Off Gas Coolers—
The off gas coolers are of the single-pass, water cooled, counter-current
flow double pipe type and are mounted vertically. The inner tube (gas flow)
is 1 inch IPS Sched. 80, 316 SS, the outer shell (water flow) is 2 inches
IPS Sched. 40, CS.
The off gas first flows through a precooler intended to reduce the off
gas temperature from about 760°C (1400°F) to approximately 370°C (700°F). A
second heat exchanger (the off gas cooler) is used to reduce further the off
gas temperature from 370°C (700°F) to 150°C (300°F). The precooler has a
fixed heat exchange area (gas side) of 0.18 m2 (1.9 ft2). The off gas cooler
has provision for drawing off the cooling water at nine locations along the
shell. This results in a variable area (gas side) ranging from 0.023 m2
(0.225 ft2) to 0.268 m2 (2.88 ft2). The variable area cooler will provide pre-
cise control of off gas temperature at all combustor operating conditions.
This will insure that the temperature of the off gas does not fall below
its "dew point" regardless of the system's operating conditions.
Coal Feeder
The coal feeder vessel previously used has been replaced with one of
larger capacity. The new coal feeder vessel has a capacity of 0.184 m3 (6.5
ft3). This will permit the storage of 109 kg (240 Ibs) of coal and 52.7 kg
(116 Ibs) of sorbent. Solids feed rate will be about 13.6 kg/hr (30 Ibs/hr)
and will permit eight to twelve hour runs without recharging the vessel.
The control system previously employed will be retained for the new coal
feeder. As in the past, coal will be injected into the combustor directly
above the grid. In order to insure adequate mixing of the coal with the
contents of the combustor bed, the incoming coal-air stream will be surrounded
by a stream of high pressure boost air at sonic velocity as done in the past.
Details of the coal injector may be found in Figure VI-3.
Sorbent Charging and Removal Systems
The charging of a premixed coal and sorbent mixture to the combustor
would be the optimum way in which to charge sorbent to the combustor.
However, this may not be feasible because the coal feeder orifice may^be
too small to permit the passage of sorbent. Any major increase in orifice
size would have an adverse effect upon the control of coal feed rates. This
possibility will be studied immediately after the "shakedown" of the com-
bustor has been made. If simultaneous coal-sorbent feeding is not feasible,
sorbent will be fed using a pressurized lock hopper in short and intermit-
tent intervals. See Figure VI-1.
119
-------
S3
O
PC
p
a <
g
I
1. l\\_\_ 1-AR.T
i. V
\mi,t>w.i
i i
COAL IMJE.CTOR V£-V5>E-L-.
BATCH UKJlT
EXXON RESEARCH AND ENGINEERING COUPANV
MECHANICAL DIVISION
LINDEN, tt.j
-------
Intermittent charging of fresh sorbent to the combustor bed can produce
substantial changes in the bed composition unless the charging is carefuMy
controlled Frequent addition of small amounts of sorbent will limit the
changes in bed Composition to acceptable values. An upper acceptable limit
of 5/. changes in the bed composition during the addition of fresh sorbent
has been set. Two motor operated valves (MOV) spaced 30 cm apart on the
discharge line will meter the incoming sorbent; the valves are operated
by a cycle timer. For a bed height of 109 cm (small bed) each sorbent charge
will introduce 5/, of new material to the bed; for a 185 cm bed height (large
bed) each sorbent change will introduce 3% of new material to the bed.
Valve cycle times are based upon the maximum coal firing rate (14 kg/hr),
a 5% sulfur coal, and dolomite as the sorbent with a Ca/S ratio of 3.0; these
are expected to be the most severe operating conditions that will be encoun-
tered. Under these conditions 39 cycles per hour will be required. Milder
operating conditions, such as a coal firing rate of 9 kg/hr, 2% sulfur coal
and a Ca/S ratio of 1.0, would require only 4 cycles per hour. 1-1/2 inch-
carbon steel-full bore ball valves will be used to meter solids flow. Valve
ball and seats are stellite faced.
Bed inventory (or volume) will be controlled by overflow from the top
of the bed to a receiving hopper of the same volume as the charging hopper.
Bed heights of 109 cm (43 inches) and 185 cm (73 inches) will be employed.
These heights correspond to volume of 0.11 m^ and 0.19 m3 respectively and,
in conjunction with variable gas flow rates, should provide adequate varia-
tion in residence times.
The capacities of the charging and overflow hoppers are 103 kg (227 Ib)
or 0.078 m3 (2.75 ft^) . These are more than adequate for 8 hour runs under the
most severe conditions.
REGENERATOR FACILITIES
Many of the control and supply facilities for the air, nitrogen and gas
systems are common to the combustor and regenerator and were discussed pre-
viously. Only those portion of these systems that are unique to the regen-
erator will be discussed in this section.
Regenerator Vessel
The regenerator has an effective height (grid to off gas discharge) of
4.57 m (15 ft) and is constructed of 12 inch IPS, pipe, and lined with Grefco
75-28 refractory to an inside diameter of 9.52 cm (3.75 inch). The plenum
chamber below the grid is 0.69 m (2.25 ft) high and is lined with Grefco
Bubblelite refractory.
A fluidizing grid is inserted between the plenum chamber and the first
section of the regenerator. The grid serves to support the static bed, and
to provide a uniform distribution of fluidizing air when the regenerator is
in operation. The previously used grid will be replaced with_a new grid
(Figure VI-4). The new grid, having four cooling water circuits, will
operate at much cooler and more uniform temperatures than did the old grid,
with only one cooling water circuit.
121
-------
N)
NJ
a
o
.-L
23
O
Pi
§
Hi g
PJ Ml
53
w
5 a
g
s
s
g
o
13
-------
• h TP? ?TraAnr " *qUlpped With two bed overflow lines. These are 4
xnch IPS Sched 40, carbon steel pipe section that are refractory lined to
3 5 CmK I "? } ; n ™,ioWe$ Plpe Pr°vides a bed depth of 60 cm (24 inch)
and a bed volume of 0.00428 m3 (261 in3). The discharge line may be blanked
off and a refractory plug inserted when it is desired to use a greater bed
height. The high overflow line provides a bed depth of 121 cm (48 inch)
and a bed volume of 0.00863 m3 (527 in3).
The overflow lines are connected to an overflow lock hopper of 0.98 m3
(3.5 ftj) with the aid of 5.1 cm ID flexible stainless steel hose. The use
of flexible hose eliminates the rigid pipe connections between the overflow
lines and the hopper. This eliminates a possible source of leakage and
alignment problems due to thermal distortion. The lock hopper volume is
more than adequate for eight hours of operation under the highest anticipated
sorbent discharge rates.
Sorbent is charged to the regenerator from a lock hopper (of 0.98 m3
capacity also) via rigid piping and a 1-1/2 inch ID flexible hose. The sor-
bent is introduced at the bottom of the bed to insure complete mixing; the
charge point is 9.5 cm (3.75 inches) above the grid. The sorbent charging
line is equipped with two MOV's to meter in fresh sorbent in the same manner
as was discussed for the combustor.
Supplementary air and methane, over and above that supplied to the
burner, are introduced at about the middle of the bed and just above the
fluidizing grid respectively. Supplementary air and fuel are used to produce
oxidizing and reducing zones which are needed to carry out the regeneration
reactions.
The regeneration reactions are highly endothermic and unlike the com-
bustor, no bed cooling coils are used in the regenerator. Instead, a 7.5 kw
air preheater has been installed in the regenerator air supply line to provide
a auxiliary heat source, if needed.
Provisions have been made for feeding coal to the regenerator, either as
a primary or supplementary fuel. These facilities have not been installed
as yet, but injector ports have been allocated and capped off.
Thermocouples have been installed to provide bed temperature readings
every 15 cm. Above the bed, gas temperatures will be read at 30-38 cm
intervals.
The plenum chamber is mounted directly below the grid and contains the
burner. The functions and mode of operation of the burner are essentially
the same as for the combustor preheat burner. Safety and cooling system
functions in the same manner for both units.
Pressure differentials are measured across the grid and, at partial and
total bed heights, in order that bed heights, bed densities and pressure
drops across the grid may be obtained.
123
-------
Regenerator Off Gas System
The regenerator off gas is passed to a single cyclone (Figure VI-5)
whose solids are returned to a pressurized standpipe. The clean off gas is
passed to an off gas precooler (0.194 m^ area) and an off gas cooler (0.197
m-3 area). These are of the single-pass, water cooled counter-current flow,
tube and shell type, and are mounted vertically. The precooler is intended
to reduce the off gas temperature from 982°C (1800°F) to 370°C (700°F). The
off gas then passes to the cooler where the temperature is further reduced
to 150°C (300°F).
The off gas facilities for the regenerator that are downstream of the
cooler are identical to those for the combustor. These have been discussed
previously.
124
-------
FIGURE VI-5
BENCH UNIT REGENERATOR CYCLONE
! i!
i>fe
,
]
I"
125
-------
SECTION VII
COMPREHENSIVE ANALYSIS OF EMISSIONS
A program is now underway sponsored by the EPA to assess the environ-
mental effect of fluidized bed coal combustion. The program consists in
setting emission goals for all potentially harmful emissions and measuring
the concentration of these materials in all the effluents from fluidized bed
combustion units. The contractor currently coordinating the work for the
EPA is Battelle Columbus Laboratories. Exxon Research and Engineering
Company has been requested by the EPA to participate in the program by
carrying out a series of comprehensive analysis tests in the miniplant.
In these tests, specified emissions ranging from S02, NOX, etc which are
routinely measured to trace inorganic and organic materials present in the
solid and gaseous effluents from the miniplant are to be measured. The
first series of these tests (Run 50) was completed in April 1977 in coopera-
tion with Battelle. Samples were obtained and analyzed by both laboratories.
The results will be published in a summary report prepared jointly by
Battelle and Exxon early in 1978.
126
-------
SECTION VIII
ANALYSIS OF DESULFURIZATION DATA
In this section an additional analysis of desulfurization data is
discussed. Previous analysis of the data was based on the use of Ca/S ratio
calculated from the analysis of the spent sorbent and the flue gas. The
desulfurization data were again analyzed using the Ca/S ratio set on the
solids feeding system but corrected for variations in combustor temperature
and gas phase residence time. A comparison of S02 retention results obtained
in batch and continuous FBC units was also made. In this comparison,
corrections were applied to batch data to account for the residence time
distribution effects occurring in continuous units. An examination of the
approach to steady state in a combustor with continuous sorbent feed and
used sorbent withdrawal was also made.
ANALYSIS OF DESULFURIZATION RESULTS BASED
ON Ca/S RATIO SET ON SOLIDS FEEDER
In the previous report (1) and in an earlier section of this report,
desulfurization data using Pfizer dolomite sorbent were reported. Correla-
tion of the desulfurization results was based on the use of a Ca/S molar
ratio calculated from a sulfur balance using the expression
% SO Retention (SR)
Ca/S
% Ca Utilization (Xj
U
This approach was used to minimize the possible effects caused by an incom-
plete approach to steady state operation and mechanical problems with the
coal/sorbent feed blender. As pointed out in the earlier discussions, this
approach correlated the data reasonably well with an acceptably small amount
of data scatter whereas the use of the Ca/S ratio set on the solids feeder
resulted in more scatter. However, the correlation of S02 retention data with
a Ca/S ratio calculated from the same S02 retention results would be expected
to smooth the data and could give a misleading correlation. The desulfuriza-
tion data were again analyzed using the Ca/S ratios actually set on the feed
solids blender but corrected for residence time and temperature variations.
This was done to determine if an acceptable correlation could be developed
without referring to a calculated Ca/S ratio.
Test of the First-Order Reaction Rate
Residence time corrections were to be made using the first order rate
expression discussed in the previous report (1). In Run No. 51, operating
conditions were adjusted to permit a specific test of the validity of a
first order kinetic expression. This established a firmer basis for the
subsequent residence time corrections. This test was accomplished by
quickly reducing the bed height (by rejecting bed solids) while maintaining
the superficial velocity constant. As the change in bed height occurred
over a very short interval of time, the average sorbent utilization in the
bed would remain nearly constant. If the rate constant for desulfurization
has a first order dependence, a linear relationship between the In (1-SR) ,
where SR is the fraction S02 retention, and the residence time would be
127
-------
expected, with the slope of the line equal to the negative of the rate
constant. In Figure VIII-1, the values of In (1-SR) for twelve determina-
tions in Run No. 51 are shown plotted against the gas phase residence time.
The first seven determinations represent data taken as the bed height was
increasing during the course of the run and the last five points are data
taken during the rapid withdrawal of bed solids. During this rapid with-
drawal of bed solids, the change in bed height caused the gas phase resi-
dence time to drop 25 percent.
The data in Figure VIII-1 follow the expected relationship for a first
order rate constant. The calculated rate constant is 1.28 s~l (based on
the volume of settled particulate bed). The sorbent utilization in Run No.
51 measured at the time of the test was 47 percent.
Correction of S02 Retention
for Gas Phase Residence Time
The measured SC>2 retention for the runs using Pfizer dolomite are shown
in Figure VIII-2 versus the Ca/S molar feed ratio determined from the coal
and sorbent feed rates set on the solids feeder. These runs were made with
coals containing 2 and 4 percent sulfur. While uhe SC-2 retention does not
appear to be dependent on the coal sulfur content, the scatter of the
experimental data, as measured, is considerable. The data shown covers a
range of gas phase residence times between 1.5 and 3.8 s and average com-
bustor bed temperatures between 684 and 945°C. These wide variations would
be expected to cause data scatter beyond that due to uncertainty in the Ca/S
ratio.
The measured SC>2 retention (SR) values shown in Figure VIII-2, were
corrected to a constant value of 2 s for the gas phase residence time in the
expanded bed of use by the following first order rate expression.
2 In (1-SR )
on -i / meas N
SR2s = 1 - exp ( )
meas
Where t is the measured gas phase residence time in the expanded bed.
meas r
The SC>2 retention values for a 2 s gas phase residence time in the
expanded bed are shown in Figure VIII-3. While the absolute values of the
SC>2 retention have been changed (best observed by looking at the change in
values of the open circles in Figures VIII-2 and VIII-3), the scatter in
the experimental values still exists.
Temperature Dependence of the Desulfurization Performance
Four of the S02 retention values corrected to a 2 s gas phase residence
time shown in Figure VIII-3 which deviate markedly from the trend (labelled
L.T.) correspond to data obtained at combustor bed temperatures of 690,
684, 762 and 829°C. As pointed out in the previous section, desulfurization
was found to decrease at temperatures under 800°C.
128
-------
FIGURE Vlll-l
Ln (I-SR) VS. GAS PHASE RESIDENCE TIME
T
T
-1.10-
2 -1.20
LU
o;
CM
O
CO
-1.40
12
-1.50
DATA: Run No. 51 VELOCITY = 1.52 m/S
Points 1-7 Bed Height Increasing
Points 7-12 Bed Height Decreasing
1.1
1.2 1.3 1.4
RESIDENCE TIME (S)
1.5
1.6
-------
100
90
FIGURE VIII-2
SULFUR REMOVAL EFFICIENCY VS. Ca/S
o
Q
UJ
o:
Z3
CO
<
LU
^
oo
o
^
UJ
UJ
>
O
C£
C£
CO
80
70
<=> 60
50
40
30
S
O
r
o
PF\ZER DOLOMITE
• 2% S Coal
O 4% S Coal
1
0.5 1.0 1.5 2.0
Ca/S MOLAR PATIO
2.5
3.0
130
-------
FIGURE VIII-3
SULFUR REMOVAL EFFICIENCY @ 2 SEC. VS. Ca/S
100
PFIZER DOLOMITE
2% S Coal
4% S Coal
L.T. - LOW
TEMPERATURE
POINTS
Ca/S MOLAR RATIO
131
-------
The temperature dependence of the S02 retention corrected to 2 s gas
phase residence time in the expanded bed is shown in Figure VIII-4 for two
Ca/S feed ratios. The use of a constant gas phase residence time for this
analysis removes much of the variation in the data which would normally
mask the effect of temperature. Clearly a temperature dependence is evident.
The experimentally determined desulfurization rate constant for runs made
at a Ca/S feed ratio of 1.5 was then evaluated for the temperature depen-
dence using an Arrhenius plot, as shown in Figure VIII-5. An activation
energy of 13.3 kcal/gm mole was determined which is in good agreement with
the value reported by Borgwart of 10-20 kcal/gm mole for various limestones
(14).
Correction of the SC>2
Retention for Temperature
The measured S02 retention values were corrected to a constant 2 s gas
phase residence time in the expanded bed and to a constant reactor bed tem-
perature of 900°C using the activation energy of 13.3 kcal/g mole. The S02
retention at a 2 s gas phase residence time and 900°C is shown in Figure
VIII-6. It is evident that most of the experimental scatter has now been
removed. The trend in SC>2 retention determined by this analysis is in good
agreement with that shown in Figure 111-21 although the Ca/S feed ratio used
in this analysis is based on calcium and sulfur feed rates while the data
from Figure 111-21 is based on calculated Ca/S feed ratios using the measured
values of SC>2 removal and sorbent utilization.
COMPARISON OF S02 RETENTION RESULTS
OBTAINED IN BATCH AND CONTINUOUS UNITS
In the previous report (1) it was shown that the values of S02 reten-
tion at a given calcium to sulfur feed ratio obtained in the continuous
pressurized fluidized bed combustors from different laboratories were in
agreement when appropriate corrections were made to reduce the data to a com-
mon gas phase residence time. Agreement of the results obtained in batch
fluidized bed combustors with that observed in continuous units, i.e., the
miniplant has been less satisfactory (1). This lack of agreement between
the results obtained in the batch units and in the continuous units is not
surprising when one considers the difference in the operating environment
between these units. In the batch units, a charge of fresh sorbent particles
initially removes essentially all the S02 released during combustion and in
the process becomes partially sulfated, with presumably each particle reach-
ing the same level of utilization at the same time. As the particle utiliza-
tion increases, the ability to remove S02 decreases and the S02 retention
drops. In the continuous units, fresh sorbent is continuously being added
to the reactor and an average mixture of sorbent particles is being removed
from the bed, such that a steady state value of both sulfur removal
efficiency and sorbent utilization is established.
With the increased availability of continuous reactors and batch reactors
for fluidized bed combustion studies and the increasing use of thermo-
gravimetric analyies (TGA) to study sorbent utilization, it would be highly
desirable to be able to relate the desulfurization performance obtained in
the various units. If such a relationship can be obtained, it would enable
132
-------
UJ
co
CNJ
o
o
§
o
cc
ZD
co
100
90
FIGURE VIII-4
SULFUR REMOVAL EFFICIENCY @ 2 SEC. VS. TEMPERATURE
I 1
• Ca/S =1.5
A Ca/S = 0.75, 2% S Coal
A Ca/S = 0.75, 4% S Coal *
80
70
60
50
A
401—
600
700 800
900
1000
TEMPERATURE - DEC. CELSIUS.
-------
FIGURE VIII-5
EFFECT OF TEMPERATURE ON DESULFURIZATION RATE CONSTANT
+0.5
-0.5
-1.0
O
8.5
O
9.0
T
T
T
PFIZER DOLOMITE
O Ca/S =1.5
Calculated
Activation Energy
13,300 Cal/gm Mole
O
_L
_L
9.5
10.0
10.5
1/T x 10+4 (DEC. KELVIN)
134
-------
FIGURE VIII-6
COMPARISON OF SET VS. CALCULATED Ca/S EFFECTS
100, , ,
90
o
o
O
O
o
o
LU
CM
®
OH
CM
O
CO
80
70
60
50
40
30
0
/ O
/
/
O/
8
0.5
1.0
PFIZER DOLOMITE
• 2% S Coal
O 4% S Coal
-.-. Ca/S Calculated from
Analyses (Fig.111-21)
1
1
1.5
2.0
3.0
Ca/S MOLAR RATIO
135
-------
an expanded use of the TGA and batch units to screen new sorbents which might
be considered for use in the larger continuous units. Such a relationship
would also increase the confidence that the performance data obtained in
process development scale continuous units, i.e., the miniplant, can be used
for predicting the performance of commercial size fluidized bed combustors.
