Office.of Research and Development Laboratory
Research Triangle Park, North Carolina 27711
EPA-600/7-77-107
_ . . . r\TT
SeteiTIDGr 19/7
STUDIES OF THE PRESSURIZED
FLUIDIZED-BED COAL
COMBUSTION PROCESS
Interagency
Energy-Environment
Research and Development
Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into seven series. These seven broad categories
were established to facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously planned to foster
technology transfer and a maximum interface in related fields. The seven series
are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the effort
funded under the 17-agency Federal Energy/Environment Research and Development
Program. These studies relate to EPA's mission to protect the public health and welfare
from adverse effects of pollutants associated with energy systems. The goal of the
Program is to assure the rapid development of domestic energy supplies in an environ-
mentally-compatible manner by providing the necessary environmental data and
control technology. Investigations include analyses of the transport of energy-related
pollutants and their health and ecological effects; assessments of, and development
of, control technologies for energy systems; and integrated assessments of a wide
range of energy-related environmental issues.
REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect the
views and policies of the Government, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.
-------
EPA-600/7-77-107
September 1977
STUDIES OF THE PRESSURIZED
FLUIDIZED-BED COAL
COMBUSTION PROCESS
by
R.C. Hoke, R.R. Bertrand, M.S. Nutkis, D.D. Kinzler,
LA. Ruth, M.W. Gregory, and E.M. Magee
Exxon Research and Engineering Co.
P.O. Box 8
Linden, New Jersey 07036
Contract No. 68-02-1312 and -1451
Program Element No. EHE623A
EPA Project Officer: D. Bruce Henschel
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, 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"
unit. Supporting studies were also carried out in a smaller batch combustion
unit. Emissions of S02, 803, NOX, CO and particulates from the combustors were
measured as a function of combustion conditions. Dolomite and limestone
requirements needed to keep S02 emissions within new source performance
standards were estimated based on the experimental results. SO^ and CO emis-
sion levels were generally low. NOX emission levels were well within the
current new source performance standard. Particulate emission exceeded the
new source performance standard and will require the use of an additional
particulate removal device beyond two stages of conventional cyclones to meet
the emission standard. Carbon combustion efficiencies were also measured.
Levels of 99% were achieved at higher combustion temperatures.
Shakedown of the miniplant regenerator section was completed by a run
in which the combustor and regenerator were both operated and solids were
transferred between the combustor and regeneration sections continuously
for a 24 hr. period. A solids transfer system was developed prior to the
run, permitting the continuous transfer of the hot solids between the com-
bustor and regenerator.
Additional studies are planned to develop additional combustion data,
to study the removal of particulates from the combustor flue gas at high
temperature and high pressure, to demonstrate continuous combustion and
sorbent regeneration for a 100 hr. period and make a comprehensive analysis
of potentially harmful emissions from the miniplant unit.
This work was submitted in fulfillment of Contract Numbers 68-02-1312
and 68-02-1451 by Exxon Research and Engineering Company under sponsorship
of the Environmental Protection Agency. Work was completed in July 1976.
iii
-------
CONTENTS
Page
Abstract i:Li
vi
List of Figures
ix
List of Tables
xi
Acknowledgements
Sections
I Summary 1
II Introduction °
III Combustion Studies 13
Miniplant Combustor 13
Experimental Results -^
Batch Combustor 81
IV Miniplant Regenerator Shakedown 104
Equipment
Materials HI
Batch Operation 112
Coupling of Regenerator to Combustor 125
V Discussion of Results 137
Comparison of Batch Unit and Miniplant Results 137
VI Continuing Studies 142
Combustion Studies 142
Flue Gas Particulate Removal 142
Comprehensive Analysis of Emissions 146
Regeneration Studies 147
VII References 148
VIII List of Publications and Patent Memoranda 149
iv
-------
CONTENTS (CONTINUED)
Page
IX Appendix 15-1
A. Analytical Techniques 152
B. Miniplant Fluidized Bed Coal Combustion Run Summary 153
C. Determination of S02 and S03 by Wet Chemistry 168
D. Entrainment Rates for Grove No. 1359 Limestone
with Limited Calcination 170
E. Entrainment Rates for Grove No. 1359 Limestone
with Extensive Calcination 171
F. Entrainment Rates for Pfizer No. 1337 Dolomite 172
G. Particle Size Distribution - Miniplant Used
Limestone No. 1359 Sorbent 173
H. Particle Size Distribution - Miniplant Used
Dolomite No. 1337 Sorbent 174
I. Particle Size Distribution - Miniplant Secondary
Cyclone Capture 175
J. Particle Size Distribution - Miniplant Flue
Gas Particulates 177
K. Miniplant Solids Analyses 178
L. Miniplant Solids Composition 186
M. Summary of Batch Combustor Operating Conditions 190
N. Summary of Batch Combustor Emissions Data 193
0. Batch Fluidized Bed Combustor CO Emissions 195
P. Batch Combustor Particle Size Distribution -
Overhead Samples 196
Q- Batch Combustor Bed and Overhead Solids Analysis 197
R. Sulfur Balances for Batch Combustor 199
S. Calcium Balances for Batch Combustor 200
-------
FIGURES
No.
II-l Pressurized Fluidized Bed Coal Combustion System 9
II-2 Batch Fluidized Bed Coal Combustion Unit 10
II-3 Exxon Fluidized Bed Combustion Miniplant 12
I1I-1 Exxon Fluidized Bed Combustion Miniplant 14
III-2 Coal and Limestone Feed System 15
III-3 Combustor Vessel 17
III-4 Flue Gas Sampling System 18
III-5 Miniplant Particulate Sampling System 19
III-6 Coal Particle Size Distribution 23
III-7 Limestone No. 1359 Particle Size Distribution 25
III-8 Erosion Damage at the Upper U Bend of Coil IB 31
III-9 Baffled Coil After 15 Hours of Operation 32
111-10 Miniplant Combustor Temperature Profile 33
III-ll Miniplant S02 Emissions for Dolomite No. 1337 38
111-12 Miniplant S02 Retention Vs. Ca/S Ratio for Dolomite No. 1337 39
Hl-13 Activity Vs. Calcium Utilization for Dolomite 42
111-14 Effect of Temperature and Sorbent Particle Size on
S02 Retention - Dolomite No. 1337 43
111-15 Effect of Gas Residence Time on S02 Retention
Dolomite No. 1337 44
111-16 Effect of Gas Residence Time on Ca/S Ratio Required to
Meet EPA S02 Emission Standard with Dolomite 46
111-17 S02 Emission Vs. Ca/S Ratio for Limestone No. 1359 48
111-18 S02 Retention Vs. Ca/S Ratio for Limestone No. 1359 49
111-19 Comparison of the Degree of Sorbent Calcination with
the Dissociation Pressure of CaCCU 51
111-20 Model for the Calcination and Sulfur Reactions in a
Calcining Environment 52
111-21 Sulfur Retention Vs. Effective Ca/S Ratio Which Accounts
for Different Limestone Calcination Levels 54
111-22 Activity Vs. Calcium Utilization for Limestone No. 1359 55
VI
-------
FIGURES (CONTINUED)
No. Page
111-23 Comparison of Dolomite No. 1337 and Limestone No. 1359
As S02 Sorbents on a Mass Feed Rate Basis 59
111-24 Comparison of Monitored and Wet Chemistry
S02 Concentrations 61
111-25 NOX Emissions Vs. Excess Air 63
111-26 NOX Emissions Data - Run No. 34 64
111-27 Carbon Monoxide Emission Vs. Temperature 66
111-28 Effect of Gas Velocity and Ca/S Ratio on
Elutriation of Dolomite No. 1337 68
111-29 Combustion Efficiency Vs. Temperature 76
111-30 Heat Transfer Coefficients Vs. Temperature 80
111-31 Batch Fluidized Bed Coal Combustion Unit 82
111-32 S02 Emissions Vs. Ca/S Ratio - Eastern Coal and
Limestone No. 1359 88
111-33 S02 Emissions Vs. Ca/S Ratio - Eastern Coal
and Tymochtee Dolomite 89
111-34 S02 Retention Vs. Ca/S - Eastern Coal -
Limestone No. 1359 90
111-35 S02 Retention Vs. Ca/S Ratio - Eastern Coal -
Tymochtee Dolomite 91
111-36 S02 Retention Vs. Effective Ca/S Ratio -
Limestone No. 1359 93
111-37 S02 Retention at Constant Residence Time
Limestone No. 1359 94
111-38 S02 Retention at Constant Residence Time
Tymochtee Dolomite 95
111-39 S02 Retention Vs. Ca/S Ratio - Illinois No. 6
Coal - Limestone No. 1359 96
111-40 Batch Unit NOX Emissions 98
111-41 Batch Unit NOX Emissions Including Western Coal Results 99
111-42 Combustion Efficiency Vs. Excess Air and Temperature 101
\
111-43 Effect of Coal Feeding on Combustion Efficiency 102
vii
-------
FIGURES (CONTINUED)
No- Page.
IV-1 Miniplant Regenerator Air and Fuel Locations 105
IV-2 Miniplant Regenerator Off-Gas Handling System 107
IV-3 Miniplant Regenerator Off-Gas Sampling System 109
IV-4 Regenerator Fluidizing Grid 11°
IV-5 Typical Temperature Profile in Miniplant Regenerator Bed 117
IV-6 Original Solids Transfer System:
Combustor-Regenerator Solids Transfer Line 127
IV-7 Transfer System Pressure Balance 129
IV-8 Miniplant Solids Transfer System 131
V-l Comparison of Sulfur Dioxide Retention Measured in
Miniplant and Batch Units - Dolomite Sorbent 138
V-2 Comparison of Combustion Efficiencies Measured in
Miniplant and Batch Units 139
V-3 Comparison of NOX Emissions Measured in
Miniplant and Batch Units 140
VI-1 Ducon Granular Bed Filter 145
viii
-------
TABLES
No. Page
III-l Miniplant Coal Analyses 21
III-2 Properties of Sorbents Used in Miniplant Variables Study 24
III-3 Summary of Miniplant Test Program 27
III-4 Miniplant Combustion Process Variable Studies 36
III-5 Comparison of FBC Pressurized Units 40
III-6 45
III-7 Utilization and Calcination of Limestone No. 1359 50
III-8 Comparison of Sulfur Retentions at 930 kPa
and 600 kPa with Limestone No. 1359 56
III-9 Results of Runs at Turndown Conditions 57
111-10 Sorbent Requirement to Meet the EPA SC^ Emission Standard 60
III-ll Sulfur Balances - Inert Bed Run Series 60
111-12 Heat Loss Due to Incomplete Combustion of CO to C02 65
111-13 Sorbent Elutriation Losses 67
111-14 Miniplant Flue Gas Particulate Sampling Summary 70
111-15 Spent Sorbent Particle Size Distribution 71
111-16 Particle Size Distribution - Flyash 71
111-17 Particle Size Distribution - Fine Flue Gas Particulates 72
111-18 Typical Particulate Analyses 73
111-19 Typical Particulate Composition 74
111-20 Combustible Carbon Losses 77
111-21 Miniplant Overall Heat Transfer Coefficient
Measurements - Run 19.2 78
111-22 Heat Transfer Coefficients 79
111-23 Composition of Coals Used in Batch Fluidized Bed
Coal Combustion Program 84
111-24 Properties of Limestone and Dolomite 85
111-25 Desulfurization of Western Coal 97
111-26 Particulate Loadings. Batch Unit 103
IX
-------
TABLES (CONTINUED)
.No... Page
IV-1 Maximum Partial Pressures of SC>2 for Reduction of CaSO/,
IV-2 Regenerator Run Summary 121
IV-3 Fractional Regeneration and Sulfide Formation 122
IV-4 Sulfur Balances for Regeneration Runs 123
IV-5 Comparison of Measured and Equilibrium SC^ Concentrations 126
IV-6 Nominal Operating Conditions for Combustor and
Regenerator During Shakedown Runs 134
IV-7 Log of Events for 24 Hour Shakedown Run 134
IV-8 Shakedown Run Nominal Emission Levels from
Combustor and Regenerator 135
VI-1 Particulate Emission Control Requirements 143
-------
ACKNOWLEDGEMENTS
The authors wish to express their appreciation to the many individuals
who played major roles in the conduct of this program at Exxon Research and
Engineering Company. In particular, we wish to acknowledge the efforts of
H. R. Silakowski, the miniplant operations supervisor. His contributions
played a large part in the successful operation of the miniplant. We also
wish to acknowledge the efforts of the operating crew, T. C. Gaydos, R. E.
Long, G. E. Walsh, D. T. Ferrughelli, E. Hellwege, J. E. Bond, W. J. Spond,
J. Fowlks and our math clerk, S. Walther. We also wish to thank the
personnel of the Mechanical Division who contributed to the program, in
particular S. Pampinto, T. Sutowski, T. Morrison, D. J. Cecchini, R. A.
Van Sweringen, T. E. Artz, F. D. Huber, E. E. Poole, T. J. Morgan and
H. T. Oakley. In addition, we wish to acknowledge the help given by V. S.
Engleman and G. A. Gagliardo in assisting in the preparation of the report.
A special acknowlegement 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.
xi
-------
SECTION I
SUMMARY
The pressurized fluidlzed bed combustion of coal (PFBC) was studied in
two experimental units, a continuous unit, also called the miniplant, and a
smaller semi-batch unit. The study of the regeneration of sulfated S02 sor-
bent was also begun in the miniplant regenerator. The overall objectives of
the experimental combustion program were to characterize the SC>2, SOg, NOX,
CO and particulate emissions from a PFBC unit while varying both operating
conditions and the coal and sorbent type, to measure combustion efficiency
and heat transfer coefficients while varying operating conditions and
preceding objectives. The objective of the regeneration program during this
period was to shakedown the miniplant regenerator, developing procedures and
equipment as needed. The shakedown was completed by a continuous combustion/
regeneration run of 24 hours duration.
MINIPLANT COMBUSTION STUDIES
The miniplant combustion section consists of a combustor vessel,
refractory lined to an inside diameter of 32 cm (12.5 in). The overall
height is 10 m (33 ft). A number of vertical water-cooled tubes are mounted
in the combustor to remove the heat of combustion. Premixed coal and sorbent
are injected into the combustor at a single point 28 cm (11 in) above the
fluidized bed support grid. The combustor is capable of operating at pres-
sures 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 220 kg/hr (480 Ib/hr). Flue gas leaving
the combustor passes through two cyclones in series to remove particulate
matter. Particulates captured in the first cyclone are recycled to the com-
bustor to improve combustion efficiency. Particulates captured in the second
stage cyclone are rejected through a lock hopper. Spent sorbent is also
rejected from the combustor through a lock hopper system to maintain a con-
stant bed level in the combustor.
Runs were made with an Eastern bituminous Pittsburgh seam coal (Champion)
containing 2% sulfur and 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.
As of July 1976, the miniplant has accumulated over 1100 hours of coal
combustion time in operations of up to 240 hrs. duration. Approximately 110
runs have been completed, 40 during the shakedown phase and 70 during the
operating phase. Mechanical performance of the combustor was good. The coal
feeding system performed satisfactorily. The vertical cooling coils instal-
led to promote uniform temperatures in the combustor performed well after
modifications were made to prevent distortion and erosion. Uniform combustor
temperatures were achieved using the vertical coils with temperature dif-
ferences usually less than 35°C (65°F) across the expanded bed.
-------
The longest run duration was 240 hrs. This run, which was made in
November 1975, completed the demonstration of the miniplant combustor.
Continuous steam generation was maintained during the entire run while
several coal feeding interruptions totaling 4 hrs. did occur. The unit was
burning coal 98.5% of the time. During the last 101 hrs. of the run coal
injection was uninterrupted.
SC>2 retention studies were made with dolomite No. 1337 over a range of
operating conditions. The most important variable affecting SC>2 emissions
was the Ca/S feed molar ratio (moles of calcium fed in the sorbent to moles
of sulfur fed in the coal). The stoichiometric Ca/S ratio for the desulfur-
ization reaction is 1.0. In the experiments reported here, the Ca/S ratio
was varied from approximately 0.5 to 2.0, over which range the 862 retained
by the dolomite bed varied from approximately 40 to 95%. A desulfurization
reaction rate expression was developed which allowed the analysis of the
effects on S0£ retention of dolomite conversion (or utilization) levels and
gas phase residence time. The effect of dolomite utilization on the reaction
rate constant was developed using data obtained by a number of laboratories
operating a variety of FBC experimental units of differing sizes and
geometries. The agreement between results obtained in the various experi-
mental units was good. Increasing dolomite utilization was found to decrease
the reaction rate constant sharply. The effect of gas phase residence time,
calculated as the ratio of the expanded bed height to the superficial gas
velocity, was predicted using a first order reaction rate expression and the
predicted results were verified reasonably well by a series of runs made
over a residence time range of 0.8 to 2.8 s. The effect of residence time
is more pronounced at higher S02 retention levels and the magnitude is such
that, for 90% S02 retention, decreasing the residence time from 3 to 0.5 s
would require doubling the Ca/S ratio from 1.5 to about 3.0. The effect of
temperature on SC^ retention was measured by varying the combustor temperature
from 690 to 950°C (1270 to 1740°F). At temperatures between 840 and 950°C
(1540 to 1740°F), no effect of temperature was observed. S02 retention
levels measured at 690 to 760°C (1270 to 1400°F) were slightly lower than
those measured at the higher temperatures. The average sorbent particle size
was varied by a factor of two and no effect on S0£ retention was observed.
Pressure was varied from 600 to 930 kPa (6 to 9 atm abs.) and no effect was
found. Variation in the excess air level from approximately 5 to 100% had
no obvious effect on S02 retention.
S02 emissions were also measured with limestone No. 1359. Contrary to
the results seen- with dolomite sorbent, a marked effect of temperature was
found, with increasing temperature giving higher SC>2 retention levels. The
S02 retention levels were also lower than those observed using dolomite sor-
bent. These effects were due to the inability of the limestone to calcine
completely under pressurized combustion conditions. Calcination greatly
increases the porosity of the limestone, making the interior surface of the
stone more accessible to the S02 reactant. At higher temperature conditions
e.g., 925-950°C (1700-1740°F) the limestone underwent extensive calcination,'
and although the limestone was not as active as dolomite, it was considerably
more active than at the lower temperatures, i.e., 825-900°C (1520 to 1650°F),
where the stone was largely in the carbonate form. At much lower tempera-
tures, 670-760°C (1240-1400°F), the limestone was completely inactive. Pres-
sure and sorbent particle size were found to have no effect on S02 retention.
-------
Although an FBC utility boiler would normally be expected to operate in
the temperature range of about 850 to 950°C (1550 to 1750°F), operation at
temperatures down to about 700°C (1300°F) would be required to turn-down the
boiler output to match a decrease in the electrical power demand. A series
of runs was made using dolomite and limestone sorbent at temperatures near
700°C to determine the behavior of the FBC system at these lower temperatures.
Some runs were also made at temperatures as low as 690°C (1270°F) to determine
the lowest limit of operability. The minimum temperature at which combustion
was stable was 690°C. An attempt to decrease the temperature to 600°C
(1110°F) was unsuccessful. At that temperature, temperature control in the
combustor became erratic and carbon monoxide emissions in the flue gas
increased sharply, denoting poor combustion. The effect of the low tempera-
ture operation on SC^ emission control was noted previously. A slight
decrease in S02 retention was seen using dolomite sorbent at low temperatures.
However, limestone was completely inactive and therefore, cannot be used in
a pressurized FBC unit unless some means of increasing its activity under
low temperature "turn-down" conditions can be found. Precalcination of the
limestone is one possible way to do this and will be studied in the future.
As a result of the above studies, the sorbent requirements needed to
satisfy the current EPA new source performance standards for SO^ emissions
from a coal fired boiler (1.2 Ib SC>2/M BTU coal fired) can be estimated.
This was done for coals containing 2 to 5% sulfur using the data obtained
for the 2% sulfur coal and extrapolating to higher sulfur levels. Dolomite
requirements were found to be less than those for limestone when expressed
as the required Ca/S molar ratio. However, when expressed on a weight basis,
e.g., wt. of sorbent/wt. of coal, limestone requirements were found to be
slightly lower for coals containing up to 3% sulfur. For higher sulfur
content coals, dolomite requirements were found to be less. For example,
a coal containing 4% sulfur was estimated to require 34 kg limestone per
100 kg coal compared to 29 kg dolomite/100 kg coal. However, as cautioned
previously, limestone may not be suitable for pressurized FBC applications
unless precalcined limestone is found to be suitable for operation at the
low temperature "turndown" conditions.
803 emissions were found to be highly variable, averaging 6 ppm for one
series of runs, and 23 ppm for another with individual runs showing even
higher concentrations. No correlation was found with operating conditions
and the cause for the formation and variability of 863 emissions is not
understood. One possible explanation is sampling errors.
NOX emissions were measured and 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). Although the operating
conditions varied greatly, the only significant variables were excess air
and combustor temperature. The NOX emissions increased four fold, from 0.04
to 0.17 g/MJ over a 5 to 110% range of excess air. The temperature effect
in the 670 to 940°C (1240 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/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. utility boiler.
-------
CO emissions were low, generally in the range of 70 to 225 ppm at com-
bustor temperatures above 825°C (1500°F). As the temperature was reduced ^
below 825°C, CO emissions increased sharply to 300 to 800 ppm at 700 to 790 C
(1300 to 1400°F).
Particulate emissions and sorbent losses by elutriation were also
studied. Elutriation losses were found to increase with superficial gas
velocity, with dolomite showing significantly higher loss rates. Limestone
elutriation losses averaged about 10% of the limestone feed up to a super-
ficial velocity of 2.2 m/s (7 ft/sec), and were equivalent to a feed Ca/S
molar ratio of 0.24. Dolomite elutriation losses average about 40% of the
feed, equivalent to a Ca/S molar feed ratio of 0.4. However, the losses
were much higher at higher superficial velocities, reaching 110% of the
dolomite feed rate at a velocity of 3 m/s (9-8 ft/sec), equivalent to a Ca/S
feed rate of 0.8. This would indicate that the maximum superficial velocity
possible with dolomite sorbent may be around 2.5 m/s (8 ft/sec).
Particulate concentrations in the flue gas leaving the second stage
cyclone normally ranged from 0.9 to 4.8 g/m3 (0.4 to 2.1 gr/SCF) with a mass
median particle size of about 7 microns. The particulate captured in the
second stage cyclone had a mass median particle size of about 20 microns.
Particulate composition data were also measured. The solid material
removed from the combustor normally contained less than 1% unburned carbon,
2 to 25% coal ash and 75 to 95% used sorbent. The material captured by the
second stage cyclone normally contained 3 to 20% unburned carbon, 40 to 80%
coal ash and 15 to 60% used sorbent. These figures varied somewhat with the
sorbent type. For example, the weight fraction of used sorbent in the second
cyclone capture increased when dolomite was used, reflecting the higher
dolomite elutriation loss rates. The fine particulates passing through the
second stage cyclone contained 2 to 7% unburned carbon, 65 to 80% coal ash
and 10 to 30% used sorbent.
Carbon combustion efficiency was found to increase with combustor tem-
perature, reaching over 99% at a temperature of 940°C (1720°F). However, at
lower temperatures, the combustion efficiency was found to vary considerably
between 95 and 98%. No reason for the variation at the lower temperatures
was found. Essentially all of the combustible carbon loss was due to carbon
particulates removed in the second stage cyclone.
Heat transfer coefficients were found to vary from 250 to 420 W/m2R
(45 to 75 BTU/hr ft2 °F), increasing with temperature and elevation in the
bed and decreasing with sorbent particle size.
BATCH UNIT COMBUSTION STUDIES
The smaller and older of the two pressurized FBC experimental units is
the batch unit, so named because only coal is fed continuously to the com-
bustor. The sorbent is added batch-wise to the combustor before a run and
is removed from the combustor and analyzed after a run had been completed.
The combustor consists of a vessel refractory lined to an inside diameter"of
11.4 cm (4.5 in) with an overall height of 4.9 m (16 ft). Three vertical
cooling coils are mounted in the combustor to remove the heat of combustion.
Flue gas leaving the combustor passes through two cyclones and a final filter
in series to remove particulate matter. Particulates captured in each removal
4
-------
device are removed after the completion of a run. Since the sorbent is not
fed continuously to the combustor, the S02 emissions in the flue gas con-
tinually increase as the sorbent is sulfated. The SC>2 retention measured at
the end of a run, together with the degree of calcium sulfation measured in
the sorbent removed from the bed at the end of a run are used to calculate
an equivalent Ca/S molar feed ratio. This permits the comparison of the
desulfurization results measured in the batch unit and in continuous units
such as the miniplant.
S02 emissions were measured using an Eastern bituminous Pittsburgh seam
coal (Arkwright), containing 2.6% sulfur, an Illinois No. 6 bituminous coal
containing 4% sulfur, and a Western subbituminous coal containing 0.7% sul-
fur. Grove No. 1359 limestone and an Ohio dolomite (Tymochtee) were the
sorbents. The SO- retention data obtained with the Eastern coal and dolomite
sorbent agreed fairly well with data obtained in the miniplant after cor-
recting for differences in the gas phase residence times, and considering
the inherent differences between a batch and continuous system. However,
limestone desulfurization results measured in the batch unit exhibited a fair
degree of data scatter and were difficult to interpret. For example, data
obtained with the Eastern coal showed no effect of temperature, contrary to
the results obtained in the miniplant. Runs with Illinois coal showed a tem-
perature effect, but SC>2 retention levels were much lower than expected.
Runs made with the low sulfur Western coal using limestone sorbent gave
very low SC>2 emissions. This indicated the possibility of significant SC>2
retention by the high calcium content of the coal ash. Runs were then made
with inert bed material in place of the limestone and SO- retention levels
on the flyash of about 50% were measured.
NOX emissions were also measured in the batch unit. The emissions
increased with increasing excess air. In the range of 15 to 20% excess air,
the emissions averaged 0.09 to 0.13 g/MJ, (0.2 to 0.3 Ib/M BTU), slightly
higher than the level measured in the miniplant. However, at higher excess
air levels, the NOX emissions were significantly higher than those measured
in the miniplant. Runs made with the Western coal at excess air levels of
over 150%, indicated NOX emissions could be as high as 0.4 g/MJ (0.9 Ib/M BTU),
Further studies with Western coal will be made in a continuous unit to deter-
mine if the NOX emissions are higher than those obtained with Eastern and
Illinois coals.
CO emissions measured in the batch unit averaged about 180 ppm at higher
temperatures, within the range measured in the miniplant.
Particulate emissions from the combustor and the cyclones were measured
by weighing the particulate matter captured in the cyclones and filter.
Emissions from the combustor averaged 16 g/m3 (7 gr/SCF), 2 g/m3 (0,9 gr/SCF)
from the first cyclone and 1 g/m3 (0.5 gr/SCF) from the second cyclone.
Carbon combustion efficiencies increased with temperature and excess
air, exceeding 98% at excess air levels of 60% or more and temperatures
generally in excess of 900°C (1650°F). At 20% excess air, the combustion
efficiency was about 93% at 850 to 870°C (1560-1600°F) and 97% at tempera-
tures above 900°C (1650°F). Combustion efficiencies were generally lox^er
than those measured in the miniplant, probably due to the lack of recycle
of first stage cyclone particulates in the batch unit.
-------
MINIPLANT REGENERATOR SHAKEDOWN
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 CaS04
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 shakedown of the regenerator proceeded in two phases. In the first,
the regenerator was operated in a batch fashion, uncoupled from the combustor,
to check out the system and develop necessary equipment and procedures. In
the second phase, provisions for the continuous transfer of solids between
the combustor and regenerator were developed and'the regenerator, coupled to
the combustor, was operated continuously for a period of 24 hrs.
in the batch ope-raiding inoue, procedures were developed vvliiLn gave good
control of the bed temperature, good temperature distribution and prevented
the agglomeration of the bed by temperature excursions. This was done pri-
marily by proper adjustment and balance of the fuel and air flows into the
system and by the design of a fluidizing grid which promoted solids mixing.
S02 levels measured in the regenerator off gas during the batch tests
were 3% or lower. These concentrations were only about 50% of the concentra-
tions calculated assuming equilibrium was established between the solids and
the gas. However, the S02 concentrations measured in the batch runs agreed
closely with concentrations obtained experimentally by another investigator
(10) under equilibrium conditions indicating the calculated concentrations
may not be correct. Conversion of the sulfated sorbents was over 90% in
the batch tests and selectivity to CaO was over 99%.
The second shakedown phase began by developing a system to transfer
solids continuously between the combustor and regenerator vessels. The
system consisted in setting the pressure in the regenerator very slightly
higher than the combustor and using the pressure differential and a pulsing
flow of nitrogen to move the solids in a controlled fashion between the
vessels. A transfer line lock hopper was used to move the solids from the
combustor to regenerator. Two water cooled knife valves were designed and
built for this service. The valves were opened and closed by a timing
mechanism in such a way as to permit solids to flow by gravity from the
combustor into the transfer line. The transfer line was then isolated,
increased slightly in pressure and then drained into the regenerator. This
system performed satisfactorily during checkout and ran without failure
during the 24 hr, shakedown run. In this run, limestone was added to the
combustor in an amount required to make up for entrainment losses. S02 emis-
sions from the combustor were under 400 ppm and S02 in the regenerator off
gas was between 0.7 and 1.0%. This S02 concentration was limited not by
chemical equilibrium or reaction rates, but rather by the volume of gas
resulting from the fuel which had to be burned to maintain the regenerator
at reaction temperatures. It is not indicative of the S02 level which can
-------
be reached in a larger system with lower heat losses. As the limestone was
cycled between the combustor and regenerator, attrition and entrainment
losses decreased and no limestone was fed to the combustor for the last 10
hrs. of the test.
CONTINUING STUDIES
Additional combustion studies will be carried out in the miniplant in
which an Illinois No. 6 coal containing approximately 4% sulfur will be
burned. Precalcined limestone sorbent will also be used in another series
of tests to determine its activity at low "turndown" temperatures. The
batch combustion unit will be converted to a continuous unit and used to
support the miniplant program.
A particulate control program will be started. A granular bed filter
will be installed on the miniplant flue gas stream and shakedown will begin.
The purpose of the program will be to determine if this type particulate con-
trol device can satisfy environmental and gas turbine particulate require-
ments. It will also be used in a long term materials testing program spon-
sored by the Energy Research and Development Administration (ERDA).
A program aimed at making a comprehensive analysis of all potentially
harmful emissions from a pressurized fluidized bed combustion system will
begin. This will be done as part of the overall EPA FBC environmental
assessment program.
Regeneration studies will continue. The first goal will be to conduct
a 100 hr run demonstrating the feasibility of the continuous FBC combustion
and sorbent regeneration concept. This will be followed by an optimization
study aimed at measuring SC>2 content in the regenerator off gas and regener-
ated sorbent activity maintenance.
-------
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 tem-
peratures), 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 CaS04 as shown in
reaction (1) .
CaO 4- S02 + 1/202 -> CaSO^ (1)
Fresh limestone or dolomite sorbent feed rates to the boiler can be
reduced by regeneration of the sulfated sorbent to CaO and recycle of the
regenerated sorbent back to the combustor. One regeneration system, studied
by Exxon Research and Engineering Company in the past, is the so-called one
step regeneration process in which sulfated sorbent is reduced to CaO in a
separate vessel at a temperature of about 1100°C according to equation (2) .
S02 in the regenerator off gas is at a 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) (2) . Figure II-2 is a photograph of the batch units. Those reports
described regeneration and combustion studies carried on in the batch unit.
The subsequent coal combustion studies carried out in the batch unit under
Contract 68-02-1451 are described in this report. The program was aimed
at the development of equipment and operating techniques, the study of the
effect of process conditions on S02, NOX and CO emissions, the measurement
-------
GAS TURBINE
STEAM
TURBINE
CONDENSER
SEPARA-
TOR
COAL AND
MAKEUP SORBENT
AIR
COMPRESSOR
SOLIDS
TRANSFER]
SYSTEM
TO SULFUR
RECOVERY
DISCARD
BOILER
'UEL
REGENERATOR
FIGURE ll-l
PRESSURIZED FLUIDIZED BED COAL COMBUSTION SYSTEM
-------
FIGURE 11-2
BATCH FLUIDIZED BED COAL COMBUSTION7 UNIT
10
-------
of combustion efficiency, particulate emissions and the measurement of heat
transfer coefficients between the fluidized bed and steam tubes. Various
coals and sorbents were also tested.
The larger unit, called the miniplant, was designed under EPA Contract
CPA 70-19 and built under Contract 68-02-0617. Figure II-3 shows a photo-
graph of the miniplant. The shakedown and operation of the unit was funded
under Contract 68-02-1312. Previous reports (3)(2) described the design,
shakedown and initial 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 S02> NOX, CO and particulate emissions, on
combustion efficiency, sorbent attrition and heat transfer coefficient was
measured. Various sorbents were tested.
This report also describes the regeneration section of the miniplant
and the initial results of the regenerator shakedown activity.
The period of performance discussed in this report is July 1, 1975 to
July 30, 1976.
11
-------
FIGURE II-3
EXXON FLUIDIZED BED COMBUSTION MINIPLANT
12
-------
SECTION III
COMBUSTION STUDIES
Combustion studies have been carried out in two pressurized fluidized
bed units. The smaller and older of the two units is the batch unit. This
unit is capable of continuous coal addition, but does not have provisions
for continuous sorbent addition or continuous removal of sulfated sorbent
from the fluidized bed. The miniplant is the larger of the two units and
has provisions for continuous addition of coal, and sorbent and continuous
withdrawal of sulfated sorbent. This section of the report describes
combustion studies carried out in both experimental units.
MINIPLANT COMBUSTOR
The miniplant is shown schematically in Figure III-l. As of July 1976,
the combustor has been operated for a total of approximately 1100 hours in a
series of individual runs of 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 regenerator section
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, and 7) combustor safety and alarm system. A detailed descrip-
tion of each of these systems can be found in an earlier report and only a
brief discussion will be included here (2). • '
Figure III-2 shows a schematic of the miniplant coal and sorbent feeding
system. Solids are fed under atmospheric pressure from storage bins through
a feeder and blender into a feed vessel. The coal/sorbent mixture is held
in this vessel until refill of the injector vessel becomes necessary. The
injector vessel is maintained at a pressure slightly higher than the com-
bustor. Solids are fed from the bottom of the injector vessel through an
orifice into the transport line. Transport air is then used to inject the
solids into the combustor through an air-cooled nozzle. The solids feed
rate is controlled by controlling the pressure difference between the
injector vessel and combustor. The higher the pressure difference, the
greater the solids feed rate,'as long as the transport air rate is held con-
stant. The solids feed rate is adjusted by a cascade control system which
maintains a constant temperature in the combustor at a point close to the
solids entry port. This control system continually readjusts the pressure
differential between the injector vessel and the combustor to vary the coal
injection rate in such a way as to maintain constant temperature in the
combustor.
13
-------
ORIFICE
p
COOLING
WATER
CITY
WATER
AUXILIARY
AIR
COMPRESSOR
NATURAL GAS
COMPRESSOR
COAL
&
LIMESTONE
FEED
SUPPLY
LIQUID FUEL STORAGE
MAIN AIR
COMPRESSOR
(1400SCFM@
150PSIG)
FIGURE Ill-l
EXXON FLUIDIZED BED COMBUSTION
IPLANT
-------
Limestone Bin
Coal Bin
Blender
D
Feed
Vessel
FIGURE 111-2
COAL & LIMESTONE FEED SYSTEM
Vent
High Pressure Air
Controllers
A P Signal -o0e-»-
AP Temp
I ^-
Injector
Vessel
•24'
Load
Cells
A
TC
i/in/u
Combustor
1/2 S.S. Pipe
-------
The combustor consists of a 61 cm I.D. steel shell refractory lined to
an actual internal diameter of 31.8 cm. The 9.75 m high unit is designed in
flanged sections and contains various ports 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.
Combustion air to the unit is provided by a main air compressor having a
capacity 40 stnd. m3/min at 1030 kPa gauge (1400 scfm at 150 psig). The bot-
tom plenum section houses the natural gas burner used for initial pre-heating
of the bed during unit start-up. Fuel to the burner is provided by a natural
gas compressor with a capacity of 0.57 stnd. m3/min at 1379 kPa gauge (20
scfm at 200 psig). Once the fluidized bed temperature reaches 430°C, a
liquid fuel system is used to heat the bed to the coal ignition temperature.
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 m2 and consists of vertically-oriented loops constructed of 1.3 cm
Schedule 40 316 stainless steel pipe. A closed-loop arrangement is used in
which a feedwater reservoir provides the supply of demineralized cooling
water. A high pressure pump is used to pump the water through the coils.
The flow rate and exit temperature from each coil can be separately con-
trolled and monitored. The steam-water mixture that exits from each coil is
condensed, returned to .the feedwater reservoir, and recirculated.
The combustion gases that exit the combustor go through 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 combus-
tion efficiency. Particles trapped by the secondary cyclone are dropped
into a lockhoppar and disposed of on a batch basis.
The technique used to control combustor pressure consists of allowing
the system pressure drop to occur across an appropriately sized silicon
carbide nozzle located in the flue gas exit line. Back pressure is con-
trolled by regulating the flow of a secondary air stream to the nozzle
inlet. After passing through the nozzle, the gases are cooled and then
enter a scrubber for final cleanup before being discharged into the
atmosphere.
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
gas composition is essentially unaltered from that of the original flue gas.
A schematic is shown in Figure III-4. A flue gas particulate sampling sys-
tem was also installed and is shown schematically in Figure III-5.
Solids rejection from the combustor is required to maintain a steady
bed height when a mixture of coal and sorbent is fed. Solids are rejected
through a port located 230 cm above the fluidizing grid. Solids flow by
gravity through a refractory lined pipe into a "pulse pot" from where they
are penumatically transported by controlled nitrogen pulses to a pres-
surized lockhopper.
16
-------
FIGURE III-3
COMBUSTOR VESSEL
SHELL- Z4'H. 375 WALL STEEL flf£(f
•5MCLL. Fi-AMCiES • Z&" tf° * STEEL, ff.
-------
00
:lue Gas
Line
1
V
Vt
i
Continuous p
jnt
7
Analyzers ^*
Heated Sample Conditioning Box —7
Wet Chemistry
Analysis
t * Y
b 1 ltd'
_J— 1 f— tX]
Cooler
^•M
T
1
}
Filter
PI
Y
TC>-
F
ilte
#1
PI PI
? -T1?
„ Filter TC
r #2
••^^H
i
^2 Purge
t i
i i
1 r^^^^"^
Permapure
Drier
\ •
[Filter
/
R f
,X T^ I ^ Hydrocarbon
Analyzers
FIGURE 111-4
FLUE GAS SAMPLING SYSTEM
-------
FLUE GAS FROM SECONDARY CYCLONE
HIGH TEMPERATURE
VALVE
-Xh
PRESSURE
GAUGE
TC
XL
PRESSURE
GAUGE
TC
-txj-J
ROTAMETER
HEAT COLLECTION STEAM FLOW CONTROL VALVE
EXCHANGER FILTER HEAT EXCHANGER
PURGE
NITROGEN
FIGURE 111-5
MINIPLANT PARTICULATE SAMPLING SYSTEM
-------
Data characterizing total system operation are recorded on three
multipoint recorders. In addition, at one minute intervals, the same data
are recorded by a data logger system consisting of a Digitrend 210 data
logger with printer and a Kennedy 1701 magnetic tape recorder. The magnetic
tape, containing about 3600 items of data per hour of run time, is fed to
a computer which converts the data logger output signals to flow rates,
pressures, etc. The data are then averaged and standard deviations cal-
culated over pre-selected time intervals.
The unit is equipped with a process alarm system which 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 shut-
down of the system.