This subsection describes an attempt to relate the performance observed
in these various units.
Residence Time Averaging for the Effective
Rate Constant in the Miniplant Combustor
The miniplant combustor, unlike the smaller batch combustor, will be
in a dynamic state as far as solids entering and leaving the reactor.
Fl
W
Where F^ = feed rate of uncalcined stone to the reactor - moles/hr
F? = removal rate of sulfated product - moles/hr
W = bed hold up - moles
The feed and removal rates, F^ and F2, are adjusted to maintain a uniform
quantity of material, W, in the fluidized bed. The backmixing of solids which
occurs in the fluid bed results in an age distribution for the solids, i.e.,
the solid particles being removed from the bed in the overflow stream F2
have been in the bed for various lengths of time. This variation in the
residence time of the solids is of course reflected in their degree of CaO
utilization.
The kinetic data obtained from the batch combustor gives the reaction
rate constant as a function of particle utilization, Xg. In order to use
this rate information it is necessary to know how the utilization of the
particle changes with the time the particles are in the bed, (the particle
age or particle residence time 8).
Assume an increment of fresh particles of mass, w, added to this
fluidized bed at time 6 = 0. The rate of SC>2 uptake by these particles is:
d"°so2
de ~ k(xB)'w'cso
where N = mass SO^
6 = time particle is in bed (s)
k(Xfi) = rate constant (s~ )
136
-------
CSO = average concentration of SO seen by particles
2 (mass S02/mass particles)
The rate constant is written k(XB) to indicate that it is a function of the
sorbent utilization (or sulfation level), X .
' B
At steady state, CS02 is constant during the residence time of the
particles at some average value between C* in and C* out.
2 2
Equation (1) can be expressed in terms of a concentration Cg09 given in
units of mass S02/volume of settled bed, by
dNso2 k(xB)
de - — -W'cso2
where p = bulk density of the settled bed.
S
In this analysis, all bed volumes and gas phase residence times are
based on the settled rather than the expanded bed.
The sorbent utilization XB is the S02 uptake compared to the maximum
S02 capacity of the particles.
wt S00 uptake S00
x _ £ _ £. (3)
B capacity for SO wot
where a = mass capacity of S02 per unit mass of stone.
Differentiating equation (3)
dNso
dx = 2 (4)
Substituting equation (4) into equation (2) and separating variables,
_
k(XB)
de (5)
A logarithmic mean concentration for Cgo is the appropriate averaged
concentration.
On integration between XB = 0 and XB at time 6, we obtain after
rearrangement :
137
-------
X
dX SO
R
" C
0
The integral in equation (6) is solved numerically to various values of
X and the particle residence time to obtain that Xg given by:
8 - ^ (7)
where A is the area under the X_ vs. k(XD) curve from X = 0 to X .
D D O a
In order to predict the limiting performance of the continuous unit
it is necessary to know the age distribution of the reacting solids. It has
been found that the following particle age distribution function, 1(6), is
quite accurate (13).
K0) = i exp (-e/e) (8)
o
Where 6 is the average particle residence time defined as
6 = ~
Fl
The analysis of the reaction rate constant showed that the rate constant,
k(XB), depended on the degree of CaO utilization, XB. The degree of CaO
utilization, Xg, for a particle injected into the fluid bed reactor was shown
above to depend on the length of time the particle was in the reactor, 6.
These considerations suggest that the reaction constant, k(Xg), for a given
sorbent particle will depend on the time that this particle has been in the
reactor. The expression for this dependence of the rate constant on the
particle residence time, k(6), can readily be obtained from the expressions
previously determined for the dependence of the rate constant on the degree
of CaO utilization, k(XB), and the dependence of the degree of CaO utiliza-
tion on the particle residence time Xg(6).
The desulfurization performance for the miniplant combustor will be made
up of the contribution of all the particles in the fluid bed, each with its
own particular residence time. The average rate constant for the miniplant
combustor, k", is obtained by integrating the product of the expressions for
the variation of rate constant with particle residence time, k(6), and the
age distribution for the particles, 1(0) over all values of particle
residence time, 6.
138
-------
0
This expression is solved graphically.
_ The fractional S0? retention, SR, at steady state can now be calculated
using the rate expression for the plug flow reactor model, by specifying
the operating parameters
SR = 1 - exp (-k-|5.) (11)
where Hs is the settled bed height, and U is the superficial velocity in the
miniplant combustor.
The overflow stream for the fluid bed, F2, contains, as mentioned
earlier, a wide distribution of particle ages, each with a degree of CaO
utilization dependent on its history in the bed. An analysis similar to
that used above leads to an expression for the average CaO utilization, XB,
in the commercial reactor.
XB(6)-I(0) d6 (12)
The degree of CaO utilization can also be obtained at a given value of
the calcium to sulfur molar feed ratio, Ca/S, by using the value of the
S02 retention, SR, calculated from equation (11), and the following expres-
sion which reflects a steady state sulfur balance around the reactor
x -
X ~
B SR
The expression for the degree of CaO utilization from equation (12) is
designated Xg (Distribution Function) to signify its direct calculation from
the assumed particle distribution function. The expression from equation (13)
is designated Xg (Performance) to signify its calculation from the calculated
sulfur removal efficiency at a given molar feed ratio of sorbent calcium to
coal sulfur. The adequacy of the assumed particle distribution function can
be tested by comparing the values of Xg calculated in these two ways.
Calculation of the Miniplant Desulfurization Performance
The calculation procedure described above has been used to calculate
the desulfurization performance expected in the miniplant combustor with
Pfizer dolomite as the sorbent based on the performance data measured in the
batch combustor. The batch unit performance data for 1500 micron Tymochtee
dolomite (tests No. 71 to 82) were used to calculate the necessary values
of the reaction rate constant, k. As the batch unit data was limited to
sorbent utilization levels between 0.20 and 0.46, while the calculation
139
-------
procedures described above require averaging over all values of particle
residence time (equivalent to all values of XB between 0 and 1), the necessary
integration of 1/k in equation (6) was performed after first fitting a
linear equation to values of 1/k measured in the batch unit. The values of
the reaction rate, k, measured in the batch unit and the curve fitted expres-
sion for k as a function of Xg which was used in these calculations is shown
in Figure VIII-7.
The bed hold-up, W, required by the expression for the average particle
residence time in equation (8) is determined by specifying the settled bed
height for the case to be calculated and the area of the miniplant combustor.
The S02 inlet concentration required to calculate the average SC^ concentration
seen by the particles, equation (1) , is determined by specifying the coal
feed rate, the superficial velocity and the temperature for the case to be
calculated. The sorbent feed rate, Flf is determined by specifying the Ca/S
ratio for the case to be calculated.
The numerical procedure used to calculate the desulfurization performance
of the miniplant combustor first estimates a value of the desulfurization
performance to enable the calculation of the average SC^ concentration seen by
a particle in equation (1). The entire calculation procedure is performed
and the desulfurization performance calculated by equation (11). The cal-
culated value is compared with the estimated value, a new value for the
desulfurization performance estimated, and the calculation procedure repeated
until the two values agree.
An example of the prediction of the miniplant reactor performance from
the batch combustor data is given in Figure VIII-8 for the operating condi-
tions of 900°C and a 2 second gas phase residence time (based on the time for
the gas to transit the expanded bed) and the predicted values are compared
with the experimental values measured in the miniplant. While the predicted
performance does not exactly correspond to the experimental data measured in
the miniplant, the predicted values are non-the-less quite representative
of the desulfurization performance with dolomitic sorbents. The model pre-
dicts the same desulfurization performance, irrespective of the sulfur
content of the coal feed, which agrees with the experimental determinations.
Comparison with the Performance
Measured in the TGA
The time that a given sorbent particle requires to achieve a given level
of utilization can be calculated from equation (7). A comparison of the time
to a given utilization level predicted from the data obtained in the batch
combustor with the time experimentally measured in the thermogravimetric
analyzer (TGA) for Pfizer dolomite is shown in Figure VIII-9. In these
TGA experiments the temperature was maintained at 900°C and the concentration
of SC>2 passing the sorbent particle was maintained constant at 2500 ppm.
The agreement is sufficient to distinguish clearly the predicted values as
that for a dolomite sorbent, as contrasted to the value expected for other
types of sorbents, i.e., a high calcium content limestone.
140
-------
CO
8
7-
< 5
co
^
o
0
FIGURE VIII-7
SULFATION RATE CONSTANT VS. SORBENT UTILIZATION
1 1 1 1 1
! PFIZER
1 1 1 1
DOLOMITE 1
Dp = 1500 microns
w
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
SORBENT UTILIZATION (FRACTION)
-------
FIGURE VIII-8
100
90
80
o
o
O
O
o
, 70
o
LU
CO
C\J
© 60
UJ
h-
50
CM
o
CO
40
30
0
S02 RETENTION @ 2 SEC. VS. Ca/S SET
EXXON MINIPLANT PERFORMANCE
PFIZER DOLOMITE
8
O
_L
PREDICTED PERFORMANCE!
FROM BATCH UNIT
DATA & MODEL
EXPERIMENTAL DATA
• 2% S COAL
O 4% S COAL
I
0.5 1.0 1.5 2.0
Ca/S MOLAR RATIO (SET)
2.5
3.0
142
-------
o
H
O
a:
u.
2
g
i-
<
O
O
FIGURE VIII-9
TIME FOR A PARTICLE TO REACH CALCIUM
UTILIZATION LEVEL IN ATMOSPHERE OF
2500 PPM 502 AT 900°C
PFIZER DOLOMITE
CALCULATED FROM
BATCH UNIT DATA
50 100 150 200 250
TIME IN REACTOR (MINUTES)
300
143
-------
Implications of the Model
The model provides an insight into some of the factors that control the
desulfurization performance in a continuously fed combustor like the miniplant.
Perhaps the most striking of these factors is the surprising effectiveness
of fresh sorbent particles, e.g., those particles most recently fed to the
reactor, in determining the desulfurization performance. In Figure VIII-10
are shown three histograms which represent the contribution of particles
which have been in residence in the reactor for different hourly time
increments to the total number of particles, to the sulfur removal
efficiency and to the sorbent utilization attained at steady state condi-
tions for a sorbent feed rate corresponding to a calcium to sulfur molar
feed ratio of 1.5. The contributions shown are for Pfizer dolomite as the
sorbent at a gas phase residence time in the expanded bed of 2 seconds.
It is seen from the histograms that the 15 percent of the total par-
ticles which have been in the reactor for a time interval between 0 and 1
hours, account for nearly 55% of the sulfur removal achieved. The particles
in the bed for a time duration up to three hours, while representing only
31% of the particles in the reactor, account for 85% of the sulfur removal.
It is this marked effectiveness of fresh sorbent particles for sulfur removal
that results in a psuedo steady state SC>2 concentration being obtained within
a few hours following a change in the Ca/S ratio.
While the fresh particles are very effective in determining the degree
of sulfur removal achieved, Figure VIII-10 does indicate that they contribute
little to the steady state value of sorbent utilization. Those same 31 per-
cent of the particles that accounted for 85% of the sulfur removed, account
for only 15 percent of the sorbent utilization. The reason for this is that
a small, but significant, fraction of the sorbent particles are in the reactor
for very long periods of time. In the example shown in Figure VIII-10, the
average particle residence time is 6.8 hours, up to which time 63% of the
total number of particles are accounted for. The remaining sorbent particles
have longer residence times and correspondingly higher utilization levels.
This model provides an explanation for the experimental observation
that the degree of SC>2 removal appears to approach its steady state value
within only a few hours following a change in the Ca/S feed ratio. The
analysis also points out the care that must be exercised in calculating
the Ca/S ratio from the experimental measurements of sorbent utilization
and SCsj retention and the expression for the sulfur and calcium mass balance
around the combustor at steady state conditions. Clearly, one must
allow sufficient time following any change in the feed ratios to insure that
the sorbent utilization has indeed approached its steady state value for the
calculated Ca/S feed ratio to be valid. The danger in using this method to
calculate the Ca/S ratio can better be appreciated using an example based on
the data shown in Table VIII-1. In this table the steady state values of
S02 retention and sorbent utilization for various Ca/S feed ratios are given
for Pfizer dolomite sorbent at a gas phase residence time of 2 seconds in
the expanded bed.
144
-------
FIGURE VIII-10
CONTRIBUTION OF PARTICLES IN BED FOR VARIOUS TIME PERIODS
TO STEADY STATE VALUE OF GIVEN PARAMETERS
Ca/S = 1.5, DOLOMITE, 2 SEC GAS RESIDENCE TIME
60-
50-
LU
fj 40-
LJ
h-
i* 20-
co
> 10-
t— 0-
01
1°
No. OF PARTICLES
|p
:/x
X^Xx
i^
:^x
xxx^
XXXx
^^ ^ H^ ^^ ^^ kXx^ rx/^i
•xxx x/x^ KXxx X/x^ rxxxi l^xx'^'l rxxxl |xXxS1 fx'x'/l f"y ^ ^ t
<" 012 3456789 10 11 12
£ 60-
Q
o 50-
cc.
LU
Q. 40-
UJ
^-—
P 30-
20
CO
y 10-
_i j. \,/
o
£ oJ
% c
%
^
x^
XxX
^X
Xx
XX
///,
^
!xxx
S02 RETENTION
1
w
yyy
x^ ^xJ
X/XX rxXxi f/Xxxl I///XI I/xVxi
11234 5 6 789 10 11 12
Q.
Ll_
0 50-
o
j= 40-
—)
CD
EE 30-
i—
o 20-
o
10-
n.
%
SORBENT UTILIZATION
r77!
///
^J^J:^ ^^F^^^f^1,^^1, — r-
'O i 2 3 4 5 6 7 8 9 10 11 12
TIME PARTICLE HAS BEEN IN REACTOR (HOURS)
145
-------
Table VIII-1. Comparison of Rate Constants
Ca/S _SR_ XB ^rformance) kave
(s"1) (s'1)
1.0 0.67 0.67 1.09 0.82
1.5 0.78 0.52 1.48 1.02
2.0 0.84 0.42 1.79 1.20
2.5 0.89 0.36 2.11 1.42
3.0 0.92 0.31 2.40 1.74
As an example, assume that steady state conditions were established at
a Ca/S ratio of 2.0. The sorbent utilization at this condition would be 42
percent. If the Ca/S ratio were changed to 1.0, within a few hours the
measured SC>2 level above the combustor bed would begin to approach a value
corresponding to the steady state value of sulfur removal efficiency of 67%.
It has been shown that the average sorbent utilization in the reactor bed
does not respond nearly as rapidly and will within the same time interval
more closely correspond to that of the previous established steady state
value of 42 percent for a Ca/S ratio of 2.0. It would clearly be wrong to
calculate the Ca/S ratio from the values of the sulfur removal efficiency
and the sorbent utilization measured within a few hours of this change in
Ca/S ratio.
The marked effect of the fresh particles in determining the sulfur
removal, results in the average rate constant observed at a given steady
state value of the sorbent utilization in the continuously fed combustor, i.e.,
the miniplant, being larger than the corresponding value of the rate constant
at the same value of sorbent utilization in a batch combustor. This can be
seen by comparing the values of kave and kbatch in Table VIII-1. Stated
another way, if the rate constant obtained in the batch combustor is used
directly in equation (12) the S02 retention calculated at a given gas phase
residence time will be less than that obtained using the above described
model. This is shown in Figure VIII-11 for Tymochtee dolomite at a 2 second
gas phase residence time.
The model can also be used to explain the maximum degree of sorbent
utilization that can be achieved in a given continuously fed combustor. To a
first approximation, the degree of sorbent utilization that can be achieved, is
determined by comparing the time that a sorbent particle spends on the
average in the combustor and the time that it takes an individual sorbent
particle to reach a given level of utilization. As an example in Figure
VIII-12, the average particle residence time in the miniplant combustor and
the time for a particle to reach different levels of utilization is shown for
dolomite sorbent at a 2 second gas phase residence time in the expanded bed
for different Ca/S feed ratios. It can be seen, that at a Ca/S of 1.5, (the
level necessary to meet the EPA new source performance standards for SC>2
emissions when burning a 4 percent sulfur coal) , the achievable sorbent
utilization is only 60 percent, as at this Ca/S feed ratio, the average particle
residence time in the miniplant combustor of 10 hours does not provide suf-
ficient time for the sorbent particle to achieve a higher level of utiliza-
tion. This is in good agreement with values measured during the course of
this program. The figure also enables an estimate of the effect on sorbent
146
-------
FIGURE Vlll-ll
S02 RETENTION @ 2 SEC, 900°C VS. Ca/S MOLAR RATIO
PREDICTED EXXON MINIPLANT PERFORMANCE
PFIZER DOLOMITE
100
o
o
O
O
o
LU
CO
CM
g
\-
•z.
Ul
I-
LU
CN
O
CO
90
80
70
60
50
40
30
BATCH DATA & MODEL
BATCH DATA
USING EQUATION (12)-
0
1.0 1.5 2.0
Ca/S MOLAR RATIO
147
-------
100
co
en
ID
o
o
i-
co
ID
DO
2
o
o
10
FIGURE VIII-12
TIME IN COMBUSTOR VS. Ca/S RATIO
EXXON MINIPLANT PERFORMANCE CALCULATION
PFIZER DOLOMITE
AVERAGE PARTICLE
RESIDENCE TIME IN
\ /COMBUSTOR
\/
CALCIUM
UTILIZATION
ACHIEVED
0.8
0.7
0.5
0.4
SUPERFICIAL VELOCITY = 6 FT/SEC
GAS PHASE RESIDENCE TIME = 2 SEC
I i I . I
0
Ca/S RATIO
148
-------
utilization if the SC>2 emission standard were to be tightened. By way of
example, if a 3 to 1 ratio of Ca/S feed were to be required to meet the new
standard, the expected sorbent utilization which could be achieved in the
miniplant would be less than 40 percent.
149
-------
SECTION IX
CONTINUING STUDIES
The future program in the miniplant will be concentrated on three major
tasks; high temperature, high pressure particulate removal, a comprehensive
analysis of all emissions from the unit and regeneration studies. After
conversion of the batch unit to a continuous bench scale unit, experimental
programs will begin centering on the study of NOX emissions, evaluation of
various coal and sorbent types and regeneration tests. These continuing
studies are described in more detail in this section.
HIGH TEMPERATURE, HIGH PRESSURE PARTICULATE REMOVAL
The particulate removal program is now concentrated on optimizing
the performance of a granular bed filter. Although operability for over a
24 hour period has been demonstrated, particulate removal efficiency must be
improved, loss of filter media during blow back decreased and certain opera-
tional problems corrected. Buildup of particulates within the filter beds
indicate the blow back has not been effective in cleaning the beds and this
will be studied. The use of various filter media of differing size and
density will be studied, filtration and blow back conditions and methods will
also be varied and the effects measured. It is also planned to use low tem-
perature transparent models to observe the action of the particulates and the
filter media during filtration and blow back. This will be done in an attempt
to learn more of what is occurring in the filter at high temperatures.
The filter will then be readied for extended testing. These tests are
part of the DOE sponsored program to study erosion, corrosion and solids
deposition effects for a series of gas turbine blade materials exposed to
typical PFBC flue gas conditions. The program, which is a part of the
cooperative EPA/DOE FBC effort, will use gas turbine blade materials and
a test passage supplied by General Electric Company. In addition to the
gas turbine materials test, boiler tubing materials will also be tested in
samples mounted in the combustor vessel within and above the expanded bed.
These samples will be supplied by Westinghouse Research Laboratory.
The extended tests will consist of a 100 hour shakedown run to test
the compatibility of the combustor, filter and materials test sections.