Materials—
Coal—Coal used in the miniplant variables study was a high volatile
bituminous coal obtained from the Consolidation Coal Company's "Champion" pre-
paration plant in Pennsylvania. The coal is a Pittsburgh No. 8 seam coal and
its analysis is shown in Table III-l. Coal samples were periodically sent to
Exxon's Baytown, Texas, coal research facilities or to the Conoco Research
and Development Company to assure accurate analysis. Grinding and sizing was
done by the Penn-Rillton Company. Essentially all of the coal was less than
2380 urn (No. 8 U.S. Mesh). Fines smaller than 40 mesh were partly removed.
Two typical distributions are shown in Figure III-6.
Sorbents—Grove limestone (No. 1359) and Pfizer dolomite (No. 1337) were
the primary sorbents used in the miniplant variables study. One run using
Tymochtee dolomite was also made. The composition of these stone is given
in Table III-2. Most of the runs were made with the stone screened to give
a distribution with a minimum of 90% between 2380 ym (No. 8 U.S. Mesh) and
841 i-im (No. 20 U.S. Mesh). A typical size distribution of the limestone is
shown in Figure 111-7. Runs to evaluate the effect of sorbent particle size
on desulfurization were also made. For these runs particle sizes of 8 x 14
and 14 x 25 mesh were used.
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, calibration of flue gas analyzers, activation of process mon-
itoring and control systems, and the turning on of all cooling water systems.
All runs were begun with an initial bed of sorbent in the combustor. This
consisted of either a fresh charge of uncalcined limestone or the bed from
the previous run.
The first operation of startup involves heating the 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
stnd. m^/min (350 scfm) or about half that used at normal operating condi-
tions 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
20
-------
TABLE III-l. MINIPLANT COAL ANALYSES
Component Weight Percent
Coal
Source of Analyses
Run No.
Moisture
Ash
Total Carbon
Hydrogen
Sulfur
Nitrogen
Oxygen (by difference)
Chlorine
Higher Heating Value (BTU/lb)
Champion
Exxon
19.4-19.9
2.21
8.85
78.14
5.08
2.21
1.61
4.12
0.08
13,649
Champion
Exxon
20.1-27.3
2.52
9.64
75.31
5.39
1.95
1.57
6.15
0.07
13,513
Champion
Conoco R&D
20.1-27.3
2.37
8.75
72.69
5.19
1.95-1.98
1.52
N.D. to
N.D.
Champion
Exxon
27.4-27.12
Dry Basis
9.52
75.26
5.13
2.08
1.62
6.39
0.06
13,577
Champion
Conoco R&D
27.4-27.12
2.95
7.91
73.34
5.17
2.17-2.21
1.50
N.D.
N.D.
(1) Not Determined
-------
TABLE III-l (continued). MINIPLANT COAL ANALYSES
Component Weight Percent
Coal
Source of Analyses
Run No.
Moisture
Ash
Total Carbon
Hydrogen
Sulfur
Nitrogen
Oxygen (by difference)
Chlorine
Higher Heating Value (BTU/lb)
Champion
Exxon
27.13-27.21,28
Dry Basis
8.27
75.83
5.11
2.14
1.60
7.06
0.06
13,711
Champion
Conoco R&D
27.13-27.21,28
2.60
8.94
72.66
5.10
1.90-2.01
1.49
N.D.
N.D.
Champion
Exxon
29-35
Dry Basis
9.12
75.64
5.09
1.78
1.60
6.87
0.06
13,628
Champion
Conoco R&D
29-35
2.73
9.35
72.89
5.16
2.10-2.19
1.46
N.D.
N.D.
Champion
Exxon
36.1
Dry Basis
8.26
72.43
5.16
2.02
1.40
10.73
0.09
13,735
-------
K3
FIGURE 111-6
COAL PARTICLE SIZE DISTRIBUTION
LU
M
CO
LU
_l
C_)
P
Q_
^
1-
CE
LU
z
LZ
^0
0^
1—
n:
LU
->
>
J. UU
90
80
70
60
50
40
30
20
10
0
—
_
—
/
/
/
~ / (
- //
- #
^r* , i
0 200400 600
ill iii
100 5040 30
^^^~ 9 •
^f ^f^
/^ •
/Y
//
o •
o •'
/
{
1 1 1 1 1 1 1 1 1 1
80010001200140018002000 2400 2800 3200 (MICRC
i I i i i i i i i ii
20 18 16 14 12 10 8 7 1/8" 6 (MESH
PARTICLE SIZE
-------
TABLE III-2. PROPERTIES OF SORBENTS USED
IN MINIPLANT VARIABLES STUDY
Designation
Chemical Analyis, wt %
Quarry Source
Grove Lime Co.
Stone Type CaO MgO SK>2 A1203 Fe203
Limestone 97.0 1.2 1.1 0.3 0.2
K5
-P-
1337
Tymochtee
Chas. Pfizer Co.
(Gibsonburg, OH)
C. F. Duff and Sons
(Huntsville, OH)
Dolomite
Dolomite
54.0 44.0 0.9 0.2 0.3
53.8 38.7 5.3 0.9 1.2
-------
FIGURE 111-7
LIMESTONE NO. 1359 PARTICLE SIZE DISTRIBUTION
N
CO
LU
_J
O
H-
Qi
<
Q-
CO
CO
UJ
100
90
80
70
60
50
40
o 30
20
10
0
T
T
J_
J_
_L
_L
_L
J_
0 200 400 600 800 10001200140016001800200022002400
PARTICLE SIZE (MICRONS)
-------
simultaneously feeding 0.57 stnd. m^/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 temperature 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 stnd. m^/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 s.et 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 opera-
ting values. Flow of water to each cooling coil is adjusted to maintain
steam/water exiting temperatures of 138-150°C. Once the desired bed tem-
perature has been reached, it is held constant by the automatic coal feed
rate control system.
Miniplant Combustor Performance
Length of Operation and Conditions Tested—
As of June 1976, the miniplant has accumulated over 1100 hours of coal
combustion time. During the shakedown phase a total operating time of 500
hours was accumulated and 37 runs conducted. Many of the initial runs were
limited to 10 hours or less duration, but as the systems were improved the
run durations increased. Included among the longer runs were four that were
24 hours in duration, one 50 hours in duration, and a 100 hour continuous run
which concluded shakedown.
During the operations phase a total operating time of approximately 600
hours was accumulated covering 66 sets of operating conditions. Included
among the longer runs were four of over 24 hours duration and a 240 hour
demonstration run. During the demonstration run, coal combustion was con-
tinuous except for a total of 4 hours, during which liquid fuel was burned to
maintain the bed temperature. During the last 7 days of the run, liquid fuel
was used for only 15 minutes. The on line factor for coal combustion was
98.5%.
The primary objective of the operations phase of the program was to
evaluate the effect of key operating parameters on the emission levels, com-
bustion efficiency, and heat transfer rates. Table III-3 presents a brief
summary of the miniplant test program, indicating the objective of each
series of runs as well as the ranges of the operating variables. As indica-
ted, the miniplant combustor proved to be very flexible and was able to run
over a wide range of operating conditions.
26
-------
TABLE III-3. SUMMARY OF MINIPLANT TEST PROGRAM
Run No.
19.1-19.9
20.1-20.2
21
26
22
Objective
Study the effect of temperature
and the Ca/S molar feed ratio
on the emission levels using
Grove limestone as the sorbent
Measure S0~ level after sorbent
feed stopped
23
25
Evaluate dolomites for use in
the variables study
27.1-27.21
28.1-28.5
Study the effect of temperature
and the Ca/S molar feed ratio
on the emission levels using
Pfizer dolomite as the sorbent
Verify the accuracy of the S02
measurements by running with
an inert alumina bed
Operating Variables
P = 930 kPa
Vel. = 2.0 m/s
Excess air = 15%
T = 820-954°C
Ca/S = 1.45-3.70
Expanded Bed Height = ~2-4m
P = 930 kPa
Vel. =1.9 m/s
Excess air = 28%
T = 870°C
Ca/S = 0
Expanded Bed Height = 2-4m
P = 930 kPa
Vel. = 1.9 m/s
Excess air = 8-12%
T = 870-885°C
Ca/S = 1.30
Expanded Bed Height = /"2-4m
P = 930 kPa
Vel. = 1.1-2.3 m/s
Excess air = 8-23%
T = 829-931°C
Ca/S = 0.0-2.5
Expanded Bed Height = 3-7m
P = 930 kPa
Vel. = 1.4-2.7 m/s
Excess air = 5-40%
T = 840-930°C
Ca/S = 0
Expanded Bed Height = 2m
-------
TABLE III-3 (continued). SUMMARY OF MINIPLANT TEST PROGRAM
t-o
00
Run No.
29-30.4
Objective
31
32.1-32.3
33
34
35
36.1-36.2
37
38.1-38.6
39.1-39.4
Better define the correlation
between the S02 emission level
and the Ca/S molar feed ratio
and temperature for uncalcined
limestone
Study the effect of operating
pressure using limestone and
dolomite
Study the effect of secondary
variables (excess air, bed
depth, superficial velocity,
and particle size) and
evaluate the performance at
turndown conditions
Operating Variables
P = 920 kPa
Vel. = 1.9-2.5 m/s
Excess air = 14-19%
T = 835-929°C
Ca/S =3.7
Expanded Bed Height = 2-4m
P = 520-600 kPa
Vel. = 1.5-2.2 m/s
Excess air = 13-23
T = 836-950°C
Ca/S = 0.8-2.5
Expanded Bed Height = 2-3m
P = 902-932 kPa
Vel. = 1.3-3.0 m/s
Excess air = 18-112%
T = 674-938°C
Ca/S = 0.8-2.5
Expanded Bed Height = 2-4m
-------
Control of Operating Conditions—
Good control of each of the operating variables was demonstrated for
sustained periods. Presented below are the averages and standard deviations
calculated for the combustor operating conditions over a 1 hour interval
during steady state conditions for a typical combustor run:
Variable Average Standard Deviation
Bed temperature (°C) 901 3
Coal feed rate (kg/hr) 119 6
Combustion pressure (kPa) 920 1
Gas superficial velocity (m/s) 3.1 0.02
Bed height (m) 6.7 1.2
Bed temperatures were well controlled using the control system described
earlier. Very good temperature profiles were established with temperature
variations across the bed generally being approximately 30°C. The tempera-
tures of the flue gas exiting the cyclones were typically close to the tem-
peratures at the top of the expanded bed.
The combustor pressure was well controlled using a fixed converging
nozzle with supplementary air addition. No pressure control problems were
noted at pressures between 520 and 930 kPa. Steady gas velocities were
maintained by achieving good control of the air flow rate, pressure, and
bed temperature.
The combustor bed level was well controlled by continuously removing
solids from the combustor through a pulse pot solids reject system as
described on page . The pulsing rate was adjusted to maintain a nearly con-
stant bed level as indicated by a constant pressure drop across the bed.
Cooling Coils—
The design of the miniplant cooling coils has undergone a number of
modifications in an attempt to overcome problems of erosion and fatigue and
to promote better solids mixing. A detailed description of the coils used
through Run 19.3 has already been presented in an earlier report (2).
Since then, new coils were constructed from 1.3 cm schedule 40 316
stainless steel pipe for protection against fatigue and, in the event of bed
agglomeration, against deformation. Support rods connecting the inlet and
outlet piping to the adjacent coils were used to hold the coils rigid.
Pipe bends having a 35 mm radius and 170 mm long tangent were butt welded
together in the construction of the coils. The avoidance of socket welded
U bends as used in prior coils allowed fewer welds. Also, the sockets of
the U bends were susceptible to damage from erosion and/or corrosion. The
surface area of each coil was 0.55 m2, approximately the same area as used
in the previous coils.
These changes significantly lengthened the lifetime of the coils, but
the coils were still susceptible to damage. These coils failed after a
total operating time of 140 hours. The damage was localized at several of
upper U bends located towards the interior of the column. These U bends
had holes and eroded surfaces on the under side of the U bends. The
29
-------
photograph in Figure III-8 is a typical example. It is believed that the
erosion occurred because jets of gas bypassed the bed by forming high
velocity annular gas streams along the length of the vertical tubes. Bed
solids accelerated by the gas stream impinged on the underside of the upper
U bends. Velocities of 15-30 m/sec or more must have been attained to
explain the erosion seen.
Prior to the 240 hour demonstration run, the coils were baffled with
bars on the underside of the more susceptible U bends. After the run it
was discovered that the cooling coils had again sustained damage due to
erosion. The most extensive damage was exhibited by coil IB. Coil 2A also
showed signs of erosion, but not as severe as coil IB. Damage was not
detected on coil 1A and only a few eroded surfaces were found on coil 2B.
Also, many of the welds connecting the vertical tubes located near the
center of the column were eroded to a smooth surface. The erosion was
again attributed to gas bypassing.
In the next effort to protect the coils by -diverting the gas jets, baf-
fles were installed on the coils by welding 2.3 X 4.2 cm rings on each tube,
one each at the upper and lower U bends and two at intermediate positions.
Half rings were primarily used, staggered at 90° intervals about the vertical
axis. The rings on adjacent tubes were staggered to maximize the cross
sectional area vailable to the upward moving solids and gases at any position
along the tube bank. The rings proved to be very effective and no signs of
erosion were noted after more than 200 hours of operation. Figure III-9
shows a photograph of a typical coil after 15 hours of use. It was sub-
sequently found that it was only necessary to use half rings at the upper
U bend to protect the coils. Coils built recently have used only the upper
half rings.
Temperature Distribution—
As discussed in an earlier report, a dramatic improvement in the bed
temperature profile was observed when vertically oriented cooling coils
were used (2). Also noted was a substantial increase in the freeboard
temperature when beds depths were increased to immerse all coils and prevent
cooling of the flue gas above the expanded beds. Flue gas temperatures
comparable to the combustion zone temperatures were thus achieved.
When baffles were added to the cooling coils to prevent erosion, the
possibility that the baffles would affect the quality of solids mixing and
the heat removal capacity of the coils was considered. Typical temperature
profiles for baffled and unbaffled coils are shown in Figure 111-10. The tem-
perature drop across the bed was slightly greater with the baffled coils
(35°C vs 27°C), but the difference is of negligible significance. The aver-
age heat transfer coefficient for the baffled coils compared very closely
with the average coefficient for the unbaffled coils. The baffles which act
as fins added about 15% to the surface area of the coils. It is likely,
however, that the stagnant areas created by the baffles reduce the bed to
tube heat transfer coefficient. The competing effects may cancel, resulting
in little or no change in the heat removal capacity.
Particulate Sampling System—
The isokinetic particulate sampling system presently being used was
shown schematically in an earlier section of the report. To assure reliable,
and accurate particulate emission measurements, the sampling system was
30
-------
FIGURE HI-8
EROSION DAMAGE AT THE UPPER U BEND OF COIL IB
31
-------
FIGURE III-9
BAFFLED COIL AFTER 15 HRS OF OPERATION
-------
FIGURE 111-10
MINIPLAIMT COMBUSTOR TEMPERATURE PROFILE
900
o
0
LU
§ 800
i-
oi
UJ
Q_
UJ
1-
700
0
1 1 1 l 1 1 1 1 1
-------
continually modified and improved. Poor performance of earlier systems were
caused primarily by:
. • plugging problems
• temperature control problems
• moisture condensation problems
A comparison of the performance of earlier sampling systems with the
present one shows a substantial reduction in the deviation from isokinetic
sampling using the present system. Other modification are planned to further
reduce deviations from isokinetic sampling and to simplify operation.
The present sampling system also allows the use of an impingement type
device to jneasure the flyash particle size distribution. The major advantage
in using this type of device is that it gives a direct measurement of the
size distribution as compared to systems which rely on collection-redisper-
sion techniques.
EXPERIMENTAL RESULTS
The primary objective of the experimental program was to measure and
compare the performance of a limestone and dolomite sorbent for S02 control
over a range of operating conditions. Variables included the sorbent feed
rate, pressure, bed temperature, superficial velocity, bed depth, excess air
level and sorbent feed particle size. The bulk of the experimental work
focused on the effect of the primary variables, the sorbent feed rate and
bed temperature. In addition to SC>2 control, a number of other measures of
process performance were determined as a function of operating conditions.
This included NOX, CO, 803 and particulate emissions, attrition and elutria-
tion of the sorbents, carbon combustion efficiency and heat transfer coef-
ficients between the fluidized bed and the cooling coils.
The effect of the primary variables on the performance of Grove No. 1359
limestone was investigated in a series of 15 runs encompassing 26 test condi-
tions. The Ca/S molar feed rate was varied from 1.5 to 3.7 and the tempera-
ture was varied from 840 to 950°C. The runs lasted from 6 to 16 hrs., except
for one 36 hour run. The effect of the primary variables on the performance
of Pfizer No. 1337 dolomite was investigated during Run 27, which was a 240
hour continuous run. Twenty-one test conditions were completed during Run
27. The Ca/S molar feed ratio was varied from 0.6 to 2.2, and the tempera-
ture was varied from 840 to 950°C.
Runs examining secondary variable effects were also run. Runs 31 and
32.1-32.3 were made at a reduced pressure of 500 to 600 kPa, compared to 900
to 940 kPa used in all other runs. Both limestone and dolomite were used.
The effect of sorbent particle size on S02 emissions was measured in runs
38.1, 38.2, 39.3 and 39.4. The standard 8 to 25 mesh sorbents were screened
into two size fractions, 8 to 14 mesh and 14 to 25 mesh. This gave a 2/1
ratio of mass average particle sizes in the two fractions. The effect on
S02 emissions of varying the gas residence time by changing the gas velocity
and bed height were examined in runs 33 to 37. Velocities up to the design
34
-------
value of 3 m/sec were used. Excess air levels as great as 100% were attained
in these tests because of a reduction in the heat transfer surface area.
Pfizer dolomite was the only sorbent used in runs 33 to 37.
The operability of the unit at "turndown" temperatures of 690 and 760°C
and the effectiveness of limestone and dolomite as sorbents at these tem-
peratures was investigated in runs 38.3 to 38.6, 39.3 and 39.4.
Run 28 was made with an inert bed of alumina to establish the accuracy
of the monitored S02 emission levels by comparison with that value calcul-
ated from the sulfur content of the coal and the excess air level.
The experimental program is summarized in Table III-4. Comprehensive
information on the individual runs is available in Appendix B.
Calculation of Ca/S Ratio by
Sulfur and Calcium Balance
A sulfur and calcium balance around a combu'stor gives the expression:
n /c /vr i /« i <* Sulfur Retention (%)
Ca/S (Mole/Mole) = 7—^—: TT .., . — ,„•>
Calcium Utilization (%)
This is an exact expression when all input and output streams are included.
At steady state (feed rates constant, bed height and composition con-
stant, output rates and compositions constants, etc.) the retention and
utilization will be a function of the Ca/S ratio (and other variables such
as residence time and temperature). The time required to displace all the
solids in the bed is long, possibly on the order of 20 to 40 hours. This
is based on the fact that the solids residence time is usually in the range
of five to ten hours and complete replacement of the solids requires four
times the residence time. Therefore, if operating conditions in the com-
bustor are changed significantly, a very long line-out period would be
required to assure that the solids composition in the combustor has reached
the new steady state value. In practice, the line-out time is usually on
the order of two to four hours and is based on the time required for the SC>2
content in the flue gas to reach a new relatively constant value. Therefore,
the solid composition may not always have achieved its new steady state
level when the data gathering portion of the new run begins.
Since it was impractical to wait until the solids has reached steady
state in each run, it was felt that a more significant correlation could be
obtained by calculating a Ca/S ratio using the retention and bed utilization
that existed at the time the data were being collected.
In practice, the use of the calculated Ca/S ratio has given somewhat
better correlations than use of the "set" value of the Ca/S ratio. The "set"
Ca/S ratio is that based on the settings of the limestone and coal feed
screws. This is especially true in cases where the set Ca/S ratio was
changed radically between runs. In other cases (e.g., runs 30.1, 30.3 and
30.4) obvious mechanical problems prevented an accurate value of Ca/S (set)
from being obtained. Generally it has been found that use of the calculated
Ca/S ratio rather than the set value shows a somewhat higher (20%) sorbent
requirement.