If the systems are compatible and the filter performs satisfactorily, long
term testing will begin. If the systems are not compatible or if the filter
does not perform satisfactorily, a decision will be made whether to proceed
with the long term tests and with what gas cleanup configuration. The long
term tests will consist of a series of exposure tests totalling 1000 hours
exposure time. Gas turbine and boiler materials will be periodically
removed and inspected by General Electric and Westinghouse respectively.
Following the completion of the filter tests, one or possibly two
alternate particulate control devices will be fabricated, installed and
tested on the miniplant in place of the granular bed filter. The choice
of the alternate devices will be made in consultation with the EPA. Cur-
rently, devices such as a high temperature electrostatic precipitator or a
high temperature bag filter are being considered.
150
-------
COMPREHENSIVE ANALYSIS
At the present time, a Level 1 analysis has been conducted on the
effluents from the miniplant combustor. The program calls for the completion
of another Level 1 analysis with the combustor coupled to the regenerator.
In addition, Level 2 and Level 3 analyses are also planned with the com-
bustor and regenerator.
In support of the comprehensive analysis and particulate removal pro-
grams, two high temperature particulate sampling systems have been designed
and are now under construction. These will be used to sample particulates
entering and leaving the granular bed filter to determine particulate con-
centration, size distribution and composition. Provisions will be included
to maintain temperature of the particulates at 870°C or higher to prevent a
change in composition due to condensation of volatile materials on the
particulates.
REGENERATION
A test program is being planned to characterize and develop sufficient
information to determine the feasibility of the combustion-regeneration
system. The primary independent variables are makeup Ca/S rate and solids
recirculation rate. Another independent variable is sorbent type (limestone
and dolomite). Dependent variables include S02 emissions from the combustor
and regenerator. Parameters which are expected to be kept fixed during the
program are combustor and regenerator temperatures, regenerator air/fuel
ratio, combustor excess air level, bed depths, and coal type (Illinois coal
will be used). The operating pressures in the combustor and regenerator
will probably be fixed at 5 to 6 atm to provide sufficient fluidization
velocity in the regenerator to permit good mixing of solids. This pressure
level is a compromise. Otherwise, significant modifications must be made
to permit operation of the system at 9 atm.
Runs of two to five days duration are planned. However, because the
activity of the sorbent is expected to decline gradually over many cycles
of sulfation and regeneration, it may not be practicable to reach a true
steady state. By removing samples of bed periodically, it may be possible
to follow the activity of the sorbent as a function of time and number of
cycles and extrapolate these results to longer times. A thermogravimetric
analyzer (TGA) may be used to run a standard "activity" test by sulfating
samples of regenerator bed. The rate of sulfation would provide a measure
of activity of the recycled stone.
BENCH SCALE UNIT
The future work in the bench scale combustor unit will evaluate various
coal and sorbent types. Coals of various rank, ash and sulfur content will
be tested as well as lignites, chars, etc. The effect of recycle of parti-
culates from the first stage cyclone to the combustor on combustion efficiency
will also be measured. Staged combustion to reduce NOX levels will be studied.
151
-------
Regeneration studies will also be carried out. Activity maintenance of
the recycled sorbent will be measured as a function of regeneration condi-
tions including the use of coal as the regenerator fuel.
152
-------
SECTION X
REFERENCES
1. Hoke, R. C. et al, "Studies of the Pressurized Fluidized-Bed Coal
Combustion Process," EPA-600/7-77-107, September, 1977.
2. "Reduction of Atmospheric Pollution," Vol. 1, p. 119, National Coal
Board (U.K.), September, 1971.
3. Keairns, D. L. et al, "Fluidized Bed Combustion Process Evaluation -
Phase II - Pressurized Fluidized-Bed Coal Combustion Development,"
EPA/650-2-75-027C, September, 1975.
4. Vogel, G. J. et al, "Bench Scale Development of Combustion and
Additive Regeneration in Fluidized Beds," Proceedings of the Third
International Conference on Fluidized-Bed Combustion, EPA-650/2-73-053
p. 1-1-1, December, 1973.
5. O'Neill, E. P. et al, "A Thermogravimetric Study of Limestone and
Dolomite - The Effect of Calcination Conditions," Thermochemica Acta
1£, 209 (1976).
6. Jonke, A. A. et al, "Sulfated Limestone Regeneration and General FBC
Support Studies," Proceedings of the Fluidized Bed Combustion Technology
Workshop, Vol. II, p. 343, CONF-770447-P-2, April 13-15, 1977.
7. Skopp, A et al, "Studies of the Fluidized Lime-Bed Coal Combustion
Desulfurization System," December 31, 1971.
8. Hoke, R. C. et al, "A Regenerative Limestone Process for Fluidized-Bed
Coal Combustion and Desulfurization," EPA-650/2-74-001, January, 1974.
9. Hoke, R. C. et al, "Studies of the Pressurized Fluidized-Bed Coal
Combustion Process," EPA-600/7-76-011, September, 1976.
10. "Pressurized Fluidised Bed Combustion," R&D Report No. 85, Interim
No. 1, National Coal Board (U.K.), September, 1973.
11. Vogel, G. J. et al, "Recent ANL Bench-Scale, Pressurized Fluidized
Bed Studies," Proceedings of the Fourth International Conference on
Fluidized-Bed Combustion, p. 21, December 9-11, 1975.
12. Hoy, H. R., Roberts, A. G., "Further Experiments on a Pilot-Scale
Pressurized Fluidized Bed Combustor at Leatherhead, England,"
Proceedings of the Fluidized Bed Combustion Technology Workshop
Vol. II, p. 43, April 13-15, 1977, CONF-770447-P-2.
13. Nauman, E. G., Collinge, C. N., Chem. Eng. Sci., 23., 1317 (1968).
153
-------
14. Borgwardt, "Kinetics of Reaction of S02 with Calcined Limestone,"
Env. Sci. & Tech., 4_, 1, January, 1970, pp. 59-63.
15. Vogel, G. J. et al, "Reduction of Atmospheric Pollution by the
Application of Fluidized-Bed Combustion and Regeneration of Sulfur-
Containing Additives," EPA-650/2-74-104, September, 1974.
16. Grumpier, T. B., Yoe, J. H., Chemical Computations and Errors, John
Wiley NY, 1949.
17. Cooper, L., "Measurement of High-Temperature, High-Pressure Processes,"
EPA-600/7-78-011, January, 1978.
18. Murthy, K. S., et al, "Comprehensive Analysis of Emissions from
Fluidized Bed Combustion Processes," Process Measurements for Environ-
mental Assessment Symposium, Atlanta, GA, February, 1978.
154
-------
SECTION XI
LIST OF PUBLICATIONS
1. Hodges, J. L., Hoke, R. C., Bertrand, R. R.t "Prediction of Temperature
Profiles in Fluid Bed Boilers," ASME/AIChE Heat Transfer Conference,
St. Louis, MO, August 9-11, 1976.
2. Hoke, R. C., "Particulate Control in Pressurized FBC-Granular Bed
Filter Applications," High Temperature and Pressure Particulate Control
Symposium. Washington, DC, November 9, 1976.
3. Hoke, R. C., Nutkis, M. S., Kinzler, D. D., "Pressurized FBC Studies I.
Combustion," Proceedings of the Fluidized Bed Combustion Technology
Exchange Workshop Vol. II, p. 157, CONF-770447-P-2, Reston, VA,
April 13-15, 1977.
4. Nutkis, M. S., Loughnane, M. D., "A Program for Hot Corrosion/Erosion
Materials Testing for Application to Fluidized Bed Coal Combustion,"
Proceedings of the Fluidized Bed Combustion Technology Exchange
Workshop, Vol. II, p.217, CONF-770447-P-2, Reston, VA, April 13-15, 1977.
5- Ruth, L. A., "Regenerable Sorbents for Fluidized Bed Combustion,"
Proceedings of the Fluidized Bed Combustion Technology Exchange
Workshop Vol. II, p. 301, CONF-770447-P-2, Reston, VA, April 13-15, 1977.
6. Ruth, L. A., Gregory, M. W. , Bertrand, R. R., "Pressurized FBC Studies
II. Sorbent Regeneration and Particulate Control," Proceedings of the
Fluidized Bed Combustion Technology Exchange Workshop Vol. II, p. 329,
CONF-770447-P-2, Reston, VA, April 13-15, 1977.
7. Nutkis, M.S., "Hot Corrosion/Erosion Materials Program and Experience
in the Pressurized Fluidized Bed Coal Combustion Miniplant," Engineering
Foundation/ASME Conference on Ash Deposits and Corrosion Due to
Impurities in Combustion Gases, Henniker, "NH, June, 1977.
8. Bertrand, R. R., "Temperature Control in the Exxon Fluidized Bed
Combustion Miniplant," 1977 ISA Symposium on Instrumentation and
Control for Fossil Demonstration Plants, Chicago, IL, July 13, 1977.
9. Hoke, R. C., "Pressurized Fluidized Bed Combustion," Proceedings of
the 12th Intersociety Energy Conversion Engineering Conference, Vol. I
p. 737, Washington, DC, August 28 - September 2, 1977.
10. Ruth, L. A., "Sorbent Regeneration in Fluidized Bed Combustion,"
Proceedings of the 12th Intersociety Energy Conversion Engineering
Conference, Vol. II, p. 758, Washington, DC, August 28 - September
2, 1977.
155
-------
11. Ruth, L. A., Varga, G. M., "Developing Regenerable S02 Sorbents for
Fluidized Bed Coal Combustion Using Thermogravimentric Analysis,"
paper presented at NATAS 7th Conference, St. Louis, MO, September, 1977,
12. Nutkis, M. S., "Hot Corrosion Erosion Testing of Materials for
Application to Advanced Power Conversion Systems Using Coal-Derived
Fuels Task II - Fluidized Bed Combustion," First International
Conference on Materials for Coal Conversion and Utilization,
CONF-771025 UC-90h, V-l, Gaithersburg, MD, October 11-13, 1977.
OTHER PRESENTATIONS
Bertrand, R. R., Hoke, R. C., Seminar on Fluidized Bed Combustion,
Massachusetts Institute of Technology, January 1977.
Hoke, R. C., Seminar on Fluidized Bed Combustion, Polytechnic Institute
of New York, October, 1977.
156
-------
SECTION XII
APPENDICES
Page
A FBC Solid Effluent Rates, Balances, Composition 159
B Second Order Correlation of S02 Retention Data 164
C Error Analysis 182
D Data Management Systems 183
E Analytical Techniques 196
F Data Logger Channel Identification 197
G FBC Miniplant Alarm/Shutdown System 199
H Tables
H-l Miniplant Fluidized Bed Coal Combustion Run Summary 203
H-2 Particle Size Distribution Spent Pfizer 1337 211
Dolomite Sorbent (Except As Noted)
H-3 Particle Size Distribution Primary Cyclone Capture 211
H-4 Particle Size Distribution Secondary Cyclone Capture 212
H-5 Particle Size Distribution GBF Capture 213
H-6 Particle Size Distribution Flue Gas 214
Particulates No Filter
H-7 Particle Size Distribution Flue 214
Particulates After Filter
H-8 Miniplant Solids Analysis 215
H-9 Miniplant Solids Composition 221
H-10 Miniplant Sample Shipments 224
J Figures
J-l Modified Filter Bed 227
J-2 Modified Filter Element 228
Fluidizing Grid
J-3 GBF Pressure Vessel 229
157
-------
APPENDICES (Continued)
Page
J-4 GBF Pressure Vessel Lining 230
J-5 Filter Pressure Vessel (Side View) 231
J-6 GBF Piping Arrangement 232
J-7 Preheat Burner System Schematic 233
J-8 Granular Bed Filter Blow Back System Flow Schematic 234
158
-------
APPENDIX A
FBC SOLID EFFLUENT RATES, BALANCES, COMPOSITION
Solid rates, concentrations and compositions were summarized for two
combustion runs made in the Exxon miniplant using Illinois No. 6 coal and
Pfizer No. 1337 dolomite. Conditions and emission results for the two runs,
Nos. 43.4 and 43.5, are given in Appendix H-l. Both runs were made at a
pressure of 940 kPa, an average bed temperature of 950°C, a superficial
velocity of 1.7 m/s and excess air of 25%. Run 43.4 was made at a Ca/S
ratio of 1.8 and 43.5 at 0.7. Run 43.4 achieved a 96% retention of S02,
run 43.5 a 44% retention.
SOLIDS RATES AND BALANCES
A calcium balance was made over the entire time period during which
runs 43.4 and 43.5 were selected. The balance was more meaningful over the
entire time period because the bed height and composition changed during
the time period. The results of the calcium balance are shown in Table
A-l. The imbalance of only 7% is very good.
Individual calcium balances were then calculated for the two runs con-
sidering only the steady state periods and the results are shown in Table
A-2. It is evident that the "loss" of calcium from runs 43.4 and
43.5 is due to bed build up.
If the bed had been maintained at a constant height, the fraction of
calcium in the bed removal line would have been 0.65 using the overall
results and 0.59 using the combined results at steady state. The overall
results, when normalized to 100%, give the ratio of bed-removal-calcium
to overhead-calcium of 58:42 or, roughly 60:40.
Since the amount of calcium removed from the bed is less in run 43.4,
this suggests that the bed build up occurred primarily, if not completely,
in run 43.4. This was verified by bed height measurements which indicated
a 30% build up in bed height in run 43.4 and a 3% build up in run 43.5.
The ratios of solids added and removed from the combustor to the coal
feed are shown in Table A-3. Since a build up occurred in 43.4, the solids
retained in the bed to coal feed ratio was estimated and is also shown in
the table. The total solids removed and retained in 43.4 equal 0.27 wt/wt
coal which is about 80% of the expected weight ratio. In run 43.5, the
total solids removed equal 0.23 wt/wt coal which is about 115% of the
expected weight ratio. Therefore, even with the correction for an increase
in bed height, the disposition of the output solids was not properly split
between the two runs. However, by combining the two runs and calculating
average input and output rates, an acceptable ratio of measured to expected
solids output of 95% was determined.
159
-------
TABLE A-l. CALCIUM BALANCE
Runs 43.4-43.5 Entire Time Period
Percent of Ca Into Bed
Fly Ash Line(1) 41
Bed Removal Line 38
Particulates 1
Increase of Ca in Bed 27
TOTAL 107
(1) Secondary cyclone
discharge
TABLE A-2. INDIVIDUAL CALCIUM BALANCES
Runs 43.4 and 43.5 Steady State Periods
Percent of Ca Into Bed
Run No.
43.4
43.5
Combined Runs
Bed Removal
Line
27
44
32
Fly Ash
Line
43
40
42
Particulates
0.4
0.7
0.6
Total
70
85
75
Combined Runs Corrected ,Q ,„ n , ino
,. Tl 1 T, • I 1 TT ->•' ^ U.D 1U2
for Bed Build Up
160
-------
TABLE A-3. SOLIDS/COAL WEIGHT RATIOS
Run No .
43.4
43.5
Combined
Coal
Fed
1.00
1.00
Sorbent Bed Removal
Fed Line
0.32 0.05
0.13 0.10
1.00 0.22 0.08
Fly Ash
Line(l)
0.13
0.13
0.13
Flue Gas Retained Z Solids Out
Particulate in Bed Expectation
0.004 -0.09 -0.80
0.004 0 1.15
0.004 "0.04 -0.95
(1) Secondary Cyclone Discharge
Run No. Bed
43.4
43.5
Removal
0.09
0.10
TABLE
Ash Removed /Ash Fed
Line Fly Ash Line
0.61
0.81
A-4. ASH BALANCE
Particulates
0.03
0.03
Retained Corrected
Total in Bed Balance
(Ash Balance)
0.73 -0.27 1.00
0.94 0 0.94
-------
Ash balances were also calculated for the two runs. The results are
shown in Table A-4. Again, if a correction is applied to 43.4 for the build
up in the bed, good balances are calculated. As shown in Table A-4, 65 to
85% of the ash is removed overhead, the balance is removed with the used
bed.
A sulfur balance was also made during the entire run period from start
up to shut down. The results are shown in Table A-5. The balance on sulfur
was 87%, while this was not as good as the calcium balance, it is nevertheless
reasonable.
SOLIDS COMPOSITION
The composition of the solids removed from the bed and the fly ash is
given in Table A-6. The percentage of the solids present as unburned carbon
and used sorbent was calculated from chemical analysis of the solids.
The percentage of ash in the solids was then obtained by difference. The
composition of the calcium portion of the bed removal solids and fly ash
was also calculated from the chemical composition and is given in Table A-6.
Solids from Run 43.5 were also analyzed for sodium and potassium by
atomic absorption. The results are given in Table A-7.
162
-------
Run No.
TABLE A-5. SULFUR BALANCE
Runs 43.4-43.5
Percent of S into Bed
Fly Ash Line 25%
Bed Removal Line 18%
Stack Gas 27%
Increase of S in Bed 17%
TOTAL 87%
TABLE A-6. SOLIDS COMPOSITION (WT %)
Carbon
Sorbent
Ash
Sorbent Calcium
Composition
CaO CaC03 CaSO/
43.4
Bed Removal Line
Fly Ash Line
43.5
Bed Removal Line
Fly Ash Line
Particulate
1
1
1
1
3
82
54
89
35
35
17
45
10
64
62
37
16
29
24
—
10
6
7
3
—
53
78
64
73
—
TABLE A-7. ALKALI CONTENT OF SOLIDS - RUN 43.5
Solid
Dolomite
Bed Material
(After 43.4,
Before 43.5)
Fly Ash
Particulates
Na (Wt %)
0.04
0.05
0.37
0.80
K (Wt %)
0.008
0.02
0.63
1.34
163
-------
APPENDIX B
SECOND ORDER CORRELATION OF S02 RETENTION DATA
In the previous report (1), sulfation data from Exxon and other
laboratories, using dolomite sorbent, were correlated by using a first order
expression for the rate constant (Eq. 1). The rate constant was then
related to the degree of sulfation of the dolomite.
kx = -7 In(l-x)
k^ = first order constant
t = residence time (bed height/superficial velocity)
x = fraction SO,, retention (SR)
This procedure gives a good correlation but has some draw backs when
used for design purposes. For example the Ca/S molar ratio does not enter
explicitly into the value of x that would be calculated for a given set of
conditions. Also, the sulfation level is not known a priori (moles SO^
in bed/moles Ca in bed).
Another correlation has been developed that not only allows the S02
retention to be calculated but also takes into account all the major design
parameters. This technique is derived by assuming that the rate of removal
of S02 and CaO is second order (first order in CaO and first order in 802).
It is assumed that sulfur in the coal is instantly converted to 862 and that
the dolomite is instantly calcined. It is also assumed that the gas/solid
reduction between S02 and the sorbent can be treated in a manner similar
to that used for homogeneous reactions.
The second order rate expression is
&L = k7(a'x')(b'-x') (2)
at ^
3
where a' = initial concentration of CaO, mols/m
3
b1 = initial concentration of S09, mols/m
L 3
x1 = concentration of product (CaSO^), mols/m
t = time, s
3
k = second order rate constant, m /mol-s
Integration of this expression gives
_ ,
- V
On dividing the numerator and denominator of the logarithmic operand
by (b')2, one obtains
164
-------
x'
a'-b' £l n-^li
u t ^•L -u ' '
D D
By definition, a'/b' = a = Ca/S molar ratio in the feed. Similarly,
x'/b' = x = fraction of S02 retained at steady state.
One then arrives at:
-rW In *:x . = k9t (5)
a -b a(l-x) 2
If b1 is factored out of the prelogarithmic term, there results
1 a—x
7-7-7—TV In —T~I r = k0t (6)
b'(a-l) a(l-x) 2
The quantity b', the initial concentration of S0_, is given by
(R)
(11.52 x 10b)(V)(A)
where S = Percent sulfur in coal,
R = Coal feed rate, kg/hr,
V = Superficial velocity of gas through the bed, m/sec,
A = Reactor cross sectional area, m2^
f\ ^
and 11.52 x 10 is a conversion factor so that b' is in mols/m .