35
-------
TABLE III-4. MINIPLANT COMBUSTION PROCESS VARIABLE STUDIES
Super. Settled Excess Feed Sorbent
Pressure Temp. Ca/S Velocity Bed Depth Air Part. Size
Test Series Run No. (kPa) (°C) (Mole/Mole) (m/s) (m) (%) (Mesh)
Primary Variables
Grove Limestone 19.1-22 930 820-940 1.45-3.7 1.7-2.3 .7-2.3 5-28 8 X 25
26-29
30.1-30.4
Pfizer Dolomite 27.1-27.21 930 930-930 .35-2.5 1.1-2.2 2.2-3.9 8-23 8 X 25
Secondary Variables
Grove Limestone 31, 32.1, 520, 600 840-950 2.5 1.6-2.2 1.3-1.9 13-32 8 X 25
39.3, 39.4 930 8 X 14, 14 X 25
Pfizer Dolomite 32.2-38.2 600, 930 840-950 .75-2.5 1.5-3 0.8-2.3 16-96 8 X 25, 8 X 14
14 X 25
Turn Down, Heat Transfer
Grove Limestone 39.1, 39.2 902 670-760 2.5 1.4-1.5 1.2 60-68 8 X 25
Pfizer Dolomite 38.3-38.6 930 690-760 0.75-1.5 1.3-1.85' 1.1-2.6 80-112 8 X 25
Inert Bed
(Alumina) 28.1-28.5 930 840-930 0 1.4-2.65 1 5-40 8 X 25
-------
obtained. Generally it has been found that use of the calculated Ca/S ratio
rather than the set value shows a somewhat higher (20%) sorbent requirement.
As more data become available from the combustor, attempts will continue
to be made to find the "best" means of correlating the retention with the
measured parameters even when the bed composition has not reached a steady
state. In this report, the calculated Ca/S ratio is used unless otherwise
noted. However, the set values have also been included in the run summary
tables.
Control of S02 Emissions with
Pfizer No. 1337 Dolomite
Effect of Ca/S Molar Ratio—
Pfizer dolomite is a very active sorbent. During Run 27, the SC>2 emission
level was reduced to below 100 ppm at a Ca/S ratio of only 2.0, as shown in
Figure III-ll. If the data are replotted as sulfur retention vs. Ca/S ratio^
as in Figure 111-12, it is seen that the retention levels exceeded 95% at a
Ca/S ratio of 2.0. The EPA S02 emission standard of 1.2 Ib SC^/M BTU was met
for the 2% sulfur Champion coal with a Ca/S mole ratio of only 0.85, cor-
responding to a 60% retention level. An examination of the retention data
shows that high calcium utilizations were achieved. At Ca/S ratios less
than 1.0, the utilization levels are in the range of 50% to 90%. In this
region the retention level decreases linearly with the Ca/S mole ratio.
The lowest utilization level was -45% at a Ca/S ratio of 2.0
Effect of Dolomite Utilization on the Rate of Sulfation—
The rate at which calcium reacts with 862 under fluidized bed combustion
conditions may be described by a reaction rate model which is first order in
S02 concentration (4,5).
The sulfation reaction rate constant can be calculated from experimental
sulfur retention levels using the equation:
„. , , , -1, 1 , ,100-Retention,
First order rate constant, k, (s ) = - In ( r—- )
gas
where t , the gas phase residence time is given by
gas
Expanded bed height (m)
gas Superficial gas velocity (m/s)
The rate constant decreases with the sorbent utilization because the
decreased porosity limits access to the active sites. This dependence is
shown in Figure 111-13. Data were included from units other than the mini-'
plant to obtain a greater number of data points over a wider range of utiliza-
tions and to judge the consistency of the data which have been generated to
date. Irrespective of the differences shown in Table III-5 with regard to the
reactor geometry, temperature, pressure, particle size, and dolomite and coal
type, a good correlation was obtained. The fraction of the explained variance
was 88% and the F ratio was 154. The correlation was refined by rejecting 4
of the 47 points which were beyond the 2a limit on the first regression. The
data were also examined for the effect of temperature and pressure but
neither was statistically significant. Two primary conclusions are that the
agreement between data from different units is good and that the effect of
37
-------
FIGURE 111-11
MINIPLANT S02 EMISSIONS FOR DOLOMITE NO. 1337
1400
CO
2
o
CO
LU
CM
O
CO
1200
1000 -
800 _
200
600 ~
400 -
0
RUN 27
PRESSURE - 930 kPa
TEMPERATURE - 840-930°C
RESIDENCE TIME - 2.5-3 S
J_
2 3
Ca/S (MOLE/MOLE)
38
-------
FIGURE 111-12
MINIPLANT S02 RETENTION VS. Ca/S RATIO
FOR DOLOMITE NO. 1337
100
90-
80
70
o:
CM
o
CO
60
50
40
30
20
0
1 2
Ca/S (MOLE/MOLE)
39
-------
TABLE III-5. COMPARISON OF FBC PRESSURIZED UNITS
Laboratory
Argonne
BCURA^
BCURA(8>
Unit Size
(m) Tests
0.15 dia. VAR1/VAR9
EA1/EA9
0.6 x 0.9 1, 2, 3.1,
3.2
1.2 x 0.6 1.1/1.5
2.1/2.4
3.1
Press.
kPa
810
810
600
350
350
510
Temp.
°C
785-900
900
890-950
800
800
800
Gas
Residence
Time (sec)
0.6-1.4
0.7
2
1.9
1.9
1.9
Ca/S
(Mole/Mole)
1-3.2
1.1-2.9
1.4-2.2
0.7-0.9
1.4-1.95
1.5-2.0
Sorbent
T DOL
T DOL
P DOL
UK DOL
P DOL
P DOL
Coal
ARK
ARK
PITTS &
ILL.
WELBECK
PITTS
PITTS
Bed Mass
Avg. Part.
Size (ym)
800
800
800
700
1200-1500
900
3.2
4.1, 4.2
EXXON 0.3 dia. 27.1/27.21
Miniplant 32.2, 32.3
35, 36.1, 36.2
27
930 840-950
1-3
0.5-2.5
P DOL
CHAMP
800-1200
T DOL - Tymochtee dolomite
P DOL - Pfizer No. 1337 dolomite
UK DOL - United Kingdom dolomite
ARK - Arkwright
PITTS - Pittsburgh
ILL - Illinois
CHAMP - Champion
-------
temperature and pressure is minor. Utilizations as great as 70-80% were com-
monly obtained in the miniplant when deep beds were used in conjunction with
low Ca/S ratios. The utilizations in the Argonne unit were lower, mostly
35-50%, primarily because of shallow beds and low residence times. The BCURA
data overlaps the data from the other two units. However, most of the util-
izations range between 45-60%,
Effects of Other Variables—
Temperature— Statistical analysis of emission data from Run 27 where the
temperature ranged from 840-950°C did not reveal a temperature effect on
activity, nor did statistical analysis of the first order rate constant vs.
utilization data given in Figure 111-13. Emission data from "turndown" runs
38.3 to 38.6 did, however, show that activity was slightly lower at the
reduced temperature levels of 690°C and 760°C. The turndown data are compared
to data from Run 27 in Figure 111-14.
Parameters which may affect activity and have a temperature dependence
are the intrinsic rate constant, diffusivity and pellet porosity. The por-
osity depends on the degree of calcination. Depending on the operating con-
ditions, dolomite undergoes either half calcination or full calcination when
fed to the combustor. At a pressure of 930 kPa, the temperature required for
full calcination is ^-900°C. Many tests were made under conditions which were
marginally unfavorable for full calcination. In the case of the turndown
tests, they were very unfavorable. The high activity at turndown temperatures
shows that half calcined dolomite is sufficiently porous to have good activity
and that the reaction between calcium carbonate and sulfur dioxide is rapid.
The small temperature dependence over a fairly broad temperature range is
characteristic of diffusion controlled reactions.
Gas Residence Time—The effect of gas residence time on sulfur retention
was investigated in Runs 35 to 37. The gas velocity and bed height were
varied to obtain residence times between 1 and 3 s. The emissions data are
plotted in Figure 111-15 together with the retention profile obtained from
Run 27 where the gas residence time was 2.5-3 s. As expected, the retention
levels of the low residence time data generally falls below the profile from
Run 27.
A prediction of the residence time effect can be made using the following
approach. A sulfur balance gives:
, S02 Retention (%)
Calcium Utilization (%)
The S02 retention level, R, is derived from first order kinetics to be:
R = 100 (l-e~ktgas)
and the rate constant, k, as a function of utilization can be obtained
directly from Figure 111-13 or can be calculated using the empirical equation:
k = 10.11 - 0.234 (utilization) + 0.0014 (utilization)2
These three equations with five variables have two degrees of freedom. There-
fore, if two of the variables are set, such as residence time and Ca/S ratio,
41
-------
o
LJ
CO
FIGURE 111-13
ACTIVITY VS. CALCIUM UTILIZATION FOR DOLOMITE
7.2
6.4
5.6
4.8
4.0
3.2
o
o
LJ
I-
Di
LJ
a
or
o
/ 2.4
QL
1.6
0.8
0
• EXXON R&E
a ARGOIMNE N.L.
• NATIONAL COAL BOARD
A NATIONAL COAL BOARD
0
, 1 , , 100 - Retention
t I ~ n f
u [ 100
n
20 40 60 80
CALCIUM UTILIZATION (%
100
42
-------
100
FIGURE 111-14
EFFECT OF TEMPERATURE AND SORBENT PARTICLE SIZE
ON S02 RETENTION - DOLOMITE No. 1337
90
80-
70
Temp 690°C
Temp 690°C
Temp 760 °C
Sorbent 8-25 Mesh
60
o
z 50
UJ
CM
o 40
CO
Temp 760°C Sorbent 8-25 Mesh
Temp 900°C Sorbent 8-14 Mesh
Temp 900°C Sorbent 14-25 Mesh
30
20
10
Oj
Data Run 27
Press. 930 kPa
Temp. 840-950°C
Sorbent 8-25 Mesh
Ca/S (MOLE/MOLE)
43
-------
FIGURE 111-15
EFFECT OF GAS RESIDENCE TIME DIM S02 RETENTION
DOLOMITE NO. 1337
100
PARAMETERS ARE
GAS PHASE
RESIDENCE TIME (S)
DATA (RUNS 35-37)
DATA (RUN 27)
PREDICTED
0
0
Ca/S (MOLE/MOLE)
44
-------
the system is defined and the other three variables, rate constant, utiliza-
tion and retention can be calculated. The calculation is a trial and error
procedure. Predicted retention profiles generated for residence times of
0.5, 1 and 2 s are included in Figure 111-15. The profiles do not differ much
in the 1-3 s range. This is especially true at Ca/S ratios below 1.0 where
the profiles converge. The effect does become increasingly important, however,
when the residence time is below 1.0 s and must be considered when very shal-
low beds are employed. A comparison of the emissions data to the predicted
retention levels is given in Table III-6.
TABLE III-6.
Gas Residence Time Measured Retention Predicted Retention
Run No. (s) (%) (%)
32.2 0.8 60 75
32.3 0.8 66 75
35 2.8 62 64
36.1 1.6 66 75
36.2 1.0 76 73
37 1.4 46 35
The measured retention levels from runs 35 to 37, made at 9 atm total
pressure are in good agreement with the predicted retention levels. The data
from Runs 32.2 and 32.3 could be low because of the reduced pressure of 600
kPa.
The dependence on gas residence time of the Ca/S ratio required to meet
the EPA S02 emission standard of 1.2 Ib S02/M BTU is shown in Figure 111-16.
These curves were calculated using the above equations. The effect is not
large, but is evident. As an example, for a 2% sulfur coal, an increase in
gas residence time from 1 s to 3 s reduces the Ca/S requirement from -0.95
to -0.75.
Particle Size Effect— The particle size effect was examined in Runs
38.1 and 38.2. A dolomite feed stream with a size range of 8-14 mesh was
used in one test while the other test used a 14-25 mesh feed. The mass
average size ratio of the two feed streams was ~2. The emission level in the
two tests did not differ significantly as is shown in Figure 111-14. The
final size distribution of the bed material in Runs 38.1 and 38.2 could not
be determined. An operating problem with the bed removal system did not
allow a sample to be taken. The mass average size of the beds could very
well have been more similar than that of the feed streams. Since the
dolomite attrits more readily.
Pressure and Excess Air—No pressure or excess air dependencies were
identified when the data in Figure 111-13 were examined by regression
analysis. The pressure ranged from 350 kPa in the BCURA 1.2 X 0.6 m unit
to 930 kPa in the miniplant. Excess air effects were not evident in the
data from Run 27 either, when those data were examined alone.
The SOo retentions were lower than expected in Runs 32.2 and 32.3 made
at a reduced pressure, 600 kPa. However, insufficient data are available at
low pressure operations to determine if a pressure dependence exists.
45
-------
FIGURE 111-16
EFFECT OF GAS RESIDENCE TIME ON Ca/S RATIO REQUIRED TO MEET
EPA SOo EMISSION STANDARD WITH DOLOMITE
UJ
_i
o
UJ
_i
o
COAL SULFUR
CONTENT
CT3
O
0
0
GAS RESIDENCE TIME (SEC)
-------
Control of S02 Emissions with Grove No. 1359 Limestone
Effect of Ca/S Ratio and Temperature—
502 emissions for runs using Grove limestone are shown in Figure 111-17
as a function of the Ca/S molar feed ratio. Data were obtained at a pressure
of 930 kPa, at Ca/S molar feed ratios of 1.1 to 4 and over a temperature
range of 825 to 950°C. The emissions were significantly higher at tempera-
tures less than 900°C because of the stone's inability to calcine extensively.
Calcination greatly increases the stone's porosity making the active sites
more accessible to the SC>2 reactant. In the following discussion and figures,
calcining conditions refer to combustion temperatures above 900°C while
carbontaing conditions refer to temperatures below 900°C. The data in Figure
III->17 are replotted in Figure 111-18 as S02 retention vs. Ca/S molar feed
ratio. A 60% retention level was required to meet the EPA S02 emission
standards of 1.2 Ib S02/M BTU for the 2% sulfur Champion coal. The standard
was met with a Ca/S ratio of approximately 1.4 under calcining conditions
and approximately 2.3 under carbonating conditions.
Calcination Effects—
The Grove limestone variable study runs showed that the bed would undergo
full calcination when the temperature was raised from 880°C to 930°C. Cor-
relations of the dissociation pressure of CaCC>3 as a function of temperature
predict that calcination would have begun at a temperature of 916°C (see
Figure 111-19). The composition of the bed during the interval of steady
state operation is given in Table III-7 for many of the runs. At the lower
temperatures, the bed was largely composed of CaC03 and CaS04. Limited cal-
cination did occur near the fluidizing grid in a zone where the C02 concen-
trations were low. The extent of calcination of a particle with each pass
through the calcining zone depends upon the particle's residence time in that
zone, upon the CC^ concentration profile and on the thickness of the sulfa-
tion shell. A working model describing the progress of the calcination and
sulfation reactions is shown in Figure 111-20. Calcination occurs at the
boundary between the reaction shell and carbonate core. When a particle has
been recently introduced into the combustor, C02 within the particle can
readily diffuse from the particle and calcination proceeds rapidly when the
particle is in the calcining zone. Once the particle has been sulfated,
however, the C02 diffusion and calcination rates are much slower. The
increased porosity at the point of calcination makes the CaO accessible to
S02, however, C02 also competes for these sites once the particle reenters
the carbonating zone. The relative rates of the sulfation and carbonation
reactions in part, determine how rapidly limestone sulfates. The particles
reside mostly in the carbonating zone. Therefore, much of the CaO liberated
by calcination is reacted before the particles reenter the calcining zone.
The reaction zone in the more highly sulfated particles is likely small and
the rate of calcination might be the rate determining step. The operating
conditions can affect the solids recirculation rate between calcining and
carbonating zones, the extent of the calcining and carbonating zones, and
the C02 concentration profile. All of these factors affect the extent of
calcination. As an example, a higher gas velocity increases the recircula-
tion rates, causing a particle to see the calcining zone more frequently.
The utilizations in the shell of sulfation must range from 60-90% to
account for the overall utilizations of 20-30%. A greater porosity near the
particle's exterior and a more favorable pore structure generated by a slow
vs. fast calcination step may make the high utilizations possible.
47
-------
1600
FIGURE 111-17
S02 EMISSION VS. Ca/S RATIO FOR LIMESTONE NO. 1359
I I
T
1400
1200h
925-950°C
875-900°C
825-840°C
1000
E
Q.
Q.
CO
CO
LJ
C\J
O
in
800H
600
400
200
CARBONATING CONDITIONS
CALCINING
CONDITIONS
0
2 3
Ca/S (MOLF/MCLE)
48
-------
FIGURE 111-18
S02 RETENTION VS. Ca/S RATIO FOR LIMESTONE NO. 1359
100
UJ
I-
LU
ce:
90
80
70
60
50
CM 40
o
CO
30
20
10
0
0
CALCINING
CONDITIONS
CARBONATING
CONDITIONS
No. 1359 Limestone
A 925-950°C
• 875-900°C
• 825-840°C
123
Ca/S (MOLE/MOLE)
49
-------
TABLE III-7. UTILIZATION AND CALCINATION OF
LIMESTONE NO. 1359
Bed Composition
Run No.
19.4
19.5
19.6
19.7
19-9
20.2
26
29.1
30.1
30.2
30.3
30.4
31
32.1
Temp.
°C
880
890
820
940
950
950
885
875
885
930
885
835
838
950
Ca/S
Mole/Mole
2.65
2.7
3.2
4.0
2.56
3.75
2.75
2.8
1.5
3.3
1.6
1.1
3.0
3.2
CaO
Mole %
12
11
5
77
64
76
29
8
3
72
9
3
29
72
CaC03
Mole %
62
65
77
1
4
1
37
72
64
1
52
67
49
5
Mole %
26
25
18
21
32
23
33
20
33
27
39
30
22
24
50
-------
600
500
400
300
200
TO
CL
LU 100
= 90
80
70
13
60
CM
O
O
50
40
30
20
10
FIGURE 111-19
COMPARISON OF THE DEGREE OF SORBENT CALCINATION WITH
THE DISSOCIATION PRESSURE OF CaC03
Limited Calcination
Extensive Calcination
Dissociation Pressure of CaCO,
750
830
910
1000
AVE. BED TEMPERATURE, °C
51
-------
Carbonate
Core
(CaC03)
Sulfation
Reaction Shell
Shell /CaS04\
CaO \ CaO
'CaCOSl VCaCOo/
X •*
CO-
RADIAL POSITION IN A PARTICLE
FIGURE 111-20
MODEL FOR THE CALCINATION AND SULFATION
REACTIONS IN A CALCINING ENVIRONMENT
52
-------
Data obtained under both calcining and carbonating conditions can be
correlated using an empirically modified form of the Ca/S ratio which accounts
for the different calcination levels. If the calcium carbonate is considered
to be an inert component, the effective Ca/S ratio is given by:
Effective Ca/S = Ca/S
where X^CaQ + CaSQ \ is the mole fraction of the calcium present as the oxide
or sulfate. The correlation is given in Figure 111-21. As seen, all data are
well correlated by a single line, irrespective of the operating temperature.
Effect of Limestone Utilization on the Rate of Sulfation—
The sulfation reaction first order rate constant can be calculated with
experimental sulfur retention levels using the equation:
First order rate constant, k, (s'1) = - -+— In (100-Retention)
gas
, = Expanded Bed Height (m)
"gas Superficial Gas Velocity (m/s)
The dependence of the rate constant on the utilization is shown in
Figure 111-22. The activity decreases rapidly as sulfation occurs and
becomes negligible well before the stone is completely reacted. Utilizations
ranged between only 18 and 33% as compared to the sulfation level of 45-50%
which appears to be the maximum attainable in a fluidized bed combustor.
Limestone has an activity under calcining conditions that is three times
greater than that under carbonating conditions.
Effect of Other Variables—
Pressure—Runs 31 and 32.1 were made at reduced pressures of 520 kPa and
600 kPa, respectively. The SC>2 emissions from these runs are compared to
emission levels obtained from a correlation of data obtained at 930 kPa pres-
sure in Table III-8. The agreement is good enough to show that the pressure
effect is small or negligible.
Particle Size—In Runs 39.3 and 39.4, a limestone feed stream with a
size range of 8-14 mesh was used in the first test while a 14-25 mesh feed
was used in the other test. The mass average size ratio of the two feed
streams was approximately 2. The sulfur retention levels were 68.7 with the
fine feed and 70.3 with the coarse feed, a difference which is not signi-
ficant.
Operation at "Turndown" Conditions
The commercial application of pressurized fluid bed combustion (PFBC)
would require that the system be operated under "turndown" conditions. To
check the operability and sulfur removal capacity of PFBC under such con-
ditions runs were carried out at nominally, 700° and 690°C using Pfizer
dolomite and Grove limestone as sorbents. Results are shown in Table II1-9.
53
-------
FIGURE 111-21
Ln
-P-
100
80
P 70
60
0
SULFUR RETENTION VS. EFFECTIVE Ca/S RATIO WHICH ACCOUNTS
FOR DIFFERENT LIMESTONE CALCINATION LEVELS
0.5
1.0
_L
J_
1.5 2.0 2.5
EFFECTIVE Ca/S
820-830°C
870-885°C
930-950°C
3.0
3.5
-------
FIGURE 111-22
ACTIVITY VS. CALCIUM UTILIZATION FOR LIMESTONE NO. 1359
1.6
-1
1 4
in -1 -^
Jai"
•*
K-
Z
< 1.2
CO
•z.
o
0
u 1.0
<
c£
UJ
Q
Hi .8
o
I—
co
[L
.6
.4
.2
0
\
A TEMPERATURE
x • 825-840°C
\ • 875-900°C
A\ A A 925-950°C
\
CALCINING ^^ \
~ CONDITIONS \
A
\
\
A
\
\
—
\
\
• •X **
~ CARBONATING ^^ \ •
CONDITIONS ' N
\
\
" \
i i i i i
0 10 20 30 40 50
UTILIZATION (%)
55
-------
TABLE III-8. COMPARISON OF SULFUR RETENTIONS AT 930 kPa
AND 600 kPa WITH LIMESTONE NO. 1359
Ca/S Pressure Sulfur Retention
Run No. (Mole/Mole) Temp. (°C) (kPa) (%)
31 3.0 835 520 66
* 3.0 835-880 930 66
32.1 3.2 950 600 76
* 3.2 950 930 83
* Obtained from a correlation of S0? emissions for Grove limestone
given in Figure
56
-------
Ol
TABLE III-9. RESULTS OF RUNS AT TURNDOWN CONDITIONS
Operating Conditions: 38.1 38.3 38.4 38.5 38.6 39.1 39.2 39.3
A.vg. Bed Temperature, °C
Ca/S Molar Feed Ratio
Sorbent
Flue Gas Emission:
S02 1 ppm
NOX, ppm
CO, ppm
Sulfur Retention, %
Lb S00/M BTU
891
0.61
PD
761
135
175
40
2.29
762
0.94
PD
398
125
325
48
1.42
762
1.1
PD
223
149
375
71
0.78
690
1.5
PD
180
114
800
76
0.64
684
1.4
PD
163
102
400
82
0.5
750
2.5
GL
912
120
350
0
2.69
674
2.5
GL
900
174
—
6
2.64
938
3.5
GL
329
125
100
70.3
.80
PD = Pfizer Dolomite (No. 1337)
GL = Grove Limestone (No. 1359)
-------
Operations were very good during these tests; temperature control was
excellent and combustion was sufficient to give reasonable carbon monoxide
emissions. Near the end of the runs with dolomite, the temperature was
reduced further to determine the lowest feasible operating temperature. At
600°C, the temperature control was becoming erratic and carbon monoxide emis-
sions were much higher so no further reduction in temperature was attempted.
As indicated in Table III-9, the sulfur retention was maintained under
turndown conditions using dolomite as absorbent but retention was very poor
at the lower temperatures when limestone was used. The poor results with
limestone are due to inadequate calcination at the lower temperatures.
These results could have a strong influence on the choice of dolomite vs.
limestone in a commercial unit. The use of precalcined limestone may permit
its use under turndown conditions. This will be studied in the future.
Sorbent Feed Required to Meet the EPA S02 Emission Standard
A plot of sulfur retention vs. the sorbent mass feed rate (Figure 111-23)
shows that calcined Grove limestone (930°C combustion temperature) is far
more effective than the partly calcined stone (880°C combustion temperature).
It also shows that Pfizer dolomite is more effective on a weight basis than
calcined limestone at sulfur retention levels greater than 70%. While far
higher utilizations can be achieved with dolomite than with limestone,
dolomite contains only ~50 weight % CaC03 as compared to a ~100% CaCO^ con-
tent for limestone. Figure 111-23 can be used to estimate the sorbent feed
rate required to meet the EPA standard for any sulfur coal by calculating
the equivalent retention level. While the data was acquired with a 2% sul-
fur coal, the sulfation reaction kinetics are approximately first order in
S02, meaning that the conversion profiles should be independent of the
initial 862 concentrations or coal sulfur content. Table 111-10 giving the
sorbent feed requirements for 2, 3, 4 and 5% sulfur coals with 32.5 kj/g
heating value was prepared in this manner. At a coal sulfur level of 2%,
20% less limestone is needed as compared to dolomite to obtain the required
59% retention level. The sorbents are equally effective for a 3% sulfur coal.
At higher coal sulfur levels, dolomite becomes increasingly more attractive.
Thirty percent more limestone than dolomite is required to meet the standard
with a 5% sulfur coal. With this coal, the Ca/S molar feed ratios would be
3.25 for limestone and 1.3 for dolomite. These values form approximate
upper bounds on the molar feed rate requirement, as 90% of the U.S. coal
reserves have a sulfur content under 5%. The data given in Table 111-10
were estimated for a gas phase residence time of 3 s.
Determination of the Accuracy of the S02 Emission Data
Inert Bed Runs—
As shown in Table III-ll, excellent sulfur balances were calculated for
the five test conditions of the inert bed run series, 28.1-28.5. The balances
average 103%. Only 2.5% of the sulfur fed with the coal was retained by the
ash, the bulk of the sulfur appeared as S02 in the flue gas. The bed sulfur
analysis was 3.1%, higher than anticipated given the low surface area of the
alumina. Approximately 10% of the total sulfur feed was absorbed by the bed.
Almost certainly, much of the sulfur retention occurred during the initial
stages of the run and did not significantly affect the sulfur balances, which
are for a 2 to 3 hour period of steady state operation.
58
-------
FIGURE 111-23
COMPARISON OF DOLOMITE NO. 1337 AND LIMESTONE NO. 1359
AS S02 SORBENTS ON A MASS FEED RATE BASIS
UJ
U_
100
80
60
40
20
0
0
DOLOMITE
Limestone - 930°C
(Calcined)
Limestone - 880°C.
(Partly Calcined)
8
10
Kg FEED SORBENT
Kg COAL SULFUR
-------
TABLE 111-10. SORBENT REQUIREMENT TO MEET
THE EPA S02 EMISSION STANDARD (3)
Coal
Sulfur Retention Level Ca/S (Mole/Mole) kg Sorbent/100 kg Coal
(%) Req'd (%) (1) Limestone(2) Dolomite Limestone(2) Dolomite
2.0 59 1.3 0.8 8.2 10
3.0 73 2.1 1.0 20 20
4.0 79 2.8 1.2 34 29
5.0 84 3.2 1.3 51 40
(1) For coal with 32.5 kj/g (14,000 BTU/lb) higher heating value.
(2) Calcining Conditions
(3) Estimates are for 3 s gas residence time.
TABLE III-ll. SULFUR BALANCES - INERT BED RUN SERIES
kg Sulfur out OQ
Run No. kg Sulfur in
28.1 103
28.2 102
28.3 100
28.4 107
28.5 101
Avg. 103 + 3 (la)
Wet Chemistry S02 Determinations—
The S02 concentration in the flue gas was determined with a wet chemistry
technique during many of the runs. The purpose was to verify the accuracy of
the monitored S02 emissions. In most instances, the sample analyzed was drawn
from a position a short distance downstream of the flue gas sampling port,
before the gas had contacted the filter, pressure regulator or dryer used to
prepare the flue gas for the continuous analyzers. The g£s temperature at
the sampling.points was 130 to 230°C. In earlier runs, the sample was not
filtered. In later runs, a glass fiber filter was installed on the sample
line. No significant effect on SO- measurements was seen. The sample was
passed from a heated line to a series of bubblers containing an isopropyl
alcohol solution to absorb 803, and a hydrogen peroxide solution to absorb
the S02« The amounts of S02 absorbed in the hydrogen peroxide solution were
determined titrimetrically using sodium hydroxide as the titrant and methyl
orange as the indicator. Four determinations on S02 calibration gases gave
an agreement of 100 +5 (la) %.
The results are listed in Appendix Table C and the agreement between the
monitored (UV) and wet chemistry S02 concentrations is examined in Figure 111-24.
The line of best fit for the data is offset by about 70 ppm from the line ol
100% agreement with the monitored values higher. This implies that either a
flue gas constituent was causing interference with the Dupont UV S02 analyzer
readings or that the wet chemistry determinations were low. To determine
60
-------
FIGURE 111-24
COMPARISON OF MONITORED AND WET CHEMISTRY
SOo CONCENTRATIONS
Line of Best Fit for Data
200 400 600 800 1000 1200
S02 CONCENTRATION (ppm) - WET CHEMISTRY ANALYSIS
61
-------
which was the case, sampling for wet chemistry analysis was performed both at
the sampling port on the miniplant and in the control room from a point very
close to and the instruments during runs 34-39.4. The ratio between the wet
chemistry and monitored values was 98+8 (10) % for the determinations made
just prior to the instruments while the ratio was only 60 +_ 22 (la) % for
those determinations made at the sampling port. The discrepancy appears to
be with the wet chemistry determinations, which may be affected by the
samples' high dew point. The agreement is adequate, however, to give confi-
dence to the monitored S02 concentrations.
S03 Emissions
803 emissions in the flue gas were measured using a wet chemistry method.
803 was absorbed in an isopropyl alcohol solution which was titrated with
barium perchlorate to a thorin end point. Results are given in Appendix
Table C.
A total of 22 803 determinations made during runs 19-2-26 had an average
value of 5.6+6 (la) ppm. The 22 SOo determinations made during runs 27.2-
27.21 had a higher average value of 23 + 15 (la) ppm. The reason for the
higher 803 concentrations, whether process dependent or experimental, is not
known. Subsequent determination of 803 concentrations in runs 34-39 again
gave low values. Further work will be required to determine the level of
803 emissions.
Nitrogen Oxide Emissions
Nitrogen oxide emissions ranged between 50 and 200 ppm or 0.04 to 0.17g
(as N02)/MJ (0.1 to 0.4 Ibs/M BTU). The data are shown in Figure 111-25
where NOX emissions are plotted against percent excess air. Though the opera-
tion conditions varied greatly in runs 19-2-39.4, the only statistically
significant variables were the excess air (or flue gas oxygen concentration)
and bed temperature. The NOX emissions increased about 4 fold, from 0.04 to
0.17 g/MJ over the 5-100% range of excess air. The temperature effect in the
670-940°C range was secondary and caused only a 25% increase in the emission
level. The emissions are well below the EPA standards of 0.3 g (as N02)/MJ
(0.7 Ibs/M BTU) and have an average value of only 0.09 g/MJ (0.2 Ib/M BTU)
at 15% excess air.
The data from run 34, shown in Figure 111-26, suggest that in some cir-
cumstances the sulfation level of the bed and/or SOo emission le^el affect
the NOX emission level. When the Ca/S mole ratio of the Pfizer dolomite/coal
feed was increased from 0.75 to 1.5, the S02 emissions dropped from 300 ppm
to 0 ppm (steady state conditions were not reached) and the NO emissions
increased from 100 ppm to 200 ppm. The NO emission increased slightly more
to 225 ppm when the Ca/S ratio was increased to 2.5. Use of a sulfated
dolomite feed stream during the last four hours of the run caused the NO
emissions to drop to 65 ppm, while the S02 emissions rose to 60 ppm.
CO Emissions
CO emissions were independent of the bed temperature between 825 and
950°C and ranged from 70-225 ppm. At turndown temperatures of 680°C and
760°C, the emissions were 4 to 5 times higher. The emissions are plotted
62
-------
FIGURE 111-25
NOV EMISSIONS VS. EXCESS AIR
A
U.O
0.7
_ 0.6
ID
1 0.5
\
CQ
d 0.4
X
O
^ 0.3
0.2
0.1
0
1 1 1 1 1 I
« EPA EMISSION STANDARD
•^^^"^ ' ' i_i n L_ ivi i o o i v/ 1 V *J 1 1\ \\ ur\i\ u
- -
— __
0*
^ _^ — ——^~"~—~
0 2 __.,«—• — -^"'"1"— •
" ••*••* ^^ •"^•r^ * *
'v^^* * •
_ ^^» * * -
*dl** «^* *
sf*fjf • *
"• ** •
i i i i i i
V • -^ 1 ^T
0.301
0.258
0.215
0
X
0.172 <£*
^
^
0.129
0.086
0.043
n
0 20 40 60 80 100 120 140
EXCESS AIR, %
-------
FIGURE 111-26
300
IMOX EMISSIONS DATA - RUN No. 34
27
250
225
200
E
a.
a 175
o
co 150
—
". 125
100
75
Ca/S 0.75
S02 300 ppm
Ca/S
SOo
1.5
0 ppm
Ca/S
2.5
0 ppm
Sulfated stone
S02 60 ppm
50
25
0
10:30 12:30
2:30
4:30
TIME
6:30
8:30
10:30
64
-------
in Figure 111-27. The heat losses due to incomplete combustion of CO to CC>2
were low, and ranged from 0.05-1.0% at higher temperatures to 0.6% at turn-
down temperatures as shown in Table 111-12. CO emissions were independent
of the sorbent used.
TABLE 111-12. HEAT LOSS DUE TO INCOMPLETE
COMBUSTION OF CO TO C02
Temperature Heat Loss as Percent
Run No. (°C) of Total Heat Input
38.5 690 0.59
38.6 680 0.25
38.4 760 0.27
38.3 760 0.20
39.1 760 0.21
825 - 96° °'05 - °'10
Particulate Emissions
Solid matter is emitted from the miniplant combustor from several points.
The major source of solid matter is used sorbent removed from the combustor.
This material consists largely of sulfated sorbent with a smaller quantity
of coal ash and char. The average size is usually only slightly less than
that of the fresh sorbent. Most of the coal ash and some used sorbent fines
are entrained (or elutriated) from the combustor by the flue gas and enter
the flue gas particulate removal system. Currently, this consists of two
cyclones operated in series. The first cyclone captures the large parti-
culates and returns them to the combustor to increase the carbon combustion
efficiency. The amount and particle size of the particulates entering the
first stage cyclone and the particulates recycled to the combustor cannot,
at the present time, be measured. The particulates leaving the first stage
cyclone with the flue gas consist of flyash and some attrited sorbent fines.
This particulate stream then enters the second stage cyclone and the portion
of it which is retained by the second stage cyclone is routinely measured,
analyzed and sized. Flue gas leaving the second stage cyclone is sampled
through an isokinetic sampling system and the particulate loading, size
distribution and composition are determined.
In the following sections, these particulate emissions are discussed.
Sorbent Elutriation Rates —
Elutriation rates for Grove No. 1337 limestone and Pfizer No. 1359
dolomite were calculated from the quantity of calcium in the flyash collected
by the second cyclone. The flyash sample analyzed from runs prior to 27.1
were blended to be representative of the entire run. This was done by mixing
the samples of the different lockhopper dumps in the proper proportion. In
later runs, individual dumps collected over 1 to 3 hour intervals of steady
state operation were analyzed to obtain data which would be more representa-
tive of the condition tested. Sorbent losses to a large extent, represent
the elutriation of the fines formed by attrition. Most bed particles
coarser than 100 microns are captured by the first cyclone and recycled.
65
-------
0.8
0.7
CQ 0.6
8 0.5
OQ
. 0.4
£ 0.3
LU
O
O
0.2
0.1 -
>00
FIGURE 111-27
CARBON MONOXIDE EMISSION VS. TEMPERATURE
\
\
\
\
\
\
\
\
CO Emissions Range from
70 to 225 ppm
Between 825 & 950°C
X
700
800
900
1000
TEMPERATURE °C
-------
O
O
o
h-
co
ID
CO
o
o
03
O
LU
UJ
ra
0
O
O
UJ
FIGURE 111-28
EFFECT OF GAS VELOCITY AND Ca/S RATIO ON ELUTRIATION OF DOLOMITE No. 1337
160
120
80
40
1
s /
• /A
/
/ t
I
f
t .»' *
•
*
• Ca/S - 0.5-0.7
• Ca/S - 0.9-1.3
A Ca/S - 1.5-2.1
1
0.8 1.6 2.4
SUPERFICIAL GAS VELOCITY -m/s
3.2
-------
The elutriation rates for both calcined and uncalcined Grove limestone
were low, as shown in Table 111-13. Detailed data are given in Appendix
Tables D, E and F. Over the range of Ca/S feed rates used, 1.45-4.0, and
ranged of superficial gas velocities, 1.4-2.2 m/s, the elutriation rate
represented ~10% of the sorbent feed. The limestone feed rate required to
match the loss was equivalent to a Ca/S feed ratio of about 0.24. The elutria-
tion rates for calcined and uncalcined limestone were not significantly dif-
ferent. Data from runs 19.7, 20.1 and 21 were higher than the average. The
reason is not known. Rates measured at low temperatures, 680 to 750°C suggest
the attrition and elutriation rates are lower in this temperature range (see
Appendix Table D.
TABLE 111-13. SORBENT ELUTRIATION LOSSES
Elutriation Losses
Sup. Vel. Feed Rate (Equiv. (% of (Vol. %
Sorbent (m/s) (Ca/S) Ca/S) Feed) Bed hr)
Limestone
Grove 1359
- uncalcined 1.4-2.2 1.5-2.8 0.2 12 1.1
- calcined 1.6-2.5 2.5-4.0 0.2 8 0.8
Dolomite
- low velocity 1.5-2.1 0.6-1.8 0.4 38 2.1
- high velocity 3 0.5-1.3 0.8 110 5.5
Elutriation rates for dolomite, given in Table 111-13, were substantially
greater than those for limestone. Over a range of velocities from 1.5-2.2
m/s, the rates averaged 38% of the sorbent feed and the loss was equivalent
to a Ca/S feed ratio of 0.4. At a velocity of 3 m/s, 110% of the sorbent feed
was elutriated on the average and the Ca/S equivalent was 0.8. ANL (6) was
able to correlate elutriation data well for Tymochtee dolomite by plotting
the fraction of the feed calcium entrained from the combustor against the
gas velocity. The fraction of the feed elutriated was found to be indepen-
ent of the Ca/S ratio. This approach was used to correlate the data shown
in Figure 111-28. The elutriation rates for the miniplant also appear to be
largely independent of the Ca/S ratio and to increase with gas velocity.
There is much uncertainty to the slope of the line of best fit, due to the
scatter in the data and to the lack of data at higher velocities. A compar-
ison of the Argonne and miniplant correlations shows that the elutriation
rates in the Argonne unit are higher. The primary reason for this is that in
the miniplant, most of the fines are recycled by the first cyclone. In runs
with dolomite, over one-half of the bed material had a terminal velocity
lower than the superficial gas velocity and the bed contained particles as
fine as 200 microns. The recycle rates must therefore, be substantial. The
maximum feasible operating velocity, that velocity at which 100% of the feed
is elutriated is 1.9 m/s in the Argonne unit. The maximum velocity in the
miniplant has not been determined with accuracy, but is below 3 m/s.