The reaction time is given by
t = H/V (8)
where H is the expanded bed height in meters.
Upon substitution of (7) and (8) into (6), one arrives at an equation whose
parameters are those commonly used in FBC technology:
11.52 x 106(V)(A) a-x _ ,H _ f
l) ±n a(l-x) ~ *2V ~ *
Another method of arriving at equation (9) is given in a subsequent
section. That method assumes SO. moves in plug flow but that the solids are
well mixed.
The value of k2 is found by plotting f vs t, where t = H/V. The
plot, using data from Exxon miniplant dolomite runs 19 through 51, is shown
in Figure B-l. The several points shown for run 51 were obtained in an
experiment in which the bed height was altered while other parameters
remained constant.
165
-------
FIGURE B-l
SECOND ORDER PLOT USING ACTUAL CONCENTRATIONS
5 -
4 -
(/}
1 3
v
r
o
r-H
x 2
0
O ERE Champion Coal
D ERE Illinois Coal
A ERE Run 51
A
O
O
O
O
o
o o
D
-------
3
The least square fit of the data gives a value for k2 of 9.512 x 10
m /mole-sec.
To determine the expected retention of SC»2 for any set of parameters,
equation (9) is inverted to give
x =
Where c =
ae -1
(k2) (a-1) (S) (R) (H)
- - - s-
11.52 x 10b(A)(V)
The reliability of equation (10) may be seen by calculating the expected
S02 retention and plotting the calculated values vs the measured values.
This has been done for the ER&E data and the results are shown in Figure
B-2. Included in this figure are runs where x > a. These data could not
be used in equation (9) to estimate k2. (It should be noted that in equa-
tions (9) and (10), when a = 1, indeterminate values are found. In cal-
culations, it is only necessary to set a = 0.99999 or 1.000001 and then
perform the calculation.)
To test the validity of equation 10 even further, data were used from
Argonne National Laboratories (ANL) (15) and from the National Coal Board
(NCB)(10,12). The retentions calculated from equation (10) using the second
order rate constant determined from the Exxon data were plotted against the
measured retentions and the results are shown in Figure B-3. Again, a
good fit is found.
The scope of Equation 10 can be appreciated when it is realized that
the expected retention can be calculated with only one adjustable constant,
k£, which is obtained from actual experimental data. The range of variables
in Figures B-2 and B-3 are shown in Table B-l . It can be seen that equa-
tion (10) has been shown to be valid over a wide -range of variables.
Equation (10) can be used to show the effect on sulfur retention of
changing one operating variable while the other variables are held constant.
Figures B-4 and B-5 show the effect of the variables in this fashion.
The top section of Figure B-4 shows the S02 retention vs the feed
Ca/S ratio for coals of different sulfur content. For a given Ca/S ratio,
the retention increases as coal sulfur content increases. This effect dif-
fers from that predicted by the first order model in which S02 retention is
independent of the sulfur content of the coal. Also shown in the figure is
the locus of Ca/S requirements to give 0.52 g S02/MJ (1.2 Ib S02/M BTU) for
coals of 2.8 MJ/kg (12,000 BTU/lb) and 1.4 MJ/kg (6000 BTU/lb) . It is
interesting to note that for the coal with the lower heating value, a
larger Ca/S ratio is required for the 1% coal than the 6% coal.
The superficial velocity (V) has a dramatic effect on the retention
when all other factors are constant. This effect is shown in the central
section of Figure B-4 for two levels of Ca/S. The large effect is not
surprising since V enters Equation 10 to the second power.
167
-------
FIGURE B-2
PREDICTED S02 RETENTION VS. MEASURED S09
RETENTION, EXXON DOLOMITE RUNS
o
i-
z
UJ
I-
LlJ
C\J
O
CO
Q
UJ
h-
o
Q
UJ
100
90
80
70
60
50
40
30
20
10
0
I
I
0 10 20 30 40 50 60 70 80 90 100
MEASURED S02 RETENTION, %
168
-------
FIGURE B-3
PREDICTED RETENTION VS. MEASURED RETENTION
LU
I-
LJ
C£
a,
LU
i-
o
o
LJ
100
90
80
70
60
50
40
30
20
10
0
_L
O ANL VAR Series
• NCB (12)
D NCB (10)
_L
0 10 20 30 40 50 60 70 80 90 100
MEASURED RETENTION, %
169
-------
TABLE B-l. RANGE OF VARIABLES FOR WHICH EQUATION
10 HAS BEEN SHOWN TO BE VALID
Variable ^ Range for Data Source
A, tn
S, %
Ca/S
V, m/s
H, m
R, kg/hr
ER&E
0.0794-0.0873
1.96-4.2
0.5-2.5
1.52-2.96
2.4-7
78-149
ANL
0.0183
2.82
1-3.2
0.64-1.49
0.915
7.75-23.43
NCB
0.558
2.97
1-2.18
0.7-0.762
1.34-2.44
144-273
OVERALL (NOMINAL)
0.018-0.56
2-4.2
0.5-3.2
0.64-3.0
0.92-7
7.8-270
-------
LJ
0"
o
I-
2
UJ
I-
UJ
CM
o
C/)
100
80
60
40
20
0
100
80
60
40
20
0
FIGURE B-4
PREDICTIONS USING SECOND-ORDER MODEL
EFFECT OF Ca/S RATIO ON RETENTION
O 1.96% S Coal
• 0,96% S Coal
D 4% S Coal
• 6% S Coal
Ca/S
EFFECT OF SUPERFICIAL VELOCITY ON RETENTION
O Ca/S =1.0
• Ca/S =1.5
Bed Height = 4m
Coal Feed Rate =
100 kg/hr.
1.96% S Coal
0123456
SUPERFICIAL VELOCITY m/sec.
EFFECT OF BED HEIGHT ON S02 RETENTION
Ca/S =1.0
Superficial Velocity = 2 m/sec.
Coal Feed Rate =100 kg/hr.
1.96% S Coal
468
BED HEIGHT, m
10
12
171
-------
LU
h-
LU
CM
O
100
80
60
40
20
0
0
FIGURE B-5
PREDICTIONS USING SECOND-ORDER MODEL
EFFECT OF COAL SULFUR CONTENT ON S02 RETENTION
Ca/S = 1.5
Superficial Velocity = 2 m/sec.
Bed Height = 4m
Coal Feed Rate =100 kg/hr.
_L
3 4
% SULFUR IN COAL
lOOi
I 80
LU
I-
LU
CM
O
60
40
20
0
0
EFFECT OF COAL FEED RATE ON S02 RETENTION
Ca/S = 1.0
Superficial Velocity = 2 m/sec.
Bed Height = 4 m
S Content o,f Coal = 1.96%
50 100
COAL FEED RATE, kg/hr.
150
172
-------
The lower portion of Figure B-4 illustrates the effect of bed height
on the retention. The effect is significant, especially with shallow beds.
The effects of coal sulfur content (S) and coal feed rate (R) are
portrayed in Figure B-5. Both effects arise because of the higher initial
sulfur contentration with increasing S or R.
It should be noted, on reviewing Table B-l, that Equation 10 has been
shown to hold over most of the range of Ca/S in the upper portion of Figure
B-4 (Ca/S = 0.5-3.2). The equation holds for a good portion of the central
part of Figure B-4 (V = 0.64-3.0) and for the lower section (H = 0.92-7).
Experimental data confirm Equation 10 for a large section of the upper
portion of Figure B-5 (S = 2-4.2) and for all of the lower portion of B-5
(R = 7.8-270).
Figure B-6 shows the percent S02 retention vs the Ca/S ratio for
different residence times, t, when other parameters are held constant. The
actual parameter varied is the superficial velocity for the solid lines and
change of the bed height at t = 1 s for the dashed line. As seen in Equation
(10), velocity enters as the square while H enters as the first power.
If this model is valid, there is no unique value for sulfur retention at a
given value of residence time.
Since equation (10) can be used so effectively for design purposes, it
would be very desirable to delineate the extent for which the equation is
valid. The region of low retention would be very valuable to investigate.
Figure B-7 shows how this was done for one experiment during run 51.
The bed height was varied while all other variables were constant. Figure
B-7 shows that the calculated retention fits the experimental data well over
a narrow range of bed heights. It would be useful, however, to expand the
experimental range from, for example, a bed height of 1m to 5m.
This second order kinetic model is an alternate way to analyze, cor-
relate data and predict results. It has certain advantages over this first
order model described in the previous report (1), and appears to correlate
the data well. Both methods will be evaluated further as more data becomes
available.
ALTERNATE DERIVATION OF SECOND
ORDER RATE EXPRESSION
Ca.
Ca CaS
o o
173
-------
LJ
\-
LJ
CM
O
LO
100-
90
80
70
60
50
40
30
20
10
°0
FIGURE B-6
PREDICTIONS USING SECOND-ORDER MODEL
EFFECT OF Ca/S RATIO FOR
DIFFERENT RESIDENCE TIMES
3 SEC.
2 SEC.
1 SEC
1 SEC,
Superficial velocity varied
Bed Height varied
_J
2
Ca/S
174
-------
UJ
h-
LU
or
FIGURE B-7
COMPARISON OF EXPERIMENTAL EFFECT OF
BED HEIGHT WITH THEORETICAL
100
90
80
70
60
50
40
30
20
10
0
— THEORETICAL
O RUN 51
(5/6/77)
8
10
BED HEIGHT, m
175
-------
The above diagram is a representation of a fluid bed combustor (FBC) .
The sulfur enters in the coal and the CaC03 enters as a solid. Assume that
the sulfur in the coal as instantly converted to S02 and the sorbent is
instantly calcined.
S.^ = rate of sulfur input, mols/s
= rate of S0_ input
Ca.^ = rate of Ca input, mols/s
= rate of CaO input, mols/s
x = fraction of sulfur retained
SQ = rate of sulfur out, mols/s
= rate of S02 out, mols/s
Ca = rate of CaO out, mols/s
CaS = rate of CaSO, out, mols/s
The S0? leaves as a gas and the CaO and CaSO^ leave in various solid
streams .
The assumption is made that the solids are completely mixed and the
S02 moves through the reactor in plug flow.
Now by definition,
CaS , _
-^ = * (11)
bi
A material balance and (11) give
Ca.-Ca = CaS = x S. (12)
10 o i
On rearranging (12), one obtains
Ca = Ca.-xS.
o 11
Factoring S. out of (13) gives
Ca — S. Oj ~ x;
o l
Now by definition, Cai/Si = Ca/S molar ratio in the feed. On
denoting this by a,
Ca = S. (a-x)
o i
176
(15)
-------
The initial concentration of S02 is S^APR, where AFR is the air flow
rate in m^/sec. The concentration of CaO, [Ca] is obtained by dividing (15)
by the air flow rate, AFR. (This may be visualized by assuming all CaO is
taken overhead in the gas stream.) Equation (15) then becomes
S.
[Ca] = Ca /AFR = -^(a-x) (16)
O Ar K.
Now S-^/AFR is the initial concentration of SOo in the gas stream. If
this concentration, in mols/m^, is denoted by b, Equation 6 becomes
[Ca] = b(a-x) (17)
If the reasonable assumption is made that the rate of disappearance of
S02 is proportional to the sulfur dioxide and calcium oxide concentrations,
one obtains
-^p- = -k[S][Ca] (18)
When (17) is substituted into (18), there results
= -k[S]b(a-x) (19)
U I
Now [S] at the time of exit, T, is given by
S
_ o
AFR
The value of S is
o
S.-CaSo (21)
From (10), (21) becomes
SQ = S..-S..X = S^l-x) (22)
When (22) is substituted into (20), there results
[S]T = (1~X) = b(l-x) (23)
On substitution of [S] into (19), one obtains
T bd(l-x) .,2,, .,
= = ~kb d-xXa-*) (24)
177
-------
or, on dividing by b,
d(l-x)
—fr. = -kb (1-x) (a-x) (25)
If one lets 1-x = y> (25) becomes
dy
-ft = ~kb(y)(a-l+y) (26)
This can be integrated directly to give
c '
Substitution of (29) into (28) and rearranging gives
b(a-l) *" a (1-x)
Now b, the initial concentration of SC^, is given by
b =
11.52 x 106 (V)(A)
Where (S) = % sulfur in the coal
R = coal feed rate, kg/hr
V = superficial velocity of gas, m/s
A = combustor cross section, m^
11.52 x Id** = conversion factor
(27)
Substituting 1-x = y into (27) gives
KihTlnfS = kT + c <28>
at T = o, x = o and
In *"*, = kT (30)
178
-------
When (31) is substituted into (30), one obtains
f - 11.52 x 106 (V)(A) ,_ a-x _ H ,,„,
f (S) (R) (a-1) lnl(1^0 - kV (32)
Where H = bed height (H/V = T)
The units of k are m /mol-s. The parameters in (32) are those commonly
used in FBC work.
179
-------
CO
o
Coal Feed
Run Rate (kg/hr)
135
139
133
122
27.6 123
27.7 132
27.8 134
27.9 142
27.10 135
27.11 149
27.12 133
27.13 134
27.14 137
27.15 136
27.17 136
27.18 143
27.19 137
35 80
37 110
38.1 90
38.2 90
43.2 88
43.3 93
43.4 96
43.5 95
48 90.9
50.1 74.5
51 78.5
TABLE B-2. CALCULATED VS MEASURED S02
RETENTION USING SECOND ORDER EQUATION
Exxon Data
Coal % S = 1.96
Superficial
Velocity(m/s)
2.01
2.15
1.98
1.72
1.72
1.72
1.83
2.09
1.87
2.23
1.94
2.02
2.04
1.88
1.97
2.09
2.08
1.54
2.96
2.08
2.09
1.81
1.77
1.69
1.71
1.46
1.19
0.5
0.5
0.5
2.5
0.8
1.5
1.5
1.5
0.75
0.75
0.75
0.35
0.75
1.0
0.72
1.0
1.5
0.75
0.75
0.75
0.75
0.75
0.75
1.25
0.5
1.4
1.25
1.52
1.4
Expanded Bed
^Height (m)_
4
4
3
6
5
6
7
5
5
7
7
6
6
5
6
7
7
4.4
4
2.9
3.6
5.7
3.1
3.6
3.7
3.4
3.0
2.4
% S02
Retention
Calc.
42
43
40
100
72
99
99
95
67
68
70
34
68
82
66
85
98
63
20
31
41
73
68
96
49
91
% S02
Retention
Meas.
55
57
80
97.2
98
98.5
72
70
71
46
64
71
72
84
90
62
46
40
45
60
63
96
44
58
75
75
-------
I-1
OO
TABLE B-2 (Continued). CALCULATED VS MEASURED S02
RETENTION USING SECOND ORDER EQUATION
ANL Data
Run
1
2
3
4
5
6
6R
62R
7
8
9
Superficial
Velocity
(m/s)
0.64
1.46
0.64
0.701
1.037
1.098
1.098
1.067
1.280
1.098
1.494
Coal Feed
Rate (kg/hr)
8.42
23.43
7.63
7.75
13.79
13.75
13.21
13.66
19.07
12.73
18.12
A = 0.01825 m
Coal % S = 2.82
Expanded Bed
Height (m)
0.915
0.915
0.915
0.915
0.915
0.915
0.915
0.915
0.915
0.915
0.915
o
-------
APPENDIX C
ERROR ANALYSIS
An analysis was made of the expected error in the value of the calculated
Ca/S molar feed ratio used to correlate desulfurization results. The Ca/S
ratio is calculated from two measured quantities, the S02 retention and the
degree of calcium sulfation (or utilization) in the used sorbent. The error
in each of the measured quantities was first estimated and used to determine
the error in the Ca/S ratio.
The S02 retention is calculated from the following expression
Air rate x (% SO- + SO in flue gas)
SO,, Retention = 1
- - -— — - -
coal rate x (% S in coal)
The calcium sulfation level is given by
moles S04, moles
;
( ,
; us
ra c_. _ oles Ca FA oles Ca
Ca
Where: FA = Fly ash generation rate
US = Used sorbent generation rate
and the subscripts denote the S04/Ca ratio in the fly ash and used sorbent.
The errors in the S02 retention and Ca sulfation were estimated using
standard propagation of error calculation methods (16) . The errors were
estimated based on the following estimated uncertainities in each of the
individual measured quantities used to calculate the S02 retention and Ca
sulfation.
Measured Quantity Uncertainty
Air rate 1 %
S02 emission 10 ppm
Coal rate 2 %
303 emission 10 ppm
S in coal 5 %
Fly ash rate 8.5 %
Used sorbent rate 8 %
SO^" in fly ash 6 %
Ca in fly ash 6 %
S04= in used sorbent 6 %
Ca in used sorbent 6 %
Applying these uncertainties to run 50.2 gave an estimated error in the
S02 retention of 3.3% at a retention level of 88%. The error in the calcium
sulfation was estimated as 13.4% at a sulfation level of 52%. These uncer-
tainties generated an error in the calculated Ca/S ratio of 14% at a Ca/S
ratio of 1.63.
182
-------
APPENDIX D
DATA MANAGEMENT SYSTEMS
A run in the miniplant combustor usually spans a time period of about
ten hours. During this period the conditions in the combustor approach
steady state values resulting in three to four hour data generating periods
characteristic of the chosen set of input variables. In runs were the regen-
erator is included in the system, the approach to steady state conditions is
much slower, due not only to the presence of the regenerator vessel, but also
the deactivation of the recycled sorbent. The large quantity of data gen-
erated during these extended runs requires a systematic method of data
management. It is the purpose of this section to describe systems now being
used to collect, compute, and summarize the data generated in the miniplant.
DATA GENERATION
Run 51 will be used as an example of the data that are generated during
a typical run in the miniplant. In this run, only the miniplant combustor
was operated. The regenerator and the granular bed filter systems were not
used. The primary objective of this run, which spanned a time period of
7-1/2 hours, was to measure combustion efficiency and particulate loading in
the flue gas when operating without recycle of material from the first stage
cyclone back into the combustor.
During the run, operational parameters and some of the desired output
results were continuously recorded by the data logging system at one minute
intervals. Of the 100 channels of information available on the data logger,
70 channels were in use. The remaining channels were connected to instru-
ments used only during a regeneration run. As a back up, in case of a failure
in the data logger system, most of these parameters and output data were con-
tinuously recorded on multi-point strip-chart recorders. The data logger
output, 70 channels of information for the 7-1/2 hours duration of the run,
resulted in 315,000 pieces of experimental data-from this source alone.
Eight discharges from the second cyclone lock hopper, and six discharges
from the bed solids lock hoppers were made during the run. These 12
discharges provided a measure of the rates of fly ash generation and bed
solids which are needed in determining a mass balance around the unit. Four
of the second cyclone discharges and two of the bed solids samples were later
selected as representative of the steady state conditions achieved and sub-
mitted for detailed chemical and particle size analysis. During the run,
seven samples of the combustor bed composition were taken using the bed
sampling probe and submitted for chemical and particle size analyses. Two
measurements of particulate concentration in the off gas from the second
stage cyclone were made during the run and the particulate catch submitted
for chemical and particle size analyses. The initial and final combustor bed
was also submitted for chemical analysis. These 15 solid samples, selected as
representative of the steady state operating conditions, were each analyzed
for six chemical components and the particle size distribution in 12 size
183
-------
fractions was determined. The particulate distribution in the off gas from
the second cyclone was determined in an additional 12 smaller size fractions.
Two wet chemistry measurements for S02 and SO^ were also made to check the
analytical instruments.
A total of 106 individual analysis, 204 particulate size fraction deter-
minations, and 14 solid generation rates were obtained for Run No. 51, a
short duration run with only the combustor operational. These determinations
do not include any replicate chemical analysis which frequently are required
to resolve uncertainties in the analysis.
DATA NEEDS
This quantity of generated data is used in several ways to meet the data
needs of the program. The first need is a rapid (soon after the run) method
to select that specific time interval within the run in which steady state
conditions were achieved. This selection needs to consider all of the data
collected by the data logger system to determine when operating conditions
(bed temperatures, coal feed rates, etc) and output variables (S02 and NOX
emissions, excess oxygen levels, etc.) have reached steady values. Once this
time period has been selected, a determination can be made as to which solid
samples need to be submitted for chemical and particle size analysis.
The second need is experienced when an operational problem which forces
an abrupt shutdown of the unit occurs during the run or when the value of
the output variables is not that expected from past experience. In these
instances the problem is to access and digest the necessary data over the
time interval desired when the information is part of a set of several hundred
thousand data points on the data logger tape.
A third need is to enable a comprehensive data analysis of a run. Of
the 70 points recorded by the data logger each minute during a run using
only the combustor, about 40 points determine the value of output variables
necessary for an analysis of the run. The other 30 points are operational
variables used to determine what went wrong when problems develop or to
provide internal checks on the functioning of the data recording system. The
average value of these 40 output variables over the steady state period
represent one set of inputs to the data analysis. Of course, in determining
the average value of each variable, checks on the functioning of the record-
ing system must be made to determine if the value of the variable at a given
time should, or should not be included in the average. The specification
of the chemical composition of the coal and sorbent used in the run and the
average generation rates, chemical analysis and particulate size distribu-
tions of the various output solid streams complete the data input needed for
the data analysis. Based on this information, some 30 output parameters
need to be calculated to provide a complete description of the output vari-
ables for the input variables selected for the run. A mass balance for
the major chemical elements (C, S, Ca and Mg) and for the solid inorganics
around the various input and output streams is also desirable.
The fourth need is to search selectively all or several runs in order
to provide run summaries or to determine the relationships between output
and input variables. In this context it is desirable to be able to place
the recovered data into files which can be subsequently operated on by
statistical or graphical routines.
184
-------
DATA MANAGEMENT SYSTEMS
In order to handle the data generated and to satisfy the data needs of
the program, several computer oriented data handling programs have been
developed to supplement the manual recording and storage of the data. Certain
of these programs provide a capability to input data directly from the data
logging system while also permitting an operator later to enter analytical
and particulate size distribution data into the computer storage files.
Other programs provide the capability to access the information as needed.
Three major computer systems have been developed to meet the needs of
this program. They are described in the sections that follow under their
call names of Hoke, NCSS and I/O.
Hoke System
The data logging system consists of a Digitrend data logger electron-
ically interfaced with a Kennedy magnetic tape recorder. The Hoke computer
program consists of a subprogram to convert the instrument outputs recorded
by the data logger to physical quantities such as temperatures, flow rates,
pressures, concentrations, etc. These values are subsequently used in
another subprogram to calculate derived quantities such as superficial
velocity, expanded bed height, excess air, etc. The data are then averaged
over preselected time intervals (usually 10 minutes) tabulated, and printed.
If necessary, a complete data printout of an operation can be obtained within
two hours following the completion of the run.
Figure D-l is an example of the tabular form of data printout received
from a run. This printout is then used in conjunction with the continuous
recorder charts to determine the steady state period. When a steady state
period has been determined, the magnetic tape is rerun using the Hoke steady
state program (HOKSS). This program averages selected values over the indi-
cated steady state period. Another feature of the HOKSS program is that
it will delete any value which does not fall into a fixed number of standard
deviations of the average values. Faulty values may result from failure
of a measuring device, recording device, or a momentary upset in the system.
Consequently this value is not representative of steady state and should
not be included in the average. Finally, the average values from the HOKSS
program are inserted into the Input/Output System for record keeping as
well as further calculations. The Input/Output system will be discussed
later in the text.
NCSS System
The NCSS System consists of programs to convert the instrument outputs
recorded by the data logger to physical quantities and programs to search
the physical quantities selectively for a given run or series of runs.
Further data reduction to calculate derived quantities from the measured
parameters, e.g., superficial velocity calculated from flow rate, temperature
and pressure, is also provided. In addition, the system includes auxiliary
informational and diagnostic programs. This section discusses the NCSS
System.
185
-------
FIGURE D-l
M1NIPLANT ANALYSIS PROGRAM
PAGE 1
00
01
1 TIME
2 SUPERFICIAL GAS VELOCITY - FTPS
3 COMBUSTOR AIR FLOW - SCFM
4 CA/S MOLAR FEED RATIO
5 COAL FEED RATE - LB/HR
6 TEMPERATURE GRADIENT- DEC F/FT.
7 DELTA P (PT. 15)
8 EXPANDED BED HEIGHT - IN.
9 AVE. COMB. TEMP - DEG F
10 LOWER COMB. TEMP - DEG f
11 PERCENT EXCESS AIR FROM FLUE GAS COMP.
12 PERCENT EXCESS AIR FROM AIR/FUEL
13 COMBUSTOR PRESSURE-ATM.
RUN
10
11
12
13
64.0
64.0
64.0
64.0
64.0
64.0
64.0
64.0
64.0
64.0
64.0
10
10
10
10
11
11
11
11
11
11
12
20
30
40
50
0
10
20
30
40
50
0
4.6
4.6
4.6
4.6
4.7
4.7
4.7
4.6
4.7
4.8
4.8
528.9
526.6
531.8
536.6
538.1
538.5
538.4
532.9
534.4
538.3
536.7
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
197.2
212.5
213.3
212.5
199.2
199.2
209.6
204.4
195.5
202.2
205.1
6.8
7.1
5.6
5.0
5.0
5.0
5.8
6.4
5.1
4.9
4.7
102.2
90.7
91.7
103.9
114.4
126.5
137.6
202.2
212.8
263.8
275.0
158.0
142.7
141.5
156.1
162.0
180.1
189.2
279.5
305.1
368.5
367.0
1712.7
1711.8
1721.6
1713.6
1712.1
1709.3
1710.4
1704.1
1708.2
1704.3
1708.8
1719.1
1719.7
1727.0
1718.3
1716.9
1714.9
1717.4
1713.0
1714.6
1710.0
1713.3
30.3
30.6
26.3
31.1
29.5
28.4
26.9
27.5
23.1
27.3
25.6
6.8
1.3
0.6
0.5
7.5
7.6
2.3
3.8
8.8
6.0
4.2
8.4
8.4
8.4
8.5
8.4
8.4
8.4
8.4
8.3
8.2
8.1
-------
Data Conversion Programs—
The information on the data logger tape is converted to a form that is
readable by the IBM 370 System and this tape (the "1130 tape") is processed
onto the master tape by a program package called PR1130. The PR1130 program
package mounts the "1130 tape," calls a short Fortran program to mount the
master tape, specifies the auxiliary files to be used, species the Fortran
programs to be loaded, converts the instrument readings to physical
quantities, calculates derived quantities, and stores the quantities on
the master tape. There are also a number of auxiliary files used in
creating the master tape, including:
• library of runs on master tape
• names of variables recorded on master tape
• cross-reference between variable names and data logger position
• library of calibration factors and thermocouples used in
temperature averages
The main program in the PR1130 package calls subroutines to create
the master tape and to establish the cross-reference file between the
variables and the data logger positions. The user can ask to see the
existing cross-reference file, use an old file, create a new file or
edit an old file. After the cross-reference file is established, the user
can specify any changes in the calibration factors or thermocouples used in
the temperature averages. These factors are used by subroutines which reduce
the data from the data logger tape into physical quantities, such as, tem-
peratures, flow rates, pressures, etc., and also calculate derived quantities,
such as, superficial velocity and residence time. These quantities are then
written onto the master tape.
Data Searching Programs—
The primary search tool is a program package which allows the user to:
• examine the list of available runs
• select a run to be search
• select the number of minutes to be averaged
• choose whether standard deviations are to be calculated
• select the portion of the run to be search
• check the millivolt standard for each minute and discard data
for minutes during which the millivolt standard is outside
the allowable range
• examine the list of variable choices
• select the variables to be searched
One of the variables selected can be a "key variable." When an upper bound
and lower bound are set for a key variable and that variable is outside the
given range, not only will the value for that variable be discarded but all
the values for all variables during that minute of data will be discarded.
Even though the user may select only one key variable, in effect he has
187
-------
three key variables at his disposal since both the time and the standard
millivolt checks also function as key variables. A preselected series of
variables, including limits, can be requested by using a single "calling
number" through the use of macro files. When one of these calling numbers
is used, a subroutine inserts the preselected list into the search specifica-
tions. More than one macro at a time can be requested if desired.
After the search specifications are established, the program: finds
the correct position on the tape to start the search, locates the variables
according to the search specifications, calculates the averages and standard
deviations, and prints the data. The data can be printed at the console, at
a local printer, or at a remote printer, as requested by the user. As an
illustration of the type of output available from the data searching programs,
a selection of eight recorded and calculated quantities from a 15 minute
average of data from Run No. 37 is shown, with standard deviations, in Figure
D-2. The data can also be filed for statistical analysis. If it is, the
program writes the data into a data file specified by the user and tells
the user how many observations are filed for how many variables and also
which variables are filed.
Input-Output Programs
ARG1 Series of Programs—
The ARG1 Series of programs are intended to create and support a data
bank established for the storage, calculation, and retrieval of selected
miniplant operating data. The program are written in a conversational
mode.
The program ARG1 "drives" the subroutines which file and edit the input
information, and which calculate and print selected input and calculated
parameters.
The subroutine ARG1C creates the data bank. This subroutine reads the
file which contains the properties of the coal and sorbent used in the run
and reads the data logger tape data for the run generated by the program
HOKSS, described earlier. A singular case of the ARG1C subroutine is the
creation of a list of matrix addresses (known as "the road map") for the
variables. The road map is used to edit existing data banks as additional
information becomes available.
The data bank consists of an unformatted file (ARGF1) of 52 records with
a capacity of 1000 words in each record. The data from a miniplant run is
assigned to an individual record in the file. Each record contains 228
alphameric variables and 900 numeric variables. Provisions are available in
the data bank for including data related to operation of the regenerator
and the granular bed filter, in addition to combustor operation data.
The subroutine ARG1E edits an existing data bank. Additional alphameric
and numerics may be added to the data bank as they become available. Typical
examples are chemical and physical properties of the combustor and regenerator
beds, and additional coal and sorbent data. Erroneous or obsolete informa-
tion may be replaced with new information using this subroutine. A desk
console is used to add this information to the data file.
188
-------
Figure D-2., CSS — MINIPLANT ANALYSIS PROGRAMS — FBCPRT
RUN NO. 37, PFIZER DOLOMITE, CHAMPION COAL, HIGH VELOCITY
4/7/76
1 CONSTANT MILLIVOLT STANDARD SOURCE (40 MV)
2 SOLIDS FLOW RATE (KG/HR)
3 RATIO LIMESTONE/COAL (KG/KG)
4 COMBUSTOR COAL FEED RATE (KG/HR)
5 COMBUSTOR CA/S MOLAR FEED RATIO
6 CALCULATED SULFATION LEVEL
7 S02 EMISSION INDEX (L8/MBTU)
8 COMBUSTOR-N02 EMISSION INDEX (LB/MBTU)
TIME 1 2345678
18 7 40.045 141.490 0.097 128.972 0.804 0.657 1.483 0.303
( 0.104)1 8.961)1 O.OOOH 8.170X O.OOOX 0.069)1 0.175)( 0.019)
18 22 40.049 128.349 0.097 116.993 0.804 0.708 1.354 0.332
( 0.102X 9.521)1 O.OOOX 8.682X 0.000)1 0.061X 0.154X 0.025)
18 37 40.047 131.485 0.097 119.849 0.805 0.663 1.467 0.318
( 0.103X 6.704X 0.000)1 6.111X O.OOOX 0.046X 0.116)1 0.015)
18 52 40.048 134.397 0.097 122.505 0.804 0.811 1.093 0.312
( 0.102X 8.059X O.OOOX 7.344X O.OOOX 0.054X 0.137)1 0.018)
19 7 40.045 126.333 0.097 115.158 0.804 0.761 1.220 0.324
( 0.104)1 5.791)1 O.OOOX 5.280)1 0.000)1 0.074)1 0.187X 0.016)
19 22 40.040 132.157 0.097 120.469 0.804 0.710 1.594 0.298
( 0.106)1 5.027X O.OOOX 4.580X O.OOOX 0.374X 0.078)1 0.012)
19 37 40.034 133.277 0.097 121.483 0.804 0.734 1.288 0.295
( 0.025)1 4.595)1 0.000)1 4.186X O.OOOX 0.038X 0.095X 0.009)
19 52 40.037 127.229 0.097 115.969 0.805 0.650 1.501 0.309
( 0.100X 5.939X O.OOOX 5.415X 0.000)1 0.064)1 0.163)1 0.014)
189
-------
The subroutine ARG1D is used to compute calculated values of operating
parameters. Calculations are performed for some 30 output variables
required to characterize a given run, i.e., superficial velocity, excess air,
Ca/S feed ratio, coal feed rate, etc. A calculation of a mass accountability
for carbon, sulfur, calcium, magnesium and solid inorganics over each of the
input and output streams and a mass balance for these components is also
performed.
The subroutine ARG1P prints the contents of the data bank. The user
has many print options available to him depending upon his needs, and the
amount of information available. The printing format has been selected such
that missing data is printed as a series of asterisks. An "XXXXXX" rotation
is used for those parameters (such as the sulfur content of the combustion
air) that are not obtained. These conventions distinguish between those
parameters which are necessary and those which are never obtained. Data may
be printed on a DS-40 terminal or on the IBM-1403 terminal.
As an example of the use of the ARG1 series of programs, a segment of
the computer print out obtained using the ARG1P subroutine is shown in
Figures D-3 to D-5. The first three figures are examples of experimental
data inputs to the file and the last two figures are calculated outputs.
Figure D-3 shows the average values of the operational parameters for Run
No. 50 and is illustrative of the data logger input to the data file provided
through the HOKSS program. The values of the output variables required to
characterize Run No. 50, as calculated by subroutine ARG1D, is shown in
Figure D-4, while the mass accountability and mass balance for the same
run is shown in Figure D-5.
Ancillary Programs—
The program ARGZ is used to retrieve data selected by the user from the
data file for subsequent statistical analysis or for out-filing the data in
an unformatted data file for other uses.
The program ARG4 is used to summarize mass balances for selected mini-
plant runs that have been previously filed in the data bank. The program's
output is presented in tabular form suitable for inclusion in finished
reports. An option to summarize the mass balances for all miniplant runs to
date is also provided. This option is becoming more valuable as the number
of miniplant runs on file in the data bank increases.
The program ARG5 is used to summarize preselected parameters for
selected, or all, miniplant runs filed to date in a tabular form suitable
for inclusion in finished reports. The preselected parameters cover opera-
ting conditions, flue gas emissions and overall performance. An illustrative
example is shown in Figure D-6.
190
-------
FIGURE D-3
EXXON MINIPLANT RUN NUMBER O0.220
DESCRIPTION:
COMPREHENSIVE ANALYSIS RUN V/BATT^LLE
RUN DATE MONTH 3. DAY 30. YEAR 77.
START TIME 611. HOURS DURATION 11.60 HOURS
PERIOD FOR THIS SUMMARY 1231. TO 1751. DAY 30.
NUMBER 0? COOLIE COILS IN COV/!BUSTOR= ***
COAL: FILE NO. ***** TYPE EASTERN MINS CHAMPION
PARTICLE SIZE: PLUS ***** *IESH, MINUS ***** MESH
SORBENT: TILE NO. ***** TYPE DOLOMITE SOURCE PFIZSR
PARTICLE SIZE: PLUS ***** MESH,MINUS ***** MESH
£CP MO.1337
COAL ANALYSIS
MOISTURE
ASH
TOTAL CAREON
HYDROG-iN
SULFUR
NITROGEN
CHLORINE
OXYGFN
KEATING VALUT
CALCIUM
MASN3SIUM
POTASSIUM
SODIUM
1.67 WPCT
13.10 WPCT
74.00 WPCT
5 . ?, 0 WPCT
1.85 WPCT
1.53 WPCT
******** tfPcr
2.40 WPCT
12973. ETU/LE
PFM
PPM
******** PPM
£*£#***# pp^l
SORBENT ANALYSIS (CALCINED^
CAO
MSO
SI02
AL303
FE203
POTASSIUM
SODIUM
S4 .00 WPCT
44.00 'VPCT
0.90 WFCT
.20 WPCT
0.30 WPCT
******** PPM
191
-------
FIGURE D-4
MASS BALANCES fWPCT)
RUN NUMBER
48.02!
50.11
5 .18
50.20
50.42
51.(
TOTAL MASS 99.93 100.32 99.90 99.81 99.14 99.55 99.73 99.71
SULFUR 94.75 52.63 131.71 106.24 58.42 81.15 88.10 73.37
CARBON 104.59 109.45 104.62 87.90 B9.G0 93.83 93.32 137.51
CALCIUM 108.78 15.42 149.63 104.£5 57.73 31.66 93.51 63.99
REACTIVE OXlfGFN 100.71 102.21 102.96 92.10 97.65 98.27 99.35 105.12
MAGNESIUM 90.19 10.63 9 .66 84.31 54.36 70.94 79.69 50.63
BB3 ****** ****** ****** ****** ****** ****** **-.):#*<( **#«**
SOLID INORGANICS 101.66 163.93 96.54 101.96 72.52 97.32 104.30 94.23
192
-------
FIGURE D-5
MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN
Lo
RUN NUMBER
OPERATING CONDITIONS:
RUN LENGTti(HRS)
PRESSURE(KPA)
LOWER BED TEMP(CELSIu'S )
AVERAGE BED TEMP(CELSIUS)
SUPERFICIAL V?LOCITY(M/SEC;
SETTLED BED HEIGHT(M)
INITIAL
FINAL
EXPANDED BSD HSIGHT(M)
COAL FEED RATEUG/HR)
CA/S MOLAR FEED RATE-SET
CA/S .MOLAR FEED RATE-CALC
EXCESS AIR(VPCI)
SORBENT
COAL
FLUE GAS EMISSIONS
43.00 50.11 5 .12 50.20 50.30 50.40 50.5?
SOZ(PPM)
NOX(PPM)
CO(PPM)
C02(VPCT)
02(VPCT)
RESULTS:
RBTiiNTION(WFCT)
SULFATION(WPCT)
S02(LBS/METU)
NOX(LESXMBTU)
COMBUSTION EFFICIENGT
8.25
932.
934.
874.
1 .4?
1.^?
2.15
3.37
90. 96
1.3?
0.91
39.1?
DOLO
iAST
510.
92.
114.
13.29
5.99
55.78
61.45
1.13
0.10
93.25
9.50
916.
901.
891.
1.19
•1.58
******
3.2)2
74.30
1.34
******
34.59
DOLO
iAST
373.
97.
95.
13.84
5.43
68.10
******
0.91
0.12
99.33
12.00
919.
395 .
688.
1 .21
******
******
3.93
74.54
1.34
1.29
35.91
DOLO
iAST
355.
104.
93.
13.75
5.54
70.23
54.93
0.32
.12
99.15
11.50
901 .
814.
903.
1 .22
******
******
3.19
70 .86
1 .34
1 .63
55.05
DOLD
SASP
151.
138.
131.
3 . 92
7.55
85.53
52 .58
0 .41
0.19
93.93
15.20
901 .
95"0.
885.
1.32
******
******
3.22
32.94
1.34
1.48
40.97
DOLO
SAST
62.
129.
90.
12.06
6.18
93.46
63.03
0. 14
0.15
99.89
12.50
900.
902.
399.
1.33
******
if if * If. * If
3.31
83.43
1.43
1.65
39.01
DOLO
EAST
41.
127.
55.
12.50
5.93
95.81
57.83
0.03
0.14
99.42
18.03
902.
Q?C _
89S.