67
-------
Partlculate Emission Levels—
Particulate emission levels in the flue gas leaving the second stage
cyclone were measured using the sampling system described earlier in this
section. Data are shown in Table 111-14 for runs 28 through 39. Over this
period the sampling system appeared to be operating satisfactorily with
deviations from isokinetic conditions generally less than + 15%. Except for
the earlier runs (through 30.1), particulate levels ranged from 0.9 to
4.8 g/m3 (0.4 to 2.1 gr/SCF) and averaged 2.5 g/m3 (1.1 gr/SCF). The spread
in the measured values is greater than expected and could be due to experi-
mental problems with the sampling system or variable performance of the
second stage cyclone.
The particulates captured in the second stage cyclone represent a con-
centration in the flue gas of 7 to 18 g/m3 (3 to 8 gr/SCF). Therefore, the
concentration of particulates entering the second stage cyclone, obtained
by summing the amount captured in the cyclone and the concentration measured
in the flue gas leaving the cyclone is in the range of 7 to 23 g/m3 (3 to 10
gr/SCF).
Particulate Size Distribution—
Size distributions were measured for the spent sorbent removed from the
combustor during a run, the material retained in the second stage cyclone,
and the fine particulates captured in the flue gas particulate sampling
system.
Spent Sorbent—The spent sorbent removed from the combustor during the
course of a run and the sorbent remaining in the combustor after a run was
completed were sized by sieve analyses. The results are given in Table 111-15
and are compared to the size distribution of the sorbent fed to the combustor
during the runs. Detailed data are given in Appendix Tables G and H. As
seen in Table 111-15, the spent sorbent is significantly reduced in size,
showing the effects of particle attrition and recycle of the first cyclone
capture back to the combustor. If particles were not recycled from the first
cyclone, an increase in the mass median size of the spent sorbent would be
expected due to the loss by entrainment of the smaller particles. It can
also be seen that spent limestone sorbent is significantly coarser than spent
dolomite, reflecting the higher attrition rate of dolomite. The spent sor-
bent removed from the combustor during a run is also coarser than the material
sampled from the bed after a run has been completed. The reason for this
result is not known, but may be due to particle size segregation within the
combustor. The spent sorbent is removed from a port in the combustor wall
and it is possible that coarser particles may be gathering at the wall. This
could be due to the solids circulation patterns within the combustor which
are believed to be upward in the center and downward at the wall. Smaller
particles carried up the center of the combustor may be entrained into the
cyclone system, while the larger particles are disengaged in the freeboard
section of the combustor and fall down near the wall where they can be
preferentially removed through the solids removal port.
It can also be seen that the particle size distribution data show large
variances from run to run as indicated by the standard deviations given in
Table 111-15. This could be caused by a number of operating factors such as
differences in superficial velocity, bed depth, etc.
69
-------
TABLE 111-14. MINIPLANT FLUE GAS PARTICULATE SAMPLING SUMMARY
Run No.
28.1
28.4
28.5
29
30.1
31
32.2
32.3
33.1
34
36.2
37
39.1
39.2
Sampling Time
(hrs)
70
30
25
58
2.58
00
75
50
47
4.42
2.
3.
,58
,00
2.5
3.5
Total Solids
Collected
(g)
Solids Rate in Probe
(g/min) (gr/min)
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
0.17
0.05
0.09
0.12
0.09
0.38
0.15
0.18
0.29
0.32
0.50
0.48
1.17
1.24
2.56
0.77
1.44
1.86
1.40
5.89
2.33
2.79
4.49
4.97
7.71
7.41
18.08
19.28
Deviation from
Isokinetic Sampling
0
0
0
-16
-22
-7
+4
-12
+11
+11
+10
-48
+3
+3
Solids Loading
in Flue Gas
(mg/m3) (gr/SCF)
590
370
300
530
430
2880
920
1650
1420
1580
2200
2360
4813
4813
0.26
0.16
0.13
0.23
0.19
1.26
0.40
0.72
0.62
0.69
0.96
1.03
2.10
2.10
-------
TABLE 111-15. SPENT SORBENT PARTICLE SIZE DISTRIBUTION
Particle Size (vim)
Material 10% Less Than 50% Less Than 90% Less Than
Sorbent Feed 880 1640 2280
Spent Limestone
Removed During Run 820 + 90 1360 + 130 2160 + 110
From Combustor After Run 620 + 180 1140 + 180 1980 + 150
Spent Dolomite
Removed During Run 370 + 180 1230 + 300 1940 + 390
From Combustor After Run 340 + 230 890 + 430 1820 + 310
Solids retained in second stage cyclone (flyash)—The material retained in
the second stage cyclone is a mixture of flyash, unburned carbon residue and
attrited sorbent fines. Composition data are given in a later section. This
material has passed through the first stage cyclone and is fairly fine.
Table 111-16 gives particle size distribution data.
TABLE 111-16. PARTICLE SIZE DISTRIBUTION - FLYASH
Particle Size (urn)
Number
of Runs 10% Less Than 50% Less Than 90% Less Than
42 6 + 2 23 + 16 103 + 69
29 5 + 1 18 + 3 66 + 9
12 7 + 1 30 + 11 171 + 67
As seen, if data from 42 runs are included, the particle size distribu-
tion is 6, 23 and 103 ym for the three points in the distribution. However,
if data are rejected which fall outside the 2s limits, the 29 remaining runs
show a size distribution of 5, 18 and 66 ym for the 10, 50 and 90% points.
It is then obvious that the data from the 42 runs represent two size distri-
bution populations. Twelve of the remaining 13 runs were averaged and gave
size distributions of 7, 30 and 171 ym corresponding to the 10, 50 and 90%
points in the distribution. One run was rejected even from this population.
The major differences between the two particle size populations is at the
upper end of the distributions, 66 vs. 171 ym at the 90% point in the distri-
bution. It is believed that the difference is due possibly to variable per-
formance of the cyclones. Some minor modifications have been made to the
cyclones and more recent data are typified by the smaller size distribution.
Detailed data are given in Appendix Table I.
Fine Particulates—Fine particulates sampled isoklnetically from the
flue gas were sized using a combination of wet sieving and a Coulter Counter.
The average size for six runs are given in Table 111-17. As seen, the size
71
-------
distributions are fairly uniform, and average 2, 7 and 32 jam at the 10, 50
and 90% points in the size distributions. More detailed data are given in
Appendix Table J.
TABLE 111-17. PARTICLE SIZE DISTRIBUTION -
FINE FLUE GAS PARTICULATES
Particle Size (ym)
Run No. 10% Less Than 50% Less Than 90% Less Than
31 1.8 8.0
32.2 2.0 5.8 24
32.3 2.3 5.9 32
33 3.3 7.6 26
36.2 2.5 7.5 30
37 2.0 6.5 30
Avg. 2.3 + 0.5 6.9 + 0.9 32 + 9
Particulate Composition—
Typical particulate chemical analyses for the used bed, solids removed
from the bed during the run, flyash collected in the second stage cyclone
and particulates in the flue gas leaving the second stage cyclone are given
in Table 111-18. Table 111-19 gives data for the same runs expressed as
percent carbon, ash and sorbent and also gives the CaO, CaC03 and CaSO/ con-
tent of the sorbent portions of the various solids. Runs were included which
covered a range of sorbent types, Ca/S ratios and temperatures. Detailed
data are given in Appendix Tables K and L.
In general, the spent solids removed from the combustor contain less than
1% combustible carbon except at very low temperatures where carbon contents
may be as high as 3%. Ash content in the used bed material varies from 2 to
25% and the sorbent content varies from 75 to 95%.
The flyash collected in the second stage cyclone generally contains 3 to
20% carbon. However, it tends to be lower when dolomite sorbent is used be-
cause of the dilution effect of the larger amount of entrained dolomite fines
in the flyash. In this case, carbon contents as low as 3 to 5% are seen,
especially in runs made at the higher combustor temperatures. The ash content
in the collected flyash is 65 to 80% when limestone sorbent is used compared
to 40 to 60% for dolomite sorbent. The sorbent content in the flyash gener-
ally runs from 15 to 20% for limestone compared to 30 to 60% for dolomite.
The fine particulates filtered from the flue gas contain 3 to 9% carbon.
In general, fine particulates from runs made with dolomite contain lower car-
bon concentrations, due in part at least, to the dilution effect of higher
72
-------
TABLE 111-18. TYPICAL PARTICIPATE ANALYSES
LO
Run No.
39.2
31.1
32.3
36.2
37
Sorbent
Limestone
Limestone
Dolomite
Dolomite
Dolomite
Temperature
2.5 670
3.0 840
1.5 840
1.3 900
0.5 900
Source
Ca
Bed-End of Run
Bed-Rejected Solids
Fly Ash
Fine Particulates
Bed-End of Run
Bed-Rejected Solids
Fly Ash
Fine Particulates
Bed-End of Run
Bed-Rejected Solids
Fly Ash
Fine Particulates
Bed-End of Run
Bed-Rejected Solids
Fly Ash
Fine Particulates
Bed-End of Run
Bed-Rejected Solids
Fly Ash
Fine Particulates
32
6.
3.
38
33
6.
4.
29
9.
11
26
16
6.
22
12
4.
2
9
8
2
8
2
6
-
-
-
—
-
-
-
9
3
2
13
8
3
10
7
2
Sample
-
—
-
_
-
-
-
6
7
12
20
22
7
7
29
Sample
.5
.2
11
8
37
Sample
.9
.6
21
9
46
Sample
.3
.8
23
7
Not
.6
.4
.9
.2
.7
Not
.1
Not
.5
Not
.0
Taken
31
1
0
28
12
1
0
15
.1
.2
.3
.9
Taken
1
0
12
.7
.6
Taken
1
0
2
.8
.6
.3
Taken
0
0
.8
.2
3
5
4
0
0
18
7
0
12
2
N.D.
2
3
N.D.
2
3
.3
.8
.1
.1
.5
.1
.2
.1
.2
.4
(1)
.8
.6
(1)
.8
.6
(1) N.D. - Not Determined
-------
TABLE 111-19. TYPICAL PARTICULATE COMPOSITION
Run No. Sorbent
Ca/S
(m/m)
Ca
39.2 Limestone 2.5
31.1 Limestone 3.0
32.3 Dolomite 1.5
36.2 Dolomite 1.3
37 Dolomite 0.5
Temperature
670
840
840
900
900
Source
Sorbent
- Composition (wt %) - Composition (m %)
C Ash Sorbent CaO CaC03 CaS04
Bed-End of Run
Bed-Rejected Solid
Fly Ash
Fine Particulate
Bed-End of Run
Bed-Rejected Solid
Fly Ash
Fine Particulate
Bed-End of Run
Bed-Rejected Solid
Fly Ash
Fine Particulate
Bed-End of Run
Bed-Rejected Solid
Fly Ash
Fine Particulate
Bed-End of Run
Bed-Rej ected Solid
Fly Ash
Fine Particulate
3.3
5.8
4.1
0.1
0.5
18.1
7.2
0.1
12.2
2.4
N.D.^
2.8
3.6
N.D.™
2.8
3.6
24
78
79
10
26
65
80
9
58
71
2
41
73
13
48
79
- Sample Not
73
16
16
90
73
16
13
90
O 1 ItT 4-
~ O elLLlp J_ G IN O t
30
26
(2) gg(2)
56
23
(2) 8?(2)
O 1 j-\ M 4-
•"• oa.in.'pJ.G INOu
49
17
T ilf on
J.£LK.Gn
26
38
-41
29
48
43
11
23
rp -i
J.3.K.GH
42
64
13
38
30
3
rp -I
J.3.K.GT1
14
35
65
12
3
49
24
13
14
34
12
4
30
7
6
7
5
2
9
50
138
22
27
44
76
43
46
32
58
55
64
90
81
63
(1) N.D. - Not Determined
(2) Carbon Free Basis
-------
concentrations of sorbent fines. The sorbent portion of the fine particulates
generally ranges from 15 to 30% for dolomite sorbent and 10 to 20% for lime-
stone sorbent. The ash content of the fine particulates generally ranges
from 70 to 80% for both limestone and dolomite sorbent.
Combustion Efficiency
Combustion efficiency data are correlated as a function of the bed temp-
erature in Figure 111-29. The efficiency was calculated as one hundred minus
the percentage of the combustible carbon in the feed lost with the flyash.
Three distinct profiles were obtained depending on the run series. The pro-
files converge to 99 + % at 940°C. However, at lower temperatures, the tem-
perature effect differs greatly. The lowest efficiencies were obtained from
runs 19.2-26, and the highest efficiencies were obtained from runs 34-39.4.
Results from runs 27.1-32.2 were between those from the other two series. The
average efficiency at 880°C was, respectively, 96.5, 98.2 and 99.3%. The
average efficiencies from the turndown tests were 98.2% at 760°C and 98.8% at
680°C. These values are much greater than the efficiencies which would be
predicted from the two lower profiles. The data from runs 27.1-32.2 were
examined for secondary variable effects using regression analysis. The
excess air level, gas velocity, bed height and pressure were, however, not
statistically significant.
The three temperature dependencies shown by the data in Figure 111-29
have yet to be explained. A variety of explanations were considered, includ-
ing process variables, the equipment, and sampling and analysis. The process
variables examined were the coal feed rate, combustion intensity, excess air,
velocity, sorbent, and coal chemical and size characteristics. None gave any
indication of affecting the efficiency. Variations in the performance of the
first and second stage cyclones and the first cyclone's flyash recycle system
might affect the combustion efficiency. If the refractory lining was eroded
or the barrel damaged, the collection efficiency would be impaired. This
would mean a lower recycle rate from the first cyclone causing lower combus-
tion efficiencies. The efficiencies were higher rather than lower in the
later runs, however. The performance of the first cyclone could also be im-
paired by a backward gas flow in the cyclone's dipleg if a seal was not formed.
Particle size analysis of the second stage cyclone collection and temperatures
in the first cyclone dipleg during operation suggest that this did not happen.
A poor second stage cyclone collection does cause carry over to the off gas of
some carbon bearing flyash particles. This elutriated carbon is not included
in the combustion efficiencies calculated for Figure 111-29. The magnitude
of this loss was evaluated from the chemical composition of the isokinetically
sampled off gas particles and from a flyash balance. Table 111-20 shows that
carbon losses with the off gas particulates are small (0.2-0.4% of the feed
carbon) and are independent of the flyash carbon content and cyclone collec-
tion efficiency. The different temperature dependencies in Figure 111-29 are,
therefore, not due to varying carbon losses in the off gas from the second
cyclones.
The sampling procedure differed in Runs 19.2-26 from that used in later
runs. In the earlier runs the carbon content of the entire flyash collection
for a run was used in the calculation of the efficiency, including the non-^
representative amounts of unburned carbon remaining from startup and operating
upsets. In the later runs, only the flyash collections made during a one to
three hour period of stable operation were used for the carbon determination.
There was an exception, Run 19.3, where three one-hour long collections were
75
-------
O
z
LJ
O
LJL
LJL
UJ
z.
o
h-
co
ID
CD
^
O
O
100
99
98
97
96
95 -
941
640
FIGURE 111-29
COMBUSTION EFFICIENCY VS. TEMPERATURE
• Runs 27.1 - 32.3
• Runs 34 - 39.4
O Runs 19.2 - 26
1
* / ' •
•>' /
/%
0
o
/
/ o
700
760
820
880
940
1000
AVE. TEMPERATURE - °C
-------
made and analyzed for carbon. The combustion efficiency for these collections
averaged 96.8% at 880°C as compared to 96.1% for several runs made at the same
temperature but with the entire flyash collection used in the calculation.
Apparently the difference in the sampling procedure cannot explain the dif-
ference in the profiles for Runs 19.2-26 and Runs 27.1-32.2". The sampling
procedure was identical for Runs 27.1-32.2 and 34-39.4. The accuracy of the
carbon determinations was confirmed with a known sample. Also, several fly-
ash samples from early runs were reanalyzed with good agreement with the
original results.
TABLE 111-20. COMBUSTIBLE CARBON LOSSES
Combustible Carbon in the Combustible Carbon
Second Cyclone Flyash Second Cyclone Collection in Off Gas
Run No. Collection Eff. (%) (% of Feed Carbon) (% of Feed)
31.1 83 3.1 0.2
3.5 0.2
32.2 53 0.7 0.4
0.4 0.4
32.3 55 3.0 0.2
2.2 0.2
36.2 71 0.5 0.2
0.7 0.2
37 60 0.6 0.2
Heat Transfer Coefficients
Heat transfer coefficients were measured by maintaining the cooling water
flow through the coils in tne liquid state. The flow to each coil was meas-
ured with an orifice meter. The inlet and outlet temperatures of the cooling
water were measured with thermocouples inserted within the piping. The flows
and temperatures were recorded on the data logger at one minute intervals.
These data were used to calculate the average coefficient and standard
deviation for a 10 minute interval. Typical results from Run 19-2 are shown
in Table 111-21. The average of the three good measurements, for coils 1A,
IB, and 2B was 334 W/m2 K which compares closely with a value of 358 W/m2 K
calculated from an overall heat balance. The cooling coils removed 57% of
the heat input, and the remaining portion was removed as sensible heat by
the fluidizing air or lost to the surroundings. The combustion intensity
had a high value of about 5 MW/nP expanded bed.
The heat transfer measurements made since run 19.2 are tabulated in
Table 111-22. Most of the data were obtained during the turndown tests.
These data are the most consistent and were used to determine the effect of
some of the operating parameters. The data are plotted in Figure 111-30,
against the bed temperature. The overall coefficient was lowered by 10% when
the temperature was reduced from 950°C to 680°C. Some of the decrease may
have been caused by a 20% reduction in superficial gas velocity at the lower
temperatures. Much of the decrease probably occurred, however, because of a
reduction in the radiation component of the heat flux. The magnitude of the
77
-------
TABLE 111-21. MINIPLANT OVERALL HEAT TRANSFER
COEFFICIENT MEASUREMENTS - RUN 19.2
Coil #
1A
IB
2B
Average Coeff.
352 W/m2 K
318
329
Standard Deviations
of 10 Measurements
Obtained At One
Minute Intervals
7.2 W/m2 K
7.7
11.3
Surface Area of a Coil, m2 0.551
Coil Heat Flux, W/m2 280,000
Combustion Intensity, W/m Bed 5,250,000
Heat Removed by Cooling Coils in 57
Bed, % of Coal Heat Input
Overall Heat Transfer Coeff.,
Calculated from Heat Balance, W/m2K 358
78
-------
TABLE 111-22. HEAT TRANSFER COEFFICIENTS
Run
Number
19.2
28.1
29.1
32.1
38.1
38.2
38.3
38.4
38.5
38.6
39.1
39.2
39.3
39.4
Pressure
(kPa)
930
930
930
600
930
930
930
930
930
930
930
930
930
930
Temp.
°C
880
840
875
950
890
890
760
760
690
685
750
675
940
940
Sup. Vel.
(m/s)
1.9
2.1
2.2
2.2
2.1
2.1
1.85
1.85
1.7
1.3
1.5
1.4
1.6
1.6
Mass
Average
Bed Part.
Size (y)
1580
1089
1290
1060
745
745
745
745
745
745
1235
1235
1235
1235
Overall
1A
352
352
295
273
341
352
329
335
324
329
290
250
290
301
Coefficient
. IB 2A
318 179
346
363 335
341
415
420
398
398
386
380
307
295
324
341
(W/m2R)
2B
329
329
-------
FIGURE 111-30
HEAT TRANSFER COEFFICIENTS VS. TEMPERATURE
450
^
o
CvJ
410
UJ
o 370
LJ
O
O
01
UJ
U_
UJ
0=
330
290
250
COIL1A
COIL1B
O A Mass Avg. Pact. Size 745/*
(2.1)
• A Mass Avg. Part. Ax X"
Size 1235/^ XA (2.])
( ) Sup. Gas Vel. /
m/s
(1.7)
(2.1)
o ^.
(1.3)
(1.85)
o.
(1.7)
(1.6)
o
(2>1)
o
(1.85)
(1.5)
(1.4) (1.5)
(1.6
(1.6)
(1.6)
(1.4)
600 700 800 900 1,000
AVG. BED TEMPERATURE, °C
80
-------
radiation component was estimated by choosing, through trial and error, an
emissivity which would make the outside film coefficient independent of the
bed temperature. An emissivity of about 0.2 was calculated and the radiation
component was estimated to be about 12% of the heat flux at 950°C.