1.34
******
2.16
3.29
85.20
'1.49
1.45
34. 90
DOLO
E*5I
29.
119.
53.
13.10
5.51
97.02
66.89
0. 06
0.13
99.36
3.25
930.
375.
376.
1.65
1.93
******
2.32
78.78
1.69
1.31
44.76
DOLO
FAST
253.
143.
125.
13.30
6.53
59.42
52.91
.74
0.21
98.19
-------
FIGURE D-6
SUMMARY OF SELECTED PARAMETERS.
RUN NUMBER 48-00 52.11 50.12 50.20
PARAMETERS
COME EFF (PCT) 694 99.25 99.33 99.15 98.93 99.89
NOX (PPM) 440 107.65 113.50 115.40 130-09 141.35
CO (PPM) 441 133.39 111.50 108.58 171.66 93.60
S02 ;LB/MBTU) 444 1.13 0.91 0.32 0.41 0.14
NOX (LB/MBTU) 445 0.10 0.13 0.12 0.19 0.15
SULF REM EFF (PCT) 447 55.78 68-10 72.23 85.63 93.46
COAL FEED (KG/HR) 464 90.36 74.30 74.54 70.86 82.94
SORBENT FEEO (K3/HR) 695 15.35 11.38 11.42 10.86 12.71
RUN NUMBER 50.40 50.50 51.
^PARAMETERS
COMB FFF (PCT^ 694 99.42 99.86 93.19
NOX (PPM, 440 141.15 126.27 210.25
CO (PPM) 441 60.72 55.82 183.80
S02 (LB/MBTU) 444 0.09 0.06 2.74
NOX (L£/MBTU) 445 2.14 0.13 0.21
SUIF REM EFF (PCT) 447 95.Bl 97.0a 69.42
COAL FEED (rCS/HR) 464 B3.4B 86.20 78.78
SORBENT FEED (K&/HR) 695 12.83 13.21 13.71
194
-------
The program ARG6 is used to summarize selected parameters from selected,
or all, miniplant runs whose data has been filed in the data bank. The
results are presented in tabular form, suitable for inclusion in finished
reports. Unlike the two preceding programs, the user may select the para-
meters that he wishes to review. The program also permits the identification
of these parameters (by the user) with a 20 character description.
195
-------
APPENDIX E
ANALYTICAL TECHNIQUES
Analysis of Solids
Solids from combustion runs were analyzed for SO/"2, C0.,~2, Ca
Na+, carbon and total sulfur. -- - - -
described below.
+2
Mg+2,
The analytical techniques that were used are
SO
CO,
-2
-2
The sample was treated with acidic BaCl2 solution.
The BaSO^ precipitate was weighed.
HC1 was added to an acidified sample. The solution was
stripped with N2 and the gas passed through drierite,
and ascarite.
" was determined from the
Ca
+2
Mg
Na
+2
'+
Total -
Sulfur
weight gain of the ascarite.
The sample was digested by heating vigorously in a
medium of perchloric acid/nitric acid. The determination
of Ca, Mg and Na was made by atomic absorption.
(Dietert Sulfur Method) - The sample is combusted in an
oxygen atmosphere at 1250°C. The 502-803 products in
the effluent gas were analyzed by an automatic Leco
titrator .
Total - (Carbon on Catalyst Method) - The sample is combusted
Carbon in an oxygen atmosphere at 1200°C. The C02 evolved was
determined from the weight gain of ascarite.
Analysis of Flue Gas
by Wet Chemical Methods
SO- - The amount absorbed by an 80% isopropanol solution was
determined titrimetrically using 0.01N barium per-
chlorate as the titrant and thorin as the indicator.
SO - 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.
196
-------
APPENDIX F
DATA LOGGER CHANNEL IDENTIFICATION
Ft.
No.
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Loop
No.
1-0
21
12
13
14
15
16
18
1-9
1-10
10
11
36
34
35
40
44
32
22
23
24
25
26
29
29
30
33
20
1-29
1-30
1-31
1-32
1-33
1-34
1-35
1-36
1-37
1-38
1-39
1-40
1-41
1-42
1-43
1-44
1-45
1-46
1-47
1-48
Identification
Constant Millivolt Std. Signal (40 MV)
Coal, Weight in Primary Injector
Combustor, Water Flow Rate - Coil #1A
Combustor, Water Flow Rate - Coil #1B
Combustor, Water Flow Rate - Coil #2A
Combustor, Water Flow Rate - Coil #2B
Combustor, Water Flow Rate - Coil #3A
Combustor, Water Flow Rate - Coil #3B
Batch Unit, Analyzer, NOX (10,000 ppm)
Batch Unit, Analyzer, S02 (10,000 ppm)
Combustor, Flow Rate - Main Air
Combustor, Pressure
Combustor, AP - Coal Feed Vessel
Combustor, AP - Fluidizing Grid
Combustor, AP - Bed (Port 4 to 31)
Combustor, AP - Bed (Port 4 to 11)
Combustor, Water Flow Rate - Heat Transfer Loop
Spare
Regenerator, Flow Rate - Burner Air
Regenerator, Flow Rate - Burner Fuel
Regenerator, Flow Rate - Supplemental Air
Regenerator, Flow Rate - Supplemental Fuel
Regenerator, Pressure (or AP to Combustor)
Regenerator, AP - Fluidizing Grid
Regenerator, AP - Bed (Port 29 to 34)
Regenerator, AP - Bed (Port 29 to 31)
Limestone to Coal Weight Ratio
Main Air Pressure at Measuring Orifice
Run Identification Number
Combustor, Temperature - Burner Grid Metal
Combustor, Temperature - Burner Grid Cooling Water
Combustor, Temperature - Fluidizing Grid Metal
Combustor, Temperature - Fluidizing Grid Cooling Water
Combustor, Temperature - Port #3 (Burner Zone)
Combustor, Temperature - Port #5 (6")
Combustor, Temperature - Port #7 (18")
Combustor, Temperature - Port #8 (27")
Combustor, Temperature - Port #9 (48")
Combustor, Temperature - Port #12 (64")
Combustor, Temperature - Port #13 (85")
Combustor, Temperature - Port #14 (94")
Combustor, Temperature - Port #16 (103")
Combustor, Temperature - Port #20 (140")
Combustor, Temperature - Port #22 (169")
Combustor, Temperature - Port #26 (207")
Combustor, Temperature - Port #28 (252")
Combustor, Temperature - Port #32 (354")
Combustor, Temperature - Coal Injector Line
197
-------
Pt.
No.
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
Identification
1-49
1-50
1-51
1-52
1-53
1-54
1-55
1-56
1-57
1-58
1-59
1-60
1-61
1-62
1-63
1-64
1-65
1-66
1-67
1-68
1-69
1-70
1-71
1-72
1-73
1-74
1-75
1-76
1-77
1-78
1-79
1-80
1-81
1-82
1-83
1-84
1-85
1-86
1-87
1-88
1-89
1-90
1-91
1-92
1-93
1-94
1-95
1-96
1-97
1-98
Temperature
Temperature
Temperature
Temperature
Temperature
Temperature
Temperature
Cooling Water to Coils
Cooling Water from #1A
Cooling Water from #1B
Cooling Water from #2A
Cooling Water from #2B
Cooling Water from #3A
Cooling Water from #3B
Analyzer Range Switch - 862
Combustor,
Combustor,
Combustor,
Combustor,
Combustor,
Combustor,
Combustor,
Combustor,
Combustor, Analyzer Range Switch - NOX
Combustor, Analyzer Range Switch - CO
Combustor, Analyzer Range Switch - Q£
Combustor, Temperature - 1st Cyclone Gas Discharge
2nd Cyclone Gas Discharge
Off-Gas Upstream of Nozzle
Off-Gas Downstream of Nozzle
1st Cyclone Dip Leg
2nd Cyclone Dip Leg
Solids Reject Line (Before Pulse Pot)
Solids Reject Lock Hopper
Fly Ash Lock Hopper
Surface - Lower Deck
Sullair to Measuring Orifice
Combustor,
Combustor,
Combustor,
Combustor,
Combustor,
Combustor, Temperature -
Combustor, Temperature -
Combustor, Temperature -
Combustor, Temperature -
Temperature -
Temperature -
Temperature -
Temperature -
Temperature -
Combustor, Temperature -
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Combustor, Analyzer - SO
Combustor, Analyzer
Combustor, Analyzer
Combustor, Analyzer
Port #2
Port #3
Port #4
Port #5
Port #6
Port #7
Port #9
Port #10
Port #11
Port #12
Port #13
Port #14
Port #16
Port #18
Port #21
Port #24
Port #26
Cyclone Gas Discharge
Off-Gas from Cooler
2 (3,000 ppm)
NOX (10,000 ppm)
02 (25%)
C02 (25%)
CO (6250 ppm)
Combustor, Analyzer
Regenerator, Analyzer - S02 (15%)
Batch Unit, Analyzer - 02 (25%)
Batch Unit, Analyzer - C02 (25%)
Batch Unit, Analyzer - CO (5000 ppm)
198
-------
13
14
15
APPENDIX G
FBC MINIPLANT ALARM/SHUTDOWN SYSTEM
Alarm
1
2
3
4
5
6
i
i
8
9
10
11
12
Unit
Comb
Comb
Comb
Comb
Comb
Comb
Comb
Comb
Comb
Comb
System
Description
Fluidizing Air
Fluidizing Air
Cooling Coil
Water Loop
Cooling Water to
Each Coil (5)
Condenser Cooling Water
Coal Injection
1st Stage Cyclone
Dipleg
2nd Stage Cyclone
Dipleg
Cooling Tower Water
Reactor Vessel
Fluidizing Grid
Alarm Condition
Low Flow
High Flow
Low Total Flow
Low Flow
Low Flow
Low Flow
Low Flow
Low Flow
Low Pressure
High Pressure
High AP
Comb Fluidized Bed
Comb
Coal Injection
Vessel
Demineralized Cooling
Water Reservoir
High AP or
Low AP
Low AP
Low Level
Action
Flow Switch
Orifice AP Cells
Flow Switch
Ultrasonic Microphone
Thermocouple
Thermocouple
Pressure Cell
AP Cell
AP Cell-
*TD Before Alarm
*Close Coal Block Valve
and Valve AV10
Thermowatch
Capacitance Switch
4 min
X
X
/*
0*
-------
FBC MINIPLANT ALARM/SHUTDOWN SYSTEM (Continued)
Action
NJ
O
O
Alarm
#
16
17
18
19
20
21
Unit
Comb
Comb
Comb
Comb
Comb
Comb
System
Description
Natural Gas
Flame
Burner Grid
Fluidizing Grid
Bed Temperature
at 3 Points
Bed Temperature
Upper Bed
Temperature
r*/-*n1 -I i-irr TJo-f-QV
Alarm Condition
Flame Out
High Temperature
High Temperature
High Temperature
Low Temperature
High, High Temp.
Sensor /Function
UV Sensor
*Fuel Shutdown
Thermocouple
Thermocouple
Thermocouple
Thermocouple
Thermocouple
Alarm Time Shut-
Only Delay Down
X*
^
/
/
/
22
23
24
25
26
27
28
29
30
Comb
Comb
From Each Coil
Coal Injection
Probe
Comb Coal Injection Line
Comb
Comb
Comb
Comb
Comb
Coal Injection Ready
Off Gas from Cooler
Combustor to Regenerator
Transfer Line
Fly Ash Lock Hopper
Reactor Shell Surface
Comb Coal Injector Vessel
High Temperature
High Temperature
High Temperature
Low Combustor Temp,
High Temperature
Low Temperature
High Temperature
High Temperature
High AP
Thermocouple
Thermocouple
Thermocouple
*Close Coal Block Valve
Thermocouple
*Close Valve AV10
Thermocouple
Thermocouple
Thermocouple
Thermocouple
AP Cell
*Close Valve AV10
4 min
0*
0*
0*
-------
FBC MINIPLANT ALARM/SHUTDOWN SYSTEM (Continued)
Action
Alarm
#
33
34
35
37
38
39
40
41
43
44
45
46
Unit
Reg.
Reg.
Reg.
Reg.
Reg.
Reg.
Reg.
Reg.
System
Description
Main Air Compressor
Aux. Air Compressor
Site Air
Burner Air
Burner Fuel
Supplemental Air
Regenerator/
Reg . to Comb .
Burner Air
Nitrogen Compressor
Reactor
Supplemental Fuel
Fluidizing Grid
Alarm Condition
Low Pressure
Low Pressure
Low Pressure
Low Flow
High or Low Flow
High or Low Flow
High Pressure/
High Pressure Drop
High Flow
Low Pressure
High Pressure
High or Low Flow
High AP
Sensor /Function
Pressure Switch
Pressure Switch
Pressure Switch
Orifice AP Cell
Orifice AP Cell
Orifice AP Cell
Pressure Cell/
AP Cell
Orifice AP Cell
Pressure Switch
Pressure Cell
AP Cell
Alarm Time Shut-
Only Delay Down
/
/
^
0 X
S
S
'
s
S
o /
/
47
Reg. Fluidized Bed
High AP or Low AP
AP Cell
*T.D. Before Alarm
/*
49
Reg. Natural Gas Flame
50 Reg. Burner Grid
Flame Out
High Temperature
U.V. Sensor
Thermocouple
10 sec
or
5 min
-------
FBC MINIPLANT ALARM/SHUTDOWN SYSTEM (Continued)
Action
Alarm
#
51
52
53
54
55
K3
O
10 57
58
Unit
Reg.
Reg.
Reg.
Reg.
Reg.
Reg.
Reg.
System
Description
Fluidizing Grid
Bed Temperature
At 2 Points
Bed Temperature
At 2 Points
Bed Temperature
Upper 2 Points
Off Gas from
Cooler
Supplemental Fuel
Injection Ready
Reactor Shell
Alarm Condition
High Temperature
High Temperature
High Temperature
High, High Temp.
High Temperature
Low Temperature
High Temperature
Alarm Time Shut-
Sensor/Function Only Delay Down
Thermocouple /
Thermocouple /
Thermocouple /
Thermocouple /
Thermocouple /
Thermocouple /
Thermocouple V
-------
APPENDIX H-l. MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
to
o
(jO
Operating Conditions;
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m3/min
Avg. Bed Temperature, °C
Sup. Vel., m/s
Settled Bed Height, m
Initial
Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Ratio-Calc
Excess Air, %
Sorbent
Coal
Flue Gas Emission:
S02, ppm
NOX, ppm
CO,
C02,
ppm
o
2,
Results;
Retention, %
Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
41.1(7/26-27/76)*
20.5
912
19.6
889
1.6
2.3
86
0.84
44
GL
CH
350
111
116
14
6
68
0.90
0.20
41.2(7/27/76)*
5
912
19.4
875
1.5
1.9-2.2
69
0.84
3.92
37
GL
AR
0
127
133
12
6
100
26
0.00
0.30
43.1(8/24/76)
10
840
18.4
940
1.7
1.4
3.9
98
1.25
24
PD
IL
60
80
180
14
4
98
0.14
0.13
43.2(8/24-25/76)
7
880
22.9
845
1.8
2.1
5.7
88
0.75
1.23
42
PD
IL
720
120
250
9-12
6
60
49
2.58
0.31
* Combined Combustor/Regenerator Runs; Combustor Results Only.
CH = Champion Coal PD = Pfizer Dolomite (BCR No. 1337)
AR = Arkwright Coal GL = Grove Limestone (BCR No. 1359)
Run Discussed in Reference 1.
IL = Illinois Coal No. 6
-------
APPENDIX H-l (CONTINUED). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Operating Conditions; 43.3(8/25/76) 43.4(9/2/76) 43.5(9/2/76) 45-1(10/11-12/76)*
Run Length, hrs. 6 6.3 4.3 24
Pressure, kPa 885 940 940 780
Air Flow Rate, m3/min 20.8 21.1 21.2 14.2
Avg. Bed Temperature, °C 940 940 945 880
Sup. Vel., m/s 1.8 1.7 1.7 1.4
Settled Bed Height, m
Initial 1.6 1.6 — 2.3
Final 1.9 — 2.2
Expanded Bed Height, m 3.1 3.6 3.7 3.2
Coal Feed Rate, kg/hr 93 96 95 82
Ca/S Molar Feed Ratio-Set 0.75 1.25 0.50 0.74
Ca/S Molar Feed Ratio-Calc 1.44 1.80 0.68
Excess Air, % 25 25 25 37
Sorbent PD PD PD GL
Coal IL IL IL CH
Flue Gas Emissions;
S02, ppm 870 114 1328 493
NOX, ppm 90 79 75
CO, ppm 240 201 201
COo, % 12 14 14
02, % 444 6
Results:
Retention, % 63 96 44 65
Sulfation, % 45 53 65
Lb S02/M BTU 2.39 0.26 3.62 1.04
Lb NOX/M BTU 0.18 0.13 0.15
* Combined Combustor/Regenerator Runs; Combustor Results Only.
PD = Pfizer Dolomite (BCR No. 1337) CH = Champion Coal
GL = Grove Limestone (BCR No. 1359) IL = Illinois Coal No.
-------
APPENDIX H-l (CONTINUED). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
N>
O
Cn
Operating Conditions;
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m^/min
Avg. Bed Temperature, °C
Sup. Vel., m/s
Settled Bed Height, m
Initial
Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Ratio-Gale
Excess Air, %
Sorbent
Coal
Flue Gas Emission;
SO,
45-2(10/12-13/76)* 45-3(10/13-14/76)* 45-4(10/14/76)* 45-5(10/14-15/76)*
NO
ppm
'x> PPm
CO, ppm
C02, %
02, %
Results;
Retention, %
Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
30
780
14.2
885
1.4
2.8
99
0.74
40
GL
CH
107
92
0.23
14
780
14.2
890
1.4
2.9
81
1.30
33
GL
CH
217
84
0.46
12
780
14.2
870
1.4
3.4
83
1.06
30
GL
CH
249
82
0.52
28
780
14.2
875
1.4
3.6
85
0.00
28
GL
CH
635
56
1.29
* Combined Combustor/Regenerator Runs; Combustor Results Only.
GL = Grove Limestone (BCR No. 1359)
CH
Champion Coal
-------
APPENDIX H-l (CONTINUED). MINIPLANT BED COAL COMBUSTION RUN SUMMARY
K>
O
O\
Operating Conditions;
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m3/min
Avg. Bed Temperature, °C
Sup. Vel., m/s
Settled Bed Height, m
Initial
Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Ratio-Calc
Excess Air, %
Sorbent
Coal
Flue Gas Emission;
S02, ppm
NOX, ppm
CO, ppm
C02, %
02, %
Results;
Retention, %
Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
45-6(10/14-15/76)* 46.1(11/18/76)
12
780
14.2
875
1.4
2.2
3.6
85
0.00
1.69
27
GL
CH
555
62
37
1.13
3.67
940
15.0
920
1.3
1.8
3.0
81
2.50
37
GL(P)
CH
130
80
100
13
6
90
0.29
0.13
46.2(11/18/76)
4.5
940
15.0
865
1.2
>2.3
3.3
78
2.50
1.49
47
GL(P)
CH
170
120
150
11
7
86
58
0.40
0.20
46.3(11/22/76)
7.67
940
16.7
760
1.0
>2.3
2.7
65
2.50
1.01
40
GL(P)
CH
416
120
180
10
6
55
55
1.31
0.27
* Combined Combustor/Regenerator Runs; Combustor Results Only.
GL = Grove Limestone (BCR No. 1359)
GL(P) « Grove Limestone (Precalcined)
CE = Champion Coal
-------
APPENDIX H-l (CONTINUED). MINIPLANT FLUIDIZED BED COAL COMBUSTION SUMMARY
Ni
O
Operating Conditions;
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m3/min
Avg. Bed Temperature, °C
Sup. Vel., m/s
Settled Bed Height, m
Initial
Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Ratio-Calc
Excess Air, %
Sorbent
Coal
Flue Gas Emission:
S02, ppm
NOX,
CO,
co2,
o2,
ppm
ppm
Results:
Retention, %
Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
46.4(11/22/76)
3.5
940
16.7
865
1.4
2.0
3.1
81
1.50
1.14
35
GL(P)
CH
425
150
100
11
5
64
56
1.07
0.27
47(12/9/76)
3.5
1075
30.0
885
2.2
2.1
2.3
3.8
179
1.50
0.06
18
GL
CH
1400
100
175
14
3
2.7
49
2.86
0.15
48(2/24/77)
8
930
17.9
875
1.5
1.5
2.2
3.4
91
1.40
1.32
39
PD
CH
510
92
114
13
6
58
44
1.16
0.16
50.1A(3/29/77)
9.5
910
13.9
890
1.2
1.6
3.0
75
1.25
41
PD
CH/K
373
97
94
14
5
69
0.87
0.16
GL = Grove Limestone (BCR No. 1359)
GL(P) = Grove Limestone (Precalcined)
PD = Pfizer Dolomite (BCR No. 1337)
CH = Champion Coal
CH/K = Mixture of Champion Coal (2% Sulfur) and Kentucky Coal (0.6% Sulfur)
-------
APPENDIX H-l (CONTINUED). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
NJ
O
oo
Operating Conditions;
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m^/min
Avg. Bed Temperature, °C
Sup. Vel., m/s
Settled Bed Height, m
Initial
Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Ratio-Calc
Excess Air, %
Sorbent
Coal
50.1B(3/29-30/77) 50.2(3/30/77) 50.3(3/31/77) 50.4(3/31/77)
12
915
14.3
890
1.2
3.9
75
1.25
1.52
45
PD
CH/K
11.6
900
15.3
805
1.2
3.2
71
1.25
1.95
63
PD
CH/K
15
900
15.3
890
1.3
3.2
83
1.25
2.35
40
PD
CH/K
12.5
900
15.4
890
1.3
3.3
84
1.25
2.44
40
PD
CH/K
Flue Gas Emission;
NOX,
CO ,
C02,
02,
ppm
ppm
ppm
Results;
Retention, %
Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
355
102
100
14
6
70
46
0.85
0.18
151
136
130
10
8
86
44
0.41
0.26
62
128
90
12
6
95
41
0.14
0.21
41
124
56
12
6
96
40
0.09
0.20
PD = Pfizer Dolomite (BCR No. 1337)
CH/K = Mixture, of Champion Coal (2% Sulfur) and Kentucky Coal (0.6% Sulfur)
-------
APPENDIX H-l (CONTINUED). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Operating Conditions;
Run Length, hrs.
Pressure, kPa
Air Flow Rate,
Avg. Bed Temperature, °C
Sup. Vel., m/s
Settled Bed Height, m
Initial
Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Ratio-Calc
Excess Air, %
Sorbent
Coal
Flue Gas Emission;
S02,
NOX,
CO,
co2,
o2,
ppm
ppm
ppm
Results;
Retention, %
Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
50.5(3/31-4/1/77)
18
900
15.5
895
1.4
2.2
3.3
86
1.25
2.00
36
PD
CH/K
29
119
53
13
6
97
49
0.06
0.19
51(5/6/77)
8
932
18.3
871
1.5
1.9
2.3
79
1.40
1.60
46
PD
CH/K
253
143
125
14
7
71
45
0.72
0.29
52(5/24/77)
23
930
17.8
890
1.5
1.7
1.6
2.7-3.5
75-102
1.25
26-50
PD
CH/K
115-710
90-125
80-125
14
6
53
54(6/29/77)
5.5
595
14.6
910
1.9
2.7
81
0.00
29
None
CH/K
660
31
52
1.41
PD = Pfizer Dolomite (BCR No. 1337)
CH/K = Mixture of Champion Coal (2% Sulfur) and Kentucky Coal (0.6% Sulfur)
-------
APPENDIX H-l (CONTINUED). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Operating Conditions;
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m3/min
Avg. Bed Temperature, °C
Sup. Vel., m/s
Settled Bed Height, m
Initial
Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Ratio-Calc
Excess Air, %
Sorbent
Coal
Flue Gas Emission:
55(7/14/77) 56(7/26/77) 57(8/2/77) 59.1(8/11/77) 59.2(8/12/77)
S02,
NOX,
CO,
CO-,,
ppm
ppm
ppm
02, %
Results;
Retention, %
Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
3.5
570
12.5
930
2.1
2.5
83
0.00
0.37
13
None
CH/K
792
110
2'13
15
5
30
82
1.57
0.16
4
800
14.8
815
1.4
1.8
62
104
11
8
2
93
89
0.16
0.20
6
810
13.7
935
1.4
—
2.6
73
0.69
1.05
34
PD
CH/K
1.6
3.0
91
0.75
1.07
14*
PD
CH/K
150
35
145
13
3
87
81
0.29
0.05
10.5
890
17.7
956
1.6
3.7
105
0.75
33
PD
CH/K
120
78
156
88
0.58
0.27
12
873
17.7
949
1.7
1.3
7.2
109
0.00
0.80
26
PD
CH/K
360
55
163
19
4
66
83
0.79
0.08
PD = Pfizer Dolomite (BCR No. 1337)
CH/K - Mixture of Champion Coal (2% Sulfur) and Kentucky Coal (0.6% Sulfur)
* Excess Air Percent derived from flue gas composition.
-------
APPENDIX H-2. PARTICLE SIZE DISTRIBUTION SPENT PFIZER 1337
DOLOMITE SORBENT (EXCEPT AS NOTED)
Run No,
;(D
45
48
50. IB
50,
50.
50.
51
.2
.3
.5
Particle Size (ym)
5%
Material
Less
Than
10%
Less Than
25%
Less Than
50%
Less
Than
75%
Less
Than
90%
Less
Than
95%
Less Than
56
Final Bed 29
Final Bed 295
Final Bed 200
Rejected Solids
Rejected Solids 160
Rejected Solids 130
Final Bed 560
Rejected Solids 180
Rejected Solids 320
Rejected Solids 300
Final Bed 220
150
350
260
120
235
215
870
280
440
450
270
320
440
680
420
480
440
1260
470
660
560
340
490
1520
960
1040
1080
750
1060
920
540
1500
1720
1240
1540
1360
920
1400
1700
(1) Grove 1359 Limestone-Regenerated
APPENDIX H-3. PARTICLE SIZE DISTRIBUTION PRIMARY CYCLONE CAPTURE
Run No.
48
56
57
59
5%
Less Than
130
102
10%
Less Than
140
130
Particle Size (ym)
25% 50% 75%
Less Than Less Than Less Than
190
235
140
300
710
135
190
880
1150
240
365
90%
Less Than
1550
800
850
95%
Less Than
1850
1175
1250
-------
APPENDIX H-4. PARTICLE SIZE DISTRIBUTION
SECONDARY CYCLONE CAPTURE
Particle Size (pm)
5% 10% 25% 50% 75% 90% 95%
Run No. Less Than Less Than Less Than Less Than Less Than Less Than Less Than
41(1)
43<1m
43.2)^
( / )
43.3^'
43.4U;
43. 4,
I / I
43.5U;
45(D
48
50. 1A
50.2
50.3
51 Bi
ci VJ/
J-L / T \
( J )
c-i V-'/
JJ- /ON
( _J )
c -I \~/ /
52
52
52
55
56
57
57
59
59
59
5.6
9.2
7.4
3.2
5.6
12
5.4
—
4.4
6
4.3
4.5
6.3
5.4
6.0
4.2
4.1
4.1
4.2
5.2
5.4
4.8
4.7
3.8
3.9
4.0
8
12
11
8.0
9.8
18
8.8
3.1
5.6
8.2
5.8
6.1
9.6
7.4
8.0
5.5
5.6
5.7
5.7
7.3
8.0
6.8
6.4
5.2
5.2
5.8
15
21
20
16
20
49
18
8.6
8.9
14
10
10
15
13
14
9.4
11
10
11
15
15
12
12
9.8
9.6
12
36
39
57
35
45
72
39
23
16
26
18
19
53
23
27
20
19
19
20
45
31
23
23
19
18
26
58
59
79
68
70
80
77
60
29
72
39
44
180
75
94
67
36
34
38
140
54
47
46
36
35
48
80
92
115
92
95
94
110
120
68
115
72
86
320
140
210
160
54
59
58
275
76
78
69
70
70
80
98
110
145
110
112
110
145
230
90
185
100
115
375
215
320
250
72
78
76
335
94
100
84
105
105
115
Pfizer 1337 Dolomite, Champion Coal Except as Noted
(1) Grove 1359 Limestone-Regenerated
(2) Pfizer Dolomite, Illinois Coal
(3) First Cyclone Deactivated
212
-------
APPENDIX H-5. PARTICLE SIZE DISTRIBUTION GBF CAPTURE
NO
Run No .
54
55
55
55
55
55
55
55
59
59
59
59
Filter
Element
No.
—
2
2
2
2
3
3
F(1)
2
2
3
3
5%
Less Than
2.5
2.0
2.1
1.8
1.6
2.3
1.8
1.4
1.2
1.4
1.6
1.5
10%
Less Than
3.3
2.6
2.7
2.2
1.9
2.9
2.2
2.2
1.6
1.8
2.2
1.9
Particle Size
25% 50%
Less Than Less Than
5.5
4.5
4.7
4.0
3.0
4.5
3.7
4.1
2.6
2.7
3.9
3.0
10
9.5
10
13
6.5
8.0
8.8
7.3
5.2
4.9
7.1
5.4
(ym)
75%
Less Than
18
22
24
25
18
21
22
13
9.8
9.0
11
9.6
90%
Less Than
30
36
38
33
35
38
35
20
25
15
18
18
95%
Less Than
37
—
—
—
—
—
—
—
—
21
—
30
(1) Fines Collected on Exterior of Filter Element
-------
APPENDIX H-6. PARTICLE SIZE DISTRIBUTION FLUE
GAS PARTICULATES NO FILTER
Particle Size (vim)
Run No.
48
50.1
50.2
50.4
50.5
51
51
5%
Less Than
__
—
—
—
—
—
—
10%
Less Than
__
—
—
—
1
—
—
25%
Less Than
1.5
1.8
1.4
1.5
1.9
2.0
1.5
50%
Less Than
3.1
4.7
2.4
2.5
3.3
4.6
2.8
75%
Less Than
6.5
7.2
4.7
5.0
6.5
9.2
6.9
90%
Less Than
11
13
8.8
9
11
15
11
95%
Less Than
15
16
12
12
14
__
14
N5
APPENDIX H-7. PARTICLE SIZE DISTRIBUTION FLUE
PARTICULATES AFTER FILTER
Particle Size (ym)
Run No.
54
57
59
59
59
5%
Less Than
1.3
2.0
1.3
1.0
1.0
10%
Less Than
1.8
2.3
1.7
1.3
1.3
25%
Less Than
2.7
3.4
2.4
1.7
1.9
50%
Less Than
5.0
6.7
3.4
2.5
3.3
75%
Less Than
17
15
2.7
4.0
8.5
90%
Less Than
—
25
13
9.0
20
95%
Less Than
__
31
20
16
30
-------
APPENDIX H-8. MINIPLANT SOLIDS ANALYSIS
Run No.
41.1
41.2
43.1
43.2
Weight Percent
43.3
43.4
Source
Second Cyclone #2
Second Cyclone #3
Second Cyclone #4
Final Bed
Second Cyclone #5
Initial Bed
Second Cyclone #8
Second Cyclone #9
Second Cyclone #10
Second Cyclone #13
Second Cyclone #14
Second Cyclone #15
Flue Gas Particulates
Intermediate Bed
Second Cyclone #25
Second Cyclone #26
Second Cyclone #27
Flue Gas Particulates
Final Bed
Initial Bed
Rejected Solids #2
Rejected Solids #3
Rejected Solids #4
Second Cyclone #5
Second Cyclone #6
Second Cyclone #7
Flue Gas Particulates
Ca
10.8
8.8
6.6
36.4
16.8
34.5
15.8
14.8
15.6
10,
11,
10.3
6.1
25.3
9.1
9.9
10.1
7.3
26.8
26.8
23.3
26,
30.
15.3
12.4
14.0
7.2
.6
.3
s_
4.0
2.3
1.9
7.2
3.8
3.0
7.8
7.4
7.6
5.7
5.5
5.1
9.4
9.0
5.9
5.8
5.9
9.8
8.7
8.7
11.2
11.6
11.9
8.2
8.5
7.5
11.5
SQ4
12.4
9.6
8.1
22.3
12.1
14.0
27.1
26.1
26.6
19.6
18.5
26.0
27.6
29.8
18.3
20.4
18.2
25.7
28.6
28.6
28.0
31.5
33.0
25.9
26.2
26.1
30.4
C03
0.6
0.6
0.5
5.8
1.7
37.6
1.7
1.7
1.7
1.6
1.8
1.6
0.0
18.6
0.7
0.8
0.9
0.0
1.3
1.3
2.9
4.2
4.6
1.5
1.3
0.8
0.1
Total C
6.8
4.4
3.4
1.1
8.7
8.6
0.5
0.6
0.2
1.6
1.9
2.2
0.5
2.6
1.2
1.0
3.8
4.1
0,6
0.6
1.2
1.2
1.3
0.7
0.6
0.4
1.4
Other
0.5
8.4
8.3
9.0
5.6
5.7
5.5
3.1
8.9
4.9
5.5
5.5
3.2
11.5
11.5
8.3
5.8
8.9
8.7
6.8
7.7
2.7
-------
APPENDIX H-8 (CONTINUED). MINIPLANT SOLIDS ANALYSIS
Weight Percent
Run No.
43.5
45
46.1-.2
46.3
46.4
47
Source
Rejected Solids #8
Second Cyclone #11
Flue Gas Particulates
Final Bed
Second Cyclone #6
Second Cyclone #9
Second Cyclone #17
Second Cyclone #22
Second Cyclone #27
Second Cyclone #30
Second Cyclone #33
Second Cyclone #36
Second Cyclone #37
Second Cyclone #39
Flue Gas Particulates
Final Bed
Final Bed
GBF
Second Cyclone #7
Second Cyclone #8
Rejected Solids #4
Rejected Solids #5
Second Cyclone #10
Rejected Solids #6
Primary Cyclone
Final Bed
Second Cyclone #1
Second Cyclone #2
Second Cyclone #5
Flue Gas Particulates
Primary Cyclone
Final Bed
Ca
26.0
9.1
5.7
27.5
5.8
6.5
8.2
8.1
5.6
4.3
3.4
3.8
3.0
3.3
5.9
32.6
26.3
8.3
5.4
4.5
26.2
27.4
5.3
26.1
5.6
24.4
5.2
7.4
7.8
4.6
13.5
29.7
S
10.4
5.1
9.3
12.4
1.9
2.7
2.7
2.5
2.0
1.5
1.1
1.2
1.0
0.9
—
10.9
12.4
2.8
2.9
4.6
11.4
10.5
2.5
11.8
—
13.1
—
—
—
2.7
5.2
12.6
S04
40.4
16.0
28.4
34.7
7.0
10.7
8.1
7.4
5.3
4.1
2.9
2.9
2.6
2.7
17.9
28.7
36.7
9.7
7.5
7.1
34.5
33.5
7.5
35.1
10.6
37.6
12.8
8.6
7.2
7.1
12.9
35.1
CQ^
2.6
0.4
—
2.8
0.3
0.5
0.5
1.0
0.6
0.1
0.1
0.1
0.1
0.1
0.1
2.3
7.0
7.5
1.5
3.8
5.3
7.2
1.5
7.8
1.1
10.5
2.8
1.1
1.7
0.9
1.5
13.6
Total C Mg Other
1.4 10.2 Na = .04
1.1 4.9 Na=0.4
2.6 2.2 Na=0.8
0.5 11.3
13.9
12.8
8.4
13.6
11.6
12.4
13.5
12.4
15.5
11.7
1.7
0.8
1.5
7.1
16.3
17.3
1.9
2.5
8.5
2.2
1.8
2.4
36.6
12.4
8.3
36.7
3.0
4.4
-------
APPENDIX H-8 (CONTINUED). MINIPLANT SOLIDS ANALYSIS
Run No.
50.3
50.4
Weight Percent
50.5
51
Source
Second Cyclone #43
Second Cyclone #46
Rejected Solids #25
Rejected Solids #28
Second Cyclone #49
Second Cyclone Mix #1
Second Cyclone Mix #2
Rejected Solids #31
Rejected Solids Mix #1
Rejected Solids Mix #2
Flue Gas Particulates
Second Cyclone #61
Second Cyclone #63
Second Cyclone #65
Rejected Solids #44
Rejected Solids #46
Rejected Solids #47
Flue Gas Particulates
Second Cyclone #4
Second Cyclone #5
Second Cyclone #6
Second Cyclone #7
Rejected Solids #1
Rejected Solids #2
Flue Gas Particulates
Flue Gas Particulates
Initial Bed
Bed Probe #1 (> 100 Mesh)
Bed Probe #1 ( < 100 Mesh)
Bed Probe #2
Bed Probe #3
Ca
8.8
7.7
23.5
22.5
7.6
8.7
7.5
22.4
20.0
19.9
4.0
7.7
8.9
6.5
17.3
21.8
21.8
4.1
5.0
4.5
4.8
3.8
21.9
9.6
4.0
3.6
22.2
18.4
19.7
19.2
20.6
S
5.0
4.0
6.8
7.2
4.3
4.9
4.5
6.0
7.5
7.5
3.1
5.2
5.0
4.8
6.7
7.8
8.5
1.2
2.1
2.4
4.1
2.0
7.6
7.8
4.3
4.7
9.9
11.3
9.4
8.9
7.0
SQ4
14.7
13.2
21.9
22.8
12.9
14.0
12.8
19.1
18.0
22.0
10.0
14.6
14.6
14.0
21.7
24.2
24.9
6.4
6.7
6.1
6.6
5.5
26.3
24.9
15.7
13.7
35.1
32.3
29.3
28.0
22.4
C03
2.1
1.9
22.3
21.6
1.6
2.1
2.0
23.7
14.8
14.5
0.0
2.7
2.0
1.5
20.1
19.7
17.9
0.0
1.9
1.8
2.0
1.4
21.2
23.8
0.3
0.1
14.7
7.7
17.5
17.3
23.0
Total C
3.7
3.8
1.4
1.3
4.6
3.3
3.6
1.5
3.0
2.9
2.0
3.1
2.6
2.4
1.6
1.4
1.0
1.7
6.6
9.0
12.5
9.8
3.8
4.7
3.6
3.1
3.8
1.6
2.8
3.0
3.2
Mg
4.4
4.3
13.5
13.0
4.0
4.9
4.1
13.1
16.9
12.1
1.6
4.4
4.8
3.5
9.6
13.1
13.0
2.0
2.6
2.3
2.4
2.1
12.9
5.6
1.7
1.6
12.9
10.6
12.1
12.3
12.3
Other
Na = 0.6%
VT f\ f\&/
Na = 0.0%
TVT f\ O Of
Na = 0-3%
Na = 0.5%
-------
APPENDIX H-8 (CONTINUED). MINIPLANT SOLIDS ANALYSIS
Run No.