The bed was much finer in runs with dolomite because dolomite has a
higher attrition rate than limestone. The 40% difference in the mass average
size, 745 ym vs. 1235 ym, in the turndown tests with dolomite and limestone
increased the overall coefficient in the dolomite tests by 20 to 30%.
The location of the coil also had an effect on the coefficient. The
lower most cooling coil 1A, located 0.46 m to 0.92 m above the fluidizing
grid, had coefficients 15-20% lower than coil IB, located 0.92 m to 1.38 m
above the grid. The coefficient was largely independent of position among
the coils located higher in the bed, when a number of coils greater than two
was used. The difference between the coefficient for the lower two coils
probably occurs because of varying mixing patterns. The gas jets from the
fluidizing grid require a penetration depth before a uniform flow field is
established and, therefore, the mixing may be more vigorous in the vicinity
of the second coil.
After run 27 the vertical coils were baffled with rings to prevent
erosion of the coils. The baffles which acted as fins added about 15% to the
surface area of the coil. The average coefficient measured for the baffled
coil compared closely with the average coefficient for unbaEfled coils (331
vs. 335 W/m^K). It is possible that the stagnant areas created by the baffles
reduce the bed to tube heat transfer coefficient slightly, thereby offsetting
the increase in surface area.
BATCH COMBUSTOR
Equipment, Materials, Procedures
Fluidized Bed Coal Combustion Unit—
A schematic of the Exxon batch fluidized bed combustion unit is shown in
Figure 111-31. The fluidized bed eombustor vessel was constructed from four
sections of 25 cm diameter standard wall carbon steel pipe and refractory
lined to an inside diameter of 11.4 cm. The height of the vessel above the
fluidizing grid was about 4.9 m. Below the grid was a 61 cm long burner
section lined with Grefco Bubbalite. The fluidizing grid, which was made of
stainless steel, had 21-0.32 cm diameter holes to distribute the fluidizing
air. A natural gas burner, located below the grid, was used to preheat the
unit to above the ignition temperature of coal. The eombustor had three
0.95 cm diameter stainless steel vertical cooling coils extending from 27 to
144 cm above the grid. Each coil had a surface area of 0.060 m^. Thermo-
couples were located 10 cm apart in the lower section of the eombustor and
30 cm apart in the upper section.
Although sorbent material was added batchwise to the eombustor, coal
was fed continuously using a modified Petrocarb Model 16-1 ABC injector.
The injector consists of a conical-bottom tank that holds coal to be fed
and an orifice and mixing tee assembly that mixes coal with carrier gas.
The coal feed rate is controlled by injector/combustor differential pres-
sure and transport air flow rate.
81
-------
Cyclones
Drain
City Water i—i
Demineralizer
Air From
Compressor
Propane
Water
Startup
Heater
Sampling
System
Filter
Pressure
Control
Valve
Platform
Scale
Coal
«—— Injection Air
Vent
Off-Gas
Chiller
T
Knockout
FIGURE 111-31
BATCH FLUIDIZED BED COAL COMBUSTION UNIT
-------
Flow of air and fuel into the combustor and combustor pressure were under
automatic control. Gases leaving the combustor first passed through two
cyclones x*hich removed flyash and entrained stone. An off-gas cooler, which
followed the cyclones, reduced the temperature of the off-gas to the desired
level. Th& off-gas then entered a 2.5 cm diameter stainless steel expansion
coil which was electrically heated during startup to raise the temperature of
the gas above the dew point. A 3.8 cm Aerotec cyclone, following the heater,
was used to remove particulates during startup. Fine particulates were removed
by a Pall Model MEC-800-18-C filter, located upstream of the back pressure
control valve. Before being vented, the off-gas passed through a chiller and
knockout for moisture removal. A small portion of the off-gas was extracted
after the back pressure control valve and sent to the gas conditioning and
analysis system for S02, NOX, C02, and 02 measurement.
A more detailed description of the components and development of the
batch f luidized bed combustion unit can be found in an earlier report (2) .
Coal/Sorbents Tested —
Three different coals were burned in the batch f luidized bed combustion
unit. The majority of the runs were made using a high volatile (A) bituminous
Eastern coal from the Arkwright mine ground to -16 mesh. Runs were also made
using a low sulfur Western coal and a high sulfur Illinois coal. A proximate
and ultimate analysis for each of the coals are presented in Table 111-23.
Grove limestone and Tymochtee dolomite generally in the 8 X 25 mesh range
were the primary sorbents used in the experimental studies. Baker dolomite
and Pfizer dolomite were also tried but, because of high attrition rates, their
use was discontinued. An analysis of the sorbents is given in Table 111-24.
Experimental Procedures' —
Operation of the batch fluidized bed combustor can be divided into four
phases: startup, ignition and pre-heating, coal feeding, and shutdown. Start-
up consisted of those activities preliminary to ignition of the preheat
burner. These activities included equipment checkout, calibration of the gas
analyzers, charging. solids, and turning on electrical circuits, the air
compressor, the cooling water systems (fluidizing grid, burner, steam coils,
condenser) and the purge air systems (pressure taps, sight-glasses, AP cells).
To ignite the preheat burner, air and fuel flows were set and the
tion electrode was activated. Safety devices shut down all flows if igni-
tion was not obtained within ten seconds or if a flame-out occurred afterwards.
A safety interlock prevented startup for 3 minutes after an automatic shut-
down to assure adequate purging of the combustor. Subsequent to ignition,
air flow and combustor pressure were adjusted to the desired values. All
gas flows and pressure were controlled automatically. After the bed tempera-
ture was sufficiently high, supplementary fuel could be injected directly
into the bed to reduce the time required to heat the bed to the coal ignition
temperature.
Preparation of the coal feed system for a run consisted of setting the
flow of injection air and activating and adjusting the coal feeder-to-
combustor AP control system. Coal injection could be started only after the
temperature in the combustor was high enough for self-ignition of the coal
83
-------
TABLE 111-23. COMPOSITION OF COALS USED IN
BATCH FLUIDIZED BED COAL COMBUSTION PROGRAM
Proximate Analysis Ultimate Analysis
Coal Moisture
00
-p-
Arkwright
Illinois
Western
1.00
3.67
19.45
Ash Volatiles
8.11
10.25
7.08
36.86
39.50
36.89
Fixed
Carbon Moisture
54.03
46.59
36.58
1.00
3.67
19.45
Ash
8.11
10.25
7.08
Total
Carbon Hydrogen
76.26
67.23
55.12
5.30
4.79
0.60
Sulfur
2.66
4.24
0.60
Nitrogen
1.49
1.19
0.77
Chlorine
0.07
0.07
0.02
Oxygen (1)
5.11
8.64
13.02
Notes: All values are weight percent
(1) by difference
-------
TABLE 111-24. PROPERTIES OF LIMESTONE AND DOLOMITE
Chemical Analysis, Wt. %
Designation Quarry Source Stone Type CaO MgO Si02 A1203 Fe20;
1359 Grove Limestone Limestone 97.0 1.2 1.1 0.3 0.2
(Stephen City, Va)
1337 Chas. Pfizer Co. Dolomite 54.0 44.0 0.9 0.2 0.3
(Gibsonburg, Oh)
Tymochtee C. F. Duff & Sons Dolomite 53.8 38.7 5.3 0.9 1.2
(Huntsville, Oh)
-------
to occur. Flow of the preheat fuel was stopped automatically at the same
time that feeding of coal was started. An automatic safety circuit would
shut down coal injection if the combustor temperature dropped too low to
ensure combustion of the coal or if the feeder-to-combustor AP dropped below
a pre-set minimum (about 6.9 kPa).
The weight of the coal feeder vs. time was taken so that the feed rate
of coal could be determined. Another method of estimating the feed rate was
to observe the oxygen concentration in the off-gas from the combustor. A
rapid rise in oxygen concentration was usually the quickest way of determining
that a problem was developing with the coal feeding system. Combustor tem-
perature could be controlled by regulating the amount of coal burned. The
feed rate of coal could be adjusted by changing the flow of injection air or
coal feeder-to-combustor AP.
To shut down the combustor routinely, the coal feed valve was closed,
fluidizing air was stopped, and nitrogen purge was started to preserve the
solids. Flow of injection air was kept on for several minutes so that the
coal feed line could be cleared of coal. All water flows were gradually
reduced. Solids could be discharged from the reactor by blowing them out
of a port located just above the fluidizing grid after the combustor had
cooled.
Batch Combustor Performance
A detailed discussion of the development of the primary components of
the batch combustor was presented in a previously issued report (2).
Included were discussions of the development of the modified Petrocarb
coal feed system, cooling coil design, and sampling system. These
components continued to perform satisfactorily and no additional modifica^
tions were required. Additional modifications were made to other parts of
the unit to improve performance. These are described below.
Bed Preheat System—
The preheat system was modified so that propane could be injected
directly into the bed at a position approximately 15 cm above the fluidizing
grid to supplement the main propane supply during preheat. The auxiliary
propane is injected using a "sonic" air jet type nozzle to promote better
mixing of the propane in the bed. The use of the air jet nozzle was found
necessary to prevent the propane from burning above the bed. The procedure
for preheat consisted of using the main propane supply to bring the bed tem-
perature to approximately 500°C whereupon the auxiliary fuel was injected.
When the temperature started to rise rapidly, the main fuel supply was
decreased to provide sufficient air for combustion of the auxiliary propane.
This new preheat technique allowed the bed to be preheated with flow to
all three cooling coils. This prevented the coils from experiencing high
metal temperatures and was intended to help prevent possible high tempera-
ture corrosion. This preheat technique also placed less stress on the
fluidizing grid.
The preheat system was later modified to use natural gas rather than
propane.
86
-------
Bed Agglomeration - Grid Design—
A problem which occurred periodically during batch combustor operation
was that of bed agglomeration. Bed agglomeration normally occurs when the
ash in the coal begins to soften and stone particles begin to adhere. Con-
ditions during which agglomeration usually occurred resulted from either
poor solids mixing and heat removal or sudden surges of coal into the com-
bustor. To promote better solids mixing a new fluidizing grid was designed.
The new grid had 21-0.3 cm diameter air distribution holes as compared to
80-0.16 cm diameter holes in the old grid. After the installation of the
new grid, bed agglomeration was not a problem.
In order to provide better coal feed control, the location of the pres-
sure tap for the low pressure side of the coal feed AP controller was moved
from a position within the bed to the freeboard region. The purpose of this
change was to insure that this tap would not plug and thereby interfere with
the AP .control.
Batch Combustor Results
During the course of the batch unit variables study, the following
coal/sorbent combinations were tested:
- Eastern (Arkwright) coal/Grove limestone
- Eastern (Arkwright) coal/Tymochtee dolomite
- Illinois #6 coal/Grove limestone
- Western coal/Grove limestone
A summary of the run conditions is given in Appendix Table M. Results
are given in Appendix Tables N and 0.
S02 Emissions—
862 emissions measured during batch unit runs varied with time, increas-
ing as the bed became more sulfated. Runs were made to various S02 concentra-
tions in the flue gas, the bed analyzed for S0,=, and the equivalent calcium
to sulfur molar ratio calculated as
Ca/S = fraction SC^ removed/fraction of calcium sulfated
The emission at the end of the run was then plotted against the equivalent
calcium to sulfur ratio. It had been planned to calculate equivalent calcium
to sulfur ratios for periods before the end of a run by calculating a sulfur
mass balance and using the final sulfation level as an anchor point. However,
this approach was not workable since a portion of the bed was lost by attri-
tion and entrainment during the run and corrections for bed loss could not be
made accurately.
S02 emissions are given in Appendix Table N. They are shown as a func-
tion of the equivalent calcium to sulfur ratio for runs burning Arkwright
(Eastern) coal and using limestone and dolomite as the S0£ sorbent in
Figures 111-32 and 111-33. Figures 111-34 and 111-35 show the same data
expressed as percent S02 retention vs. the calcium to sulfur ratio. Data
were obtained at a pressure of 800 kPa, with the exception of two runs
made at 310 and 395 kPa, vs. the calcium to sulfur ratio. Data were obtained
at a pressure of 800 kPa, at superficial velocities of 0.9 to 1.9 m/s, over
a temperature range of 750-975°C. As shown in the figures, the data
87
-------
FIGURE 111-32
Q_
Q.
CO
co
oo
UJ
o
CO
S09 EMISSIONS VS Ca/S RATIO - EASTERN COAL AND
LIMESTONE No. 1359
1600
1400-
1200-
1000-
2 800-
600-
400-
200-
1234
CALCIUM TO SULFUR MOLAR RATIO
-------
1600
FIGURE 111-33
S02 EMISSIONS VS Ca/S RATIO-EASTERN
COAL AND TYMOCHTEE DOLOMITE
T
T
a.
CL
C/)
1400
1200
1000
2 800
tn
to
LU
CM
O
600
400
200
1
1
1
0
1234
CALCIUM TO SULFUR MOLAR RATIO
89
-------
FIGURE 111-34
100
S02 RETENTION VS Ca/S - EASTERN COAL
LIMESTONE No. 1359
1
I
T
80
60
LU
I—
LJ
C\l
O
CO
40
20
0
CALCIUM TO SULFUR MOLAR RATIO
-------
100
FIGURE 111-35
S02 RETENTION VS Ca/S RATIO - EASTERN COAL
TYMOCHTEE DOLOMITE
I
T
1
80-
60-
LLJ
I-
LU
a:
CM
O
CO
40
20
0
1
0
2 3
CALCIUM TO SULFUR MOLAR RATIO
-------
obtained exhibited considerable scatter. As a result of this data scatter,
difficulty in assessing the effect of temperature and excess air was
encountered. Part of the scatter can probably be attributed to the fact that
the batch unit does not operate in a steady state fashion. Operation of the
unit is characterized by constantly changing SC>2 emissions and bed sulfation
levels, as well as decreasing bed depths and particle sizes due to attrition
and entrainment.
Two other factors which have been shown to affect 862 retention with lime-
stone sorbent are the differences in the extent of calcination and gas phase
residence time. This was noted in previous sections describing miniplant S02
emissions results (p. 36 and 40). A significant improvement in the S02 vs Ca/S
correlation occurred when these factors were taken into account. The effect of
variable calcination levels was accounted for by using an empirically modified
form of the Ca/S ratio which considers differences in the degree of calcination.
To do this, an effective Ca/S ratio was calculated as follows:
Effect. Ca/S Ratio = (Ca/S).X(Ca0 +
where X, , is the mole fraction of the calcium present as the oxide
or sulfate.
This approach is similar to that applied to the miniplant data (p. 52).
Use of this effective Ca/S ratio decreased the extent of data scatter as
shown in Figure 111-36. However, this approach is presented only as a means
of explaining the relatively poor reproducibility when limestone sorbent is
used. It cannot be used to predict desulfurization results with limestone
sorbent, since the degree of calcination cannot as yet be predicted. Pre-
dictions of desulfurization results with limestone sorbent must be based on
correlations such as those given in Figures 111-18 or 111-34, which are not
corrected for variable calcination levels.
Correcting for variations in residence time has also produced less data
scatter in batch unit data. The corrections were based on the following
assumptions:
- a first order sulfation reaction
- complete mixing of bed solids
- plug flow of combustion gases through the combustor
Figure 111-37 shows the limestone retention results corrected to the
average retention time of 0.64 seconds while the dolomite S02 retention
results for an average retention time of 0.76 seconds are shown in Figure
111-38. As seen, the extent of data scatter was reduced for limestone.
A comparison of Figures 111-37 and 111-38 also indicates that, as had been
reported previously (p. 57), dolomite is a more effective sorbent than lime-
stone at an equivalent Ca/S ratio.
Data obtained from the Illinois #6 coal/limestone test series are pre-
sented in Figure 111-39. The effect of temperature is more visible during
these runs than during the earlier limestone runs burning Arkwright (Eastern)
coal. However, for reasons still unknown, S02 retention was lower than anti-
cipated based on data obtained during the Arkwright coal/limestone test
series.
92
-------
100
FIGURE 111-36
S0? RETENTION VS EFFECTIVE Ca/S RATIO
LIMESTONE No. 1359
80
60
VD
10
LU
1-
UJ
CM
O
CO
40
20
0
1
0
234
CALCIUM TO SULFUR MOLAR RATIO
-------
100
FIGURE 111-37
S02 RETENTION AT CONSTANT RESIDENCE TIME
LIMESTONE No. 1359
I
I
80
60
UJ
UJ
CSJ
o
CO
40
20
Residence Time - 0.64 sec,
0
0
1234
CALCIUM TO SULFUR MOLAR RATIO
-------
100
FIGURE 111-38
SCU RETENTION AT CONSTANT RESIDENCE TIME
TYMOCHTEE DOLOMITE
T
80-
60
CM
o
CO
40
20
Residence Time - 0.76 sec,
I
I
0
1234
CALCIUM TO SULFUR MOLAR RATIO
-------
FIGURE 111-39
100
S0? RETENTION VS Ca/S RATIO - ILLINOIS No. 6 COAL
LIMESTONE No. 1359
K-
LU
0
2
CALCIU
3 4
TO SULFUR MOLAR RATIO
-------
A total of 5 runs were also made in the batch unit burning a low sul-
fur (0.6%) Western coal and using Grove limestone as the SC^ sorbent. The
same bed material was used in each of the runs and a total of 106 kg of coal
was burned. Although all of the runs were made at or below the limestone
calcination temperature, the measured 802 levels were nevertheless extremely
low. A possible explanation is that the S02 might have been reacting with
the CaO contained in the ash. The calcium content of the coal was calculated
to be equivalent to a Ca/S ratio of 1.5. To determine a baseline 862 emis-
sion level for the Western coal, a run was made using an inert bed. An SC>2
reduction of ~50% was observed during this run which further indicates that
sulfur might be retained by the ash. The S02 level measured during this run
is consistent with those reported by others (9) when burning a high calcium
content lignite. These results are shown in Table 111-25.
TABLE 111-25. DESULFURIZATION OF WESTERN COAL
Run Temp.
No. Sorbent (°C) % S02 Retention
102 Limestone 840 100
103 Limestone 825 100
105 Limestone 850 88
106 Limestone 850 93
107 Limestone 845 94
104 Alundum 835 46
SO^ Emissions—
SO-j levels were determined during several runs using the method described
previously on p. 61. Samples for the wet chemical analysis were extracted
from the ducting just downstream of the combustor pressure control valve.
The measured 803 levels were higher than expected and additional measurements
must be made before any conclusions can be drawn.
NOX Emissions—
A regression analysis of NOX emission data was made and the results
indicate that most of the variation could be explained by changes in the per-
cent excess air. Figure 111-40 shows NOX emission measured during batch unit
runs (for all coals except the Western coal) plotted as a function of percent
excess air. Although there is considerable scatter, the NOX emissions seem
to tend to level out as percent excess air increases. However, when data
from runs using Western coal are included, the data seem to fit a straight
line as seen in Figure 111-41. It is not certain whether the lack of curva-
ture is really due to a difference in the emissions from the Western coal or
if the line should be straight in Figure 111-40. Unfortunately no data were
obtained at the higher excess air levels for coals other than Western.
Data from all runs are also given in Appendix Table N. The data were
obtained at a pressure of 800 kPa and at temperatures generally in the range
of 800 to 975°C.
97
-------
FIGURE
VO
oo
1.0
0.8
0.6
CO
QX 0.4
0.2-
0
20
BATCH UNIT NOY EMISSIONS
J\
o
o
o
o
o
Coal
Eastern
Imois -
Sorbent
• Limestone
O Dolomite
A Limestone
40
60 80
EXCESS AIR f°-
100
120
140
160
-------
1.0
0.8-
V£>
3 0.6 -
h-
03
0.4-
0.2
0
FIGURE 111-41
BATCH UNIT IMOX EMISSIONS INCLUDING WESTERN COAL RESULTS
o
o
o
T
o
o
o
Coal
Eastern
Illinois -
Western -
Sorbent
O Dolomite
• Limestone
A Limestone
A Limestone
20 40
60 80 100
EXCESS AIR (%)
120 140 160
-------
One important observation noted was that as the SC>2 concentration in
the flue gas increased as a run progressed, the NOX emissions decreased.
This decrease in the NOX emission may have occurred because of a reaction
between SC>2 and NO. This fact, coupled with high excess air levels, may
explain why the NOX emissions were very high in the runs burning low sulfur
Western coal since very low S02 emissions were measured in these runs.
At excess air levels of 15-20%, the range anticipated for commercial FBC
units, the emissions were generally in the range of 0.2 to 0.3 Ib NO£/M BTU.
These levels are well below the EPA emission standard of 0.7 Ib N02/M BTU.
However, as seen in Figures III—40 and 111-41, NOX levels were much higher at
higher excess air levels, reaching 0.6 Ib/M BTU at 100% excess air, and 0.9
Ib/M BTU at 150% excess air with Western coal.
The NOX produced by the batch unit combustor was predominantly NO. Less
than 5% was present as N02 and this was probably formed in the sampling system,
since the equilibrium concentration of N02 is very low at the high tempera-
tures occurring in the combustor.
CO Emissions—
The batch combustor CO level has been found to be dependent on the excess
air level, the average bed temperature, and the steadiness of the coal feed
rate. The excess air level appears to be a significant variable, especially
at lower levels. Temperature appears to become significant only at very low
levels (800°C). Prior to modifications in the coal feed system, CO emissions
were generally quite high. The high levels were attributed to the unsteady
coal feed rate which resulted in poor coal combustion. CO emissions averaged
960 ppm during these runs. Modifications to the coal feed system were made
after Run 46C and after that the CO emissions averaged 180 ppm for temper-
atures above 800°C. At 750°C, a CO emission of 3600 ppm was measured. The
effect of the steadiness of the coal feed rate will be discussed in more detail
in the following section. CO data are given in Appendix Table 0.
Combustion Efficiency—
Batch unit combustion efficiencies normally varied from 87 to 99% and
were generally in the mid-90's. Combustion efficiency as a function of
excess air and temperature is shown in Figure 111-42. Data are also given
in Appendix Table N. In one run in which both temperature and excess air
were extremely low (620°C and 0%, respectively), a low combustion efficiency
of 60% was calculated.
Another parameter which affected combustion efficiency was the steadi-
ness of coal feeding. When coal feeding was unsteady, combustion efficien-
cies were consistently lower. Figure 111-43 shows the combustion efficiency
as a function of excess air for runs made both before and after modifica-
tions were made to the coal feed system. These modifications which involved
the redesign of the orifice assembly on the coal feed vessel resulted in a
significant improvement in the combustion efficiency. Over the entire range
of excess air levels, the combustion efficiency increased between 3 and 4%.
Batch unit combustion efficiencies are now more consistent with those
reported from miniplant runs. A more detailed comparison of results from
the two units is made in Section V.
100
-------
FIGURE 111-42
COMBUSTION EFFICIENCY VS EXCESS AIR AND TEMPERATURE
100-
LU
95
o
•z.
LU
rr 90
en
=>
CQ
2
O
O
85
80
O
D
20
n
Pressure: 800 kPa
Super. Vel.: 1.0-1.5 m/s
Temperature:
n <820°C
• 850 - 870°C
O >900°C
_L
1
40 60 80
EXCESS AIR (%)
100
120
-------
O
ro
O
-z.
UJ
O
o
i-
co
=>
DQ
^
O
O
100-
85-
0
FIGURE 111-43
EFFECT OF COAL FEEDING ON COMBUSTION EFFICIENCY
20
T
I
Pressure: 800 k Pa
Superficial Vel 1.0 - 1.5 m/s
Temperature: 850-870°C
• Old Orifice
O New Orifice
40 60
EXCESS AIR (%)
80
100
120
-------
Particulate Emissions—
Particulate Loadings—Table 111-26 presents a summary of the particulate
loadings measured from batch unit runs 1C-107C. Rejecting data which were
greater than the mean by more than twice the error limit gave the following
results:
TABLE 111-26. PARTICULATE LOADINGS. BATCH UNIT
Grain Loading (gr/scf)
Bed Combustor
Material Outlet Primary Cyclone Outlet Secondary Cyclone Outlet
Grove 6.8+1.4 0.9+0.5 0.6+0.5
Limestone — — —
Tymochtee 7.0+i.o 0.9+0.5 0.4+0.3
Dolomite — ~~ ~
Alundum 5.3+0.4 1.0+0.5 0.3+0.3
The data presented above show that the combustor outlet grain loadings
were approximately the same for the limestone and dolomite runs. However,
this is believed due primarily to differences in the operating conditions used
during these runs such as superficial velocity, time of run, etc.
The overhead solids were routinely analyzed for total carbon, total
sulfur, sulfate, and carbonate. Appendix Table P presents a summary of
these analyses. Using these results, it was calculated that ^20% of the
overhead solids was elutriated bed material for the runs using Grove lime-
stone and ^40% when Tymochtee dolomite was used, clearly indicating substan-
tially higher attrition and entrainment rates for dolomite.
Particle Size Distribution—Overhead samples from selected runs were
dry sieved to determine the particle size distribution. The results are
presented in Appendix Table Q. The particle sizes for the 50% points were
approximately 65 microns and 10-20 microns for the material collected by the
primary cyclone and filter, respectively. These sizes are somewhat smaller
than those previously reported.
Cyclone Efficiency—The average collection efficiency based on the
average grain loadings calculated in the previous section was 86% for the
primary cyclone and 46% for the secondary cyclone, giving an overall com-
bined efficiency of 92%.
Components Balances—
Sulfur and calcium balances were made for a number of batch unit runs.
The results are presented in Appendix Tables R and S. The S02 balances
averaged 99.6% with a standard deviation of 20% and the calcium balances
averaged 87% with a standard deviation of 10%. The weight of sulfur into
the combustor was determined from the coal feed rate, the run length, and
the sulfur content of the coal. The calcium input was determined from an
initial bed analysis. For the solids output streams (material collected in
the cyclone diplegs, off-gas filter, and the final bed), the quantity accumu-
lated was weighed and then analyzed for sulfur, sulfate, and calcium. The
quantity of sulfur in the flue gas was determined from data obtained by
continuously monitoring the off-gas for S02«
103
-------
SECTION IV
MINIPLANT REGENERATOR SHAKEDOWN
The shakedown of the miniplant regenerator system was carried out in two
phases. The initial phase consisted in a series of shakedown runs made
while operating the regenerator in a batch fashion, uncoupled from the
combustor. The objectives of this phase were to test and develop equipment,
gain operating experience and collect some data relating the S02 level in
the regenerator off gas and sulfated sorbent conversion to the operating
conditions. The second phase was aimed at demonstrating the continuous
operation of the regenerator coupled to the combustor. The objectives of
the second phase were to develop equipment required to move sorbent con-
tinuously between the combustor and regenerator, and again, to gain opera-
ting experience and collect data. This phase was to be concluded by the
completion of a 24 hour continuous run.
This section describes the regenerator and the batch and continuous
shakedown operations.
EQUIPMENT
Regenerator Vessel
The regenerator reactor, designed for operation at 1100°C and pressures
up to 1010 kPa, consists of a 45.7 cm I.D. steel shell refractory-lined with
75-28 Grefco Litecast to an internal diameter of 21.6 cm. Numerous taps are
provided along its 6.66 m overall height to monitor both temperature and
pressure, while appropriately located ports allow for material entry and
exit.
Fuel System
There are two separate fuel systems: burner and supplementary fuel.
Burner fuel is supplied to the burner where it is mixed with an approximately
stoichiometric amount of burner air and burned. Supplementary fuel is
added directly to the regenerator column just above the fluidizing grid.
Sufficient supplementary fuel is added to produce reducing gases (CO, T^)
at the desired concentrations. Location of the fuel (and air) inlets are
shown in Figure IV-1.
Natural gas is compressed to about 180 psig by a Corken Model 590 com-
pressor with a capacity of about 1.1 s m3/min (40 SCFM). Burner and sup-
plementary fuel flow through separate lines, each containing a measuring
orifice, automatic flow control valve, and an automatic shutoff valve.
Supplementary fuel enters the regenerator just above the fluidizing grid
through a 0.64 cm (1/4 inch) O.D. stainless steel tube. A small flow of
nitrogen is added to the supplementary fuel line when fuel flow is shut off
in order to prevent bed solids from plugging the entry tube. Safety devices
prevent addition of supplementary fuel into the regenerator until temperature
is high enough for the fuel to burn satisfactorily, about 650°C.
104
-------
BURNER AIR
BED LEVEL
SUPPLEMENTARY AIR
SUPPLEMENTARY FUEL
FLUIDIZING GRID
BURNER
BURNER FUEL
FIGURE IV-I
MINIPLANT REGENERATOR AIR AND
FUEL LOCATIONS
105
-------
The ratio of flow rates of supplementary to burner fuel depends on the
concentration of reducing gases desired; generally this ratio lies between
0.2 and 0.5.
Air System
Two separate air systems can be identified on the regenerator: burner
air and supplementary air. Burner air is supplied in sufficient quantity
to burn completely the fuel (natural gas) supplied to the burner. Supple-
mentary air is added directly to the bed in order to create an oxidizing
zone in the upper portion of the bed.
The source of air for the regenerator is the main miniplant air com-
pressor. Burner air can be made to flow through either of two measuring
orifices, depending on the flow rates desired. Air is passed through a flow
control valve tied to an automatic control loop and then enters the burner.
Combustion products from the burner fluidize solids in the regenerator.
Supplementary air is also piped from the main air compressor through a.
measuring orifice and automatic flow control valve. It enters the regen-
erator through a 1.3 cm (1/2 inch) O.D. stainless steel tube whose outlet
is positioned at the inside wall. Flow rates of supplementary air are
typically about 20 percent of the burner air flow rate, but this can vary
considerably depending on the air/fuel ratios desired in the oxidizing and
reducing zones.
Off-Gas Handling
Figure IV-2 shows the off-gas handling system for the regenerator. Gases
leaving the regenerator are pressurized and hot (typically 900 kPa and about
930°C). The gas is cooled and pressure reduced before discharge into a wet
scrubber. Principal components of the off-gas system are the cyclone, off-
gas cooler, filter, pressure-control valve, and scrubber.
The regenerator cyclone removes the bulk of particulates, primarily
entrained fines originating in the bed, from the gas stream. Gas inlet
velocity to the cyclone is about 20 m/s, making it a moderately efficient
cyclone. Off-gas then enters a cooler, which is a single pass double pipe
heat exchanger 6.1 m long. The inner pipe, through which gas flows, is 3.8
cm (1-1/2 inch) Schedule 80 Type 316 stainless steel. Surface area for heat
transfer on the inside of this pipe is 0.73 m2. The outer pipe, through
which water flows (either co-current or countercurrent) is 7.6 cm (3 inch)
Schedule 40 steel pipe. Gases leave the cooler at 150-200°C, depending on
the gas flow rate.
Before reducing pressure, dust is removed by passing the off-gas through
a sintered stainless steel bayonet-type filter with a surface area of 0.14 m .
Filtration is necessary to remove fine particulates which, if present in high
concentrations, would erode the pressure reducing valve. This valve, a
Norriseal Model 510 air-to-close with a Cv = 6 orifice, is part of an automatic
pressure control loop. Pressure is sensed at the top of the regenerator
column with a Viatran pressure transmitter. Position of the pressure reducing
valve is controlled by a Taylor electronic controller operating a Fairchild
I/P converter.
106
-------
CYCLONE
HOT OFF-GAS
FROM
REGENERATOR
SOLIDS TO
LOCKHOPPER
c.w.
C.W. OUT
TO SCRUBBER
OFF-GAS
COOLER
-M-
—Cxi-
PRESSURE
CONTROL
VALVE
FILTER
ORIFICE (TO INCREASE
BACK PRESSURE)
SAMPLE TO
ANALYZERS
FIGURE IV-2
MINIPLANT REGENERATOR OFF-GAS HANDLING SYSTEM
-------
Downstream of the pressure control valve, pressure of the off-gas is
only slightly above atmospheric, and temperature is under 150°C. Regenerator
off-gas is combined with off-gas from the combustor and piped to a Research
Cottrell Scrubber for cleanup before venting. Sodium carbonate (Na2C03)
solution is added to the scrubber and reacts with the SC»2 to produce sodium
bisulfite, NaHSC>3. An excess of Na2C03 is used to keep the pH of the liquid
in the scrubber in the range of 7-9. Scrubber liquid is slowly discharged
into a diked area south of the miniplant. Calcium chloride (CaCl2> solution
is added to the liquid as it enters the diked area. Sulfite is precipitated
as CaS03'2H20 and excess Na2C03 is precipitated as CaCC^, thus keeping the
pH of liquid in the diked area near neutral.
Gas Sampling System
The system used to convey a sample of regenerator off-gas to analytical
instruments is shown in Figure IV-3. The gas sample is obtained downstream
of the pressure reducing valve so that the pressure of the sample gas is
only slightly above atmospheric. A baffle located in the off-gas piping
downstream of the sampling point serves to maintain the pressure of the
sample stream high enough so that there is sufficient flow to the analyzers.
Particulates are removed from the sample gas with a Balston Model 33 filter
fitted with an "H" type filter tube. The clean gas then enters a Perma-
Pure Model PD-1000-24S self-regenerative membrane type dryer. After leaving
the dryer, the sample passes through Teflon lined tubing to the analyzer
manifold. In order to prevent condensation of water, the sample line up to
the dryer is maintained at a temperature of about 150°C.
A sampling point downstream of the pressure reducing valve was selected
so that the sample could be obtained at low pressure. Sampling at low pres-
sure reduces the residence time of gas in the lines, minimizing the possi-
bility of gas composition changing before entering the analyzers. The need
for a pressure regulator, which can corrode and change the sample composi-
tion, is also eliminated.
Fluidizing Grid
The regenerator fluidizing grid, shown in Figure IV-4 is a stainless
steel plate 13 mm thick into which are drilled 89 holes of 3.6 mm diameter.
Because the grid is located directly above the burner, the grid must be
water-cooled. Fourteen water channels, 4.8 mm in diameter, are located
between the rows of holes to accomplish cooling. The size of 3.6 mm for the
grid holes was chosen because this was about the largest size hole through
which particles of bed material would not pass when the unit was shut down.
The number of holes, which determines the total area available for flow of
fluidizing gas, was chosen such that the pressure drop across the grid was
about 25 percent of the bed pressure drop. This should result in uniform
flow of fluidizing gas across the face of the grid.
Burner
The burner used in the regenerator is identical to that used in the
miniplant combustor. The burner was described in a previous report (2).
108
-------
TO SCRUBBER
ORIFICE
TEFLON LINED HOSE
FILTERED, DRIED SAMPLE
TO ANALYZERS
DRY N.
IN
DRYER (PERMAPURE MODEL PD-] 000-245)
ELECTRICALLY HEATED
FILTER
(BALSTON MODEL 33)
REGENERATOR
OFFGAS ( 1 ATM)
FIGURE IV-3
MINIPLANT REGENERATOR OFF-GAS SAMPLING SYSTEM
-------
22CM REGENERATOR DIAMETER
89 HOLES
3.6MM DIAM
29CM SQUARE INSERT
COOLING
WATER
OUTLETS
COOLI NG
WATER INLETS
1.1CM I.D.
53 CM DIAMETER 304 STAINLESS
STEEL PLATE, 1.3CM THICK
COOLING
CHANNELS (15 TOTAL)
4.8MM I.D.
7.9MM DIAM.
THERMOWELL
FIGURE IV-4
REGENERATOR FLUIDIZING GRID
-------
MATERIALS
Materials of Construction
Non-Metallic Materials —
Refractory Lining — The regenerator is constructed of 45.7 cm (18 inch)
Schedule 40 steel pipe, refractory lined to an inside diameter of 21.6 cm.
Thickness of the refractory insulation is about 10.6 cm. The refractory
used is a castable type, General Refractories Litecast 7528. This material
has a service limit of 1540°C (2800°F) and a bulk density of about 1.2 g/cm3
(75 Ib/ft3). Thermal conductivity at 540°C is 0.55 W/m^°C (3.8 BTU/hr ft2
Wear of the refractory lining, due to erosion caused by the bed and
other factors, has not been estimated because of the relatively short time
that the regenerator has been operated (less than 100 hours). No major
obvious damage, such as cracking or spall ing, has been noted as of July, 1976.
The section of the regenerator below the fluidizing grid (burner zone)
is lined with General Refractories Bubbalite, which is also a castable, but
with a temperature service limit of 1815°C.
Thermocouple Protection Tubes — The stainless steel sheathed type K
(chromel-alumel) thermocouples used in the regenerator would fail quickly
at the prevailing conditions of temperature and gas composition and therefore
are protected in the following manner: The thermocouple is inserted into a
6.4 mm O.D. inconel tube which in turn is placed inside a silicon carbide
sheath with an I.D. of 6.4 mm and an O.D. of 12.7 mm. The silicon carbide
sheath provides excellent resistance to high temperatures and chemical
attack and the inconel tube gives the assembly additional strength. Per-
formance of this system has been fairly good although several silicon carbide
tubes have cracked in service, probably from mechanical stresses.
Metallic Materials —
Off-Gas Cooler — The off-gas cooler accepts hot (1000°C) pressurized gases
from the regenerator and cools the gases to about 150-200°C. Process gases
flow through a water-jacketed type 316 stainless steel pipe. The cooler is
designed so that no condensation occurs during steady state operation;
however, condensation cannot be prevented during startup and shutdown. Type
316 stainless steel offers good resistance to dry and wet S02 although wet
H2S04 can cause serious corrosion even at fairly low concentrations. Regular
inspections of the cooler are made to check for the presence of corrosion.
Thus far, with under 100 hours of actual operation, no corrosion has been
detected.
Filter — A sintered stainless steel filter (bayonet type) is used to
remove particulates from the off-gas before the gas passes through the pres-
sure let-down valve. The filter element is made of Type 316 L stainless
steel. The filter is sometimes wet, even under steady-state running condi-
tions, and corrosion has been a problem. Corrosion has been minimized by
thoroughly washing and drying the element immediately after the conclusion
of each run. Allowing liquid to remain on the filter during downtime between
runs resulted in very rapid corrosion. The off -gas cooler has recently been
modified to provide higher temperatures in the filter, thus reducing condensa-
tion and corrosion problems .
Ill
-------
The filter housing is made of Type 304 stainless steel. It is con-
structed of 10.2 cm (4 inch) pipe and is much more rugged than the com-
mercially made housing that it replaced.
Gas Sampling System—The gas sampling system is shown in Figure IV-3.
The sample is conveyed through a short length of 6.4 mm (0.25 inch) tubing
into a Balston Model 33 filter. Both the tubing and filter are made of Type
304 stainless steel and both are heated to about 150°C to prevent condensa-
tion. Downstream of the filter the gas is dried and conveyed to analyzers
through Teflon lined tubing.
Bed Material
All runs made in the regenerator have used sulfated stone prepared in
the miniplant combustor. Three types of stone have been used: Grove
limestone (No. 1359), Pfizer dolomite (No. 1337) and Tymochtee dolomite.
Levels of sulfate in the bed charged to the regenerator varied from 15 to
38 weight percent.
BATCH OPERATION
Results
Eleven runs were made in which batch charges of sulfated limestone were
regenerated. The purpose of these runs was to provide operating experience,
test and develop equipment, and collect data. After the batch runs, a system
to continuously transfer solids between combustor and regenerator was devel-
oped, installed, and tested, and continuous runs were carried out. Con-
tinuous runs are described in the section on Performance of Transfer System -
24 Hour Shakedown Run. This section describes the series of "batch" runs.
Equipment Development
Supplementary Air and Fuel Systems—
From operation of the 8.3 cm diameter ("batch") regenerator in 1972-3,
it was determined that improved performance could be obtained by adding
supplementary fuel and air directly into the bed, in the manner shown in
Figure IV-1 (1). The burner, located below