48
Weight Percent
NJ
i-1
00
50.1-50.5
50.1A
50. IB
50.2
Source
Second Cyclone #4
Second Cyclone #5
Second Cyclone #6
Second Cyclone #7
Flue Gas Particulates
Flue Gas Particulates
Primary Cyclone
Initial Bed
Final Bed
Initial Bed
Final Bed
Primary Cyclone
Flue Gas Particulates
Second Cyclone Mix #1
Second Cyclone Mix #2
Second Cyclone #12
Second Cyclone #15
Second Cyclone #17
Rejected Solids #1
Rejected Solids #2
Rejected Solids #3
Flue Gas Particulates
Second Cyclone #25
Second Cyclone #27
Second Cyclone #28
Rejected Solids #8
Rejected Solids #10
Flue Gas Particulates
Ca
5.2
4.7
4.7
4.3
3.6
3.7
11.0
17.2
22.1
22.9
16.4
14.9
3.8
7.3
8.3
6.5
5.7
5.2
23.4
23.4
22.3
4.4
7.7
6.0
5.8
20.3
22.2
3.5
_S
3.1
2.9
2.6
2.8
5.1
6.8
5.4
13.6
7.1
7.4
8.4
7.3
5.3
4.6
3.8
4.3
3.2
3.1
9.4
7.9
7.9
6.7
4.0
3.3
3.3
7.6
7.5
4.6
S04
9.8
9.2
8.9
9.1
19.1
17.8
16.2
42.2
23.5
21.7
25.7
23.4
12.3
13.4
13.0
11.7
10.6
9.4
28.2
23.9
24.5
21.8
11.9
9.1
9.5
23.5
21.3
14.3
C03
0.7
0.5
0.3
0.4
0.0
0.0
1.2
2.5
22.9
22.9
20.6
2.3
0.9
5.3
4.7
1.3
0.9
0.7
11.2
20.4
21.7
0.2
3.5
2.8
3.0
16.1
22.7
0.1
Total C
8.9
6.5
6.2
6.8
2.9
1.8
2.2
0.8
4.1
3.9
2.3
1.4
1.9
5.7
7.1
8.8
8.2
7.2
2.3
3.7
1.6
3.1
9.3
9.4
9.8
2.0
2.3
3.5
Mg
2.3
2.3
2.4
2.4
1.5
1.5
6.1
9.4
13.2
13.7
9.8
9.5
1.9
3.4
3.9
3.1
2.7
2.4
6.5
9.9
12.7
1.5
4.3
3.2
2.8
11.8
13.0
1.3
Other
Fe = 10.1
Fe = 10.5
Fe = 10.4
Fe = 10.1
Fe = 7.4
Fe = 7.0
Fe = 11.7
Fe = 2.1
Fe = 0.3
-------
APPENDIX H-8 (CONTINUED). MINIPLANT SOLIDS ANALYSIS
Weight Percent
Run No.
51 (Cont)
52
KJ
54
55
Source
Bed Probe #4
Bed Probe #5
Bed Probe #6
Bed Probe #7
Final Bed
Second Cyclone #7
Second Cyclone #9
Second Cyclone #11
Rejected Solids #2
Rejected Solids #4
Rejected Solids #5
Flue Gas Particulates
Initial Bed
Bed Probe #3
Bed Probe #7
Bed Probe #10
Final Bed
Second Cyclone #4
Second Cyclone #5
Bed Probe #4
Bed Probe #5
Second Cyclone #1
Second Cyclone #2
GBF
Primary Cyclone
Final Bed
Ca
20.3
22.9
20.0
21.1
21.3
6.8
5.6
7.0
16.9
21.9
17.1
4.0
23.1
22.2
20.3
20.6
21.7
7.9
5.0
32.6
33.0
12.8
5.4
4.1
15.2
23.5
S
8.2
5.8
8.4
9.0
8.1
3.9
3.9
3.9
7.5
5.6
7.6
3.8
6.4
7.0
8.7
7.7
8.9
3.0
2.8
14.0
14.3
5.5
2.4
3.5
7.2
14.6
S04
29.4
18.0
26.8
24.9
20.8
11.5
10.7
11.9
23.4
20.4
26.6
8.6
26.7
17.4
24.4
22.6
29.2
8.2
5.7
43.4
40.8
15.1
7.8
9.9
21.8
46.3
C03
16.4
25.7
18.7
17.4
22.3
0.7
1.0
1.2
15.2
24.5
21.3
0.4
11.6
24.7
19.3
20.8
16.1
0.6
0.2
3.9
4.7
0.3
0.2
0.4
0.6
0.6
Total C
2.4
3.1
2.5
4.2
3.1
3.9
3.5
2.6
1.8
1.8
1.7
2.8
3.5
3.6
2.8
2.2
4.0
12.4
15.5
0.6
0.7
14.3
17.1
1.3
0.6
1.4
Mg Other
12.8
13.1
12.0
13.4
12.8
3.9
3.3
4.3
10.8
13.2
9.9
1.6
12.8
14.3
12.8
13.4
13.3
1.6
1.3
0.4
0.4
4.9
1.6
1.2
5.7
7.3
-------
APPENDIX H-8 (CONTINUED). MINIPLANT SOLIDS ANALYSIS
Weight Percent
Run No.
56
57
M
O
59
Source
Second Cyclone #1
Initial Bed
Bed Probe #1
Bed Probe #2
Final Bed
Primary Cyclone
Second Cyclone #4
Second Cyclone #5
Initial Bed
Bed Probe #4
Primary Cyclone
Second Cyclone #8
Second Cyclone
Second Cyclone
Flue Gas Particulates
Flue Gas Particulates
Primary Cyclone
Final Bed
Ca
13.8
24.7
22.6
21.8
20.8
19.6
9.3
8.3
24.2
19.7
16.9
4.4
4.5
3.3
3.3
1.9
11.6
15.6
S
5.6
10.4
8.5
11.2
14.3
10.4
4.6
4.1
11.7
12.5
9.1
1.6
1.7
1.5
4.1
3.1
4.9
10.3
SO&
17.6
34.2
25.5
34.8
44.3
29.3
13.0
12.8
40.6
38.5
27.0
4.5
5.0
2.9
14.4
10.2
13.1
31.0
CQ-j
9.5
6.2
18.6
12.5
0.6
4.4
0.3
0.3
8.6
3.1
0.5
—
—
—
—
—
0.4
0.1
Total C
10.3
0.9
1.7
1.5
0.5
1.1
4.0
6.2
1.9
0.6
0.8
3.1
5.2
2.8
1.9
1.2
0.4
1.0
Mg Other
7.8
10.0
12.5
11.4
6.2
11.6
5.0
4.5
5.3
7.1
10.5
2.4
2.4
1.8
0.9
0.6
7.7
9.1
-------
APPENDIX H-9. MINIPLANT SOLIDS COMPOSITION
Run
Number
41.1
41.2
43.1
43.2
43.3
43.4
43.5
45
Composition (wt.
Source
Sec. Cyclone
Final Bed
Sec. Cyclone
Sec. Cyclone
Final Bed
Sec. Cyclone
Flue Gas Part.
Final Bed
Sec. Cyclone
Flue Gas Part.
Reject Solids
Sec. Cyclone
Flue Gas Part.
Final Bed
Reject Solids
Sec. Cyclone
Final Bed
Sec. Cyclone
Flue Gas Part,
46.2
46.3
46.4
Final Bed
GBF
Sec . Cyclone
Reject Solids
Final Bed
Reject Solids
Sec. Cyclone
Prim. Cyclone
47 Final Bed
Sec. Cyclone
Flue Gas Part.
Prim. Cyclone
48 Final Bed
Sec. Cyclone
Flue Gas Part.
Prim. Cyclone
c
(Combustible)
4.7
0.0
8.3
0.1
0.0
1.6
0.5
0.3
1.8
4.1
0.5
0.3
1.4
0.0
0.9
1.0
0.4
12.5
1.7
0.1
5.6
16.2
0.8
0.3
0.6
8.2
1.6
1.7
18.7
36.5
2.7
0.0
7.2
2.7
0.6
Ash
74
26
57
41
12
55
63
18
59
59
20
45
59
12
10
64
28
75
75
27
69
69
30
26
28
77
81
18
62
50
67
11
74
74
60
Sorbent
48
74
35
59
89
43
37
81
39
37
79
55
40
88
89
35
71
12
23
72
25
15
69
73
72
15
17
81
19
13
31
89
19
23
40
Sorbent Portion
Composition (Mole %)
CaO
48
64
63
21
2
7
(1)
52
13
34
16
41
28
24
59
50
24
-9
-2
31
7
24
22
8
20
23
53
-13
12
CaC03
4
11
7
7
49
10
—
3
5
10
6
7
7
3
5
4
18
(2) 60
41
14
29
20
19
13
31
18
8
69
7
CaS04
48
26
30
72
49
83
—
45
82
56
78
53
65
73
37
46
58
49
61
55
64
56
59
80
49
58
40
44
81
31
61
221
-------
APPENDIX H-9 (CONTINUED). MINIPLANT SOLIDS COMPOSITION
Run
Number
50
50.1A
50. IB
50.2
50.3
50.4
50.5
51
Composition (wt,
Source
52
54
55
56
Final Bed
Sec. Cyclone
Sec. Cyclone
Reject Solids
Flue Gas Part.
Sec. Cyclone
Reject Solids
Flue Gas Part.
Sec. Cyclone
Reject Solids
Sec. Cyclone
Reject Solids
Flue Gas Part.
Sec. Cyclone
Reject Solids
Flue Gas Part.
Prim. Cyclone
Final Bed
Sec. Cyclone
Reject Solids
Flue Gas Part.
Bed Probe
Final Bed
Sec. Cyclone
Reject Solids
Flue Gas Part.
Bed Probe
Sec. Cyclone
Bed Probe
Final Bed
Sec. Cyclone
Prim. Cyclone
Flue Gas Part.
Final Bed
Sec. Cyclone
Prim. Cyclone
Bed Probe
c
(Combustible)
0.0
5.4
7.9
0.0
3.1
8.9
0.0
3.5
3.3
0.0
3.4
0.0
2.0
2.3
0.0
1.7
0.0
0.0
9.1
0.0
3.3
0.0
0.8
3.1
0.0
2.7
0.0
13.8
0.0
1.2
17.1
0.5
1.2
0.4
8.4
0.2
0.0
Ash
26
63
70
18
70
66
18
78
65
14
66
19
81
66
21
84
41
16
74
25
76
17
10
71
23
82
41
69
15
15
66
50
83
23
38
25
14
Sorbent
76
32
22
83
27
26
83
19
32
89
31
82
17
31
82
14
58
85
17
75
20
84
89
26
79
16
61
17
85
84
17
49
16
77
54
74
88
Sorbent Portion
Composition (Mole %)
CaO
-49
-14
13
2
2
-5
13
-4
14
4
4
-12
24
-11
16
-13
-13
-6
17
-25
-11
50
38
16
37
38
9
1
23
-3
CaC03
84
43
11
51
32
61
16
64
16
57
18
63
10
70
26
71
59
50
10
73
69
4
9
2
3
2
2
46
15
47
CaS04
65
71
76
46
66
44
71
41
70
40
78
49
65
41
58
42
54
56
73
53
43
45
54
82
60
60
89
53
62
57
222
-------
APPENDIX H-9 (CONTINUED). MINIPLANT SOLIDS COMPOSITION
Run
Number
Source
Composition (wt. %)
C Ash Sorbent
(Combustible)
Sorbent Portion
Composition (Mole %)
CaO CaCO^ CaSOA
57 Sec. Cyclone
Prim. Cyclone
Bed Probe
59 Final Bed
Sec. Cyclone
Flue Gas Part.
Prim. Cyclone
5.1
0.7
0.0
1.0
3.6
1.5
0.3
64
35
26
36
83
83
59
31
64
74
63
13
15
40
37
32
8
17
50
51
2
2
10
0
8
61
67
81
83
42
47
(1) Flue gas particulates sulfate during sampling at lower temperatures
giving erroneous results.
(2) Negative values caused by calculation procedure and possible error
in analyses.
223
-------
APPENDIX H-10. MINIPLANT SAMPLE SHIPMENTS
Requestor
Brookhaven National Labs
Uptin, NY
Curtiss-Wright
Wood-Ridge, NJ
Dravo Corp.
Pittsburgh, PA
EPA
Research Triangle Park, NC
ERE
Florham Park, NJ
Baton Rouge, LA
Baytown, Texas
Sample Description
Foster Wheeler Co.
Livingston, NJ
General Electric Co.
Schenectady, NY
MIT
Cambridge, MA
Fly Ash
Primary Cyclone
Fly Ash
Bed
Bed
Flue Gas Particulates
Limestone
Fly Ash
Illinois Coal Fines
Limestone
Limestone Fines
Limestone
Dolomite
Aragonite
Limestone
Dolomite
Primary Cyclone
Fly Ash
Fly Ash
Dolomite
Illinois Coal Fines
Illinois Goal No. 6
Sorbent
Dolomite
Dolomite
Dolomite
Run No.
43.5
43.5
37
Dolomite 23
Alumina 28
Limestone 45
Limestone 47
Dolomite 37
Dolomite 51
Amount
5 Ibs.
5 Ibe.
110 Ibs.
Date
3/14/77
3/14/77
10/29/76
5 gallons 3/4/76
2 gallons 4/1/76
13.5 grams 1/7/77
—
Limestone
Dolomite
—
—
—
—
—
—
—
—
45
43
—
—
—
—
—
—
—
2 gallons
360 Ibs.
360 Ibs.
10 Ibs.
900 Ibs.
1 Ib.
1 Ib.
1 Ib.
5 Ibs.
5 Ibs.
9/21/76
11/2/76
11/2/76
1/7/77
5/14/76
5/23/77
5/23/77
5/23/77
8/2/77
8/2/77
5 Ibs.
240 Ibs.
168 Ibs.
25 Ibs.
100 Ibs.
100 Ibs,
1/7/77
3/2/77
6/6/77
7/13/76
2/3/77
2/14/77
-------
APPENDIX H-10 (CONTINUED). MINIPLANT SAMPLE SHIPMENTS
Requestor
National Gypsum Co.
Buffalo, NY
O.R.N.L.
Oak Ridge, TN
Pratt & Whitney Co.
Middletown, CT
Ralph Stone & Co., Inc.
Los Angeles, CA
Sample Description
Bed
Research Cottrell
Bound Brook, NJ
T.V.A.
Muscle Shoals, AL
Fly Ash
Fly Ash
Fly Ash
Bed
Fly Ash
Bed
Fly Ash
Bed
Fly Ash
Flue Gas Particulates
*Flue Gas Particulates
*(under non-isokinetic conditions)
Bed
Bed
Bed
Bed
Bed
Bed
Sorbent
Limestone
Dolomite
Limestone
Dolomite
Limestone
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Limestone
Limestone
Limestone
Dolomite
Limestone
Dolomite
Run No.
19.6
38.1
45
43.4
19.7
37
43.5
43.5
37
37
50.1
50.1-50.5
19.3
30.1
30.2
32.3
19.5
43.5
Amount
1 gallon
2 Ibs.
20 Ibs.
5 Ibs.
200 Ibs.
200 Ibs.
100 Ibs.
150 Ibs.
100 Ibs.
100 Ibs.
25 grams
200 grams
500 grams
500 grams
500 grams
500 grams
7 Ibs.
7 Ibs.
Date
1/19/77
4/23/76
2/9/77
4/8/77
4/27/76
4/27/76
10/22/76
10/22/76
12/15/76
12/15/76
5/17/77
5/17/77
3/4/76
3/4/76
3/4/76
3/4/76
12/21/76
12/21/76
-------
APPENDIX H-10 (CONTINUED). MINIPLANT SAMPLE SHIPMENTS
Requestor
Sample Description
Westinghouse Research Labs. Bed
Pittsburgh, PA Fly Ash
Bed
Bed
Fly Ash
Fly Ash
Flue Gas Particulates
Fly Ash
Bed
Bed
Fly Ash
Regenerator Bed
Dolomite
Dolomite
Limestone
Limestone
Limestone
Limestone
Dolomite
Dolomite
Dolomite
Limestone
Limestone
Limestone
27
27
19.6
30.2
19.6
26
34
43.3
43.3
45
45
45
50 Ibs.
50 Ibs.
50 Ibs.
50 Ibs.
2 Ibs.
2 Ibs.
100 grams
300 Ibs.
275 Ibs.
20 Ibs.
35 Ibs.
50 Ibs.
3/8/76
3/8/76
3/8/76
3/8/76
3/8/76
3/8/76
4/12/76
11/18/76
11/18/76
1/7/77
1/7/77
1/7/77
N>
-------
APPENDIX J-l
MODIFIED FILTER BED
227
-------
APPENDIX J-2
MODIFIED FILTER ELEMENT
FLUIDIZING GRID
228
-------
APPENDIX J-3
GBF PRESSURE VESSEL
-------
APPENDIX J-4
GBF PRESSURE VESSEL LINING
K>
CO
o
-------
APPENDIX J-5
FILTER PRESSURE VESSEL
(SIDE VIEW)
231
-------
APPENDIX J-6
GBF PIPING ARRANGEMENT
-------
APPENDIX J-7
PREHEAT BURNER SYSTEM SCHEMATIC
Pressure Gauge
High Limit Gas
Low Limit
10
OJ
Pressure
Vessel
Lim
Gas
Flame
Scanner
N
i
— n f r—>
it
iting Ori
Valve
\
IP
r •3-K'K
/ V"
Burner J?L
v<
G;
fi
jriable
js Reg
:e
>
V*
1
•H\
Ignition
Transformer
Pressure Switch W ' pressu,.e
^'° / A V Switch
ulator / i J3 Y
X^l 1 IS^l ^ rXl
r
S--, Bilking V"?' "
w , Reset
Valve Shutoff Valve
JX^--- \s&
—--Pressure Gauges -""i
t
Motorized
Air Valve
p
XI 1
i/lain Gas
Shutoff \
., L
9"^s
Natural Gas
(2 psig)
telve
-ow Limit
Ur Pressure
>witch
Blower
-------
APPENDIX J-8
GRANULAR BED FILTER-BLOW. BACK SYSTEM FLOW SCHEMATIC
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
EPORT NO
EPA-600/7-78-069
!. RECIPIENT'S ACCESSIOf+NO.
4. TITLE AND SUBTITLE MiniplantStudies Qf PrgSSUHZed
Fluidized-bed Coal Combustion: Third Annual Report
5. REPORT DATE
April 1978
6. PERFORMING ORGANIZATION CODE
R.C.Hoke, R.R.Bertrand, M.S.Nutkis, L.A
Ruth, M. W. Gregory, E. M. Magee M. D. Loughnane.
R. J.Madon A.R. Garabrant, and M. Ernst
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION 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
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
Annual: 8/76 - 8/77
14. SPONSORING AGENCY CODE
EPA/600/13
is. SUPPLEMENTARY NOTES jERL-RTP project officer is D. Bruce Henschel, Mail Drop 61,
919/541-2825. EPA-600/7-76-Oil and EPA-600/7-77-107 are earlier annual reports.
. ABSTRACT
repOrt presents further results of studies of the environmental aspects
of the pressurized fluidized-bed coal combustion process , using the 218 kg coal/hr
'miniplant' continuous -combustion/sorbent-regeneration system (0. 63 MW equiva-
lent), and a 13 kg coal/hr bench-scale system. Previous combustion studies on the
miniplant combustor were extended to investigate emissions of SO2, SO3, NOx, and
particulates during combustion of a high-sulfur coal, and with the use of precalcined
limestone as the SO2 sorbent. Percentage SO2 removals obtained with the high-sulfur
coal were similar to earlier ones with intermediate-sulfur coal. Precalcined lime-
stone proved to be as effective as dolomite (on a Ca/S molar basis) in removing SO2.
The performance of the miniplant regenerator was demonstrated in an uninterrupted
125-hr run with continuous circulation of sorbent between the combustor and the
regenerator. A granular bed filter for high temperature/pressure flue gas particu-
late removal was installed on the miniplant: initial shakedown was completed with a
24-hr continuous run. Sampling was completed on the miniplant combustor for com-
prehensive analysis of emissions.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Air Pollution
Coal
Combustion
Fluidized Bed
Processing
Pressurizing
Limestone
Sorption
Regeneration
Sulfur Oxides
Nitrogen Oxides
Dust
Air Pollution Control
Stationary Sources
Particulate
13 B
21D
21B
13H,07A
08G
07D
07B
11G
13. DISTRIBUTION STATEMENT
Unlimited
Unclassified
234
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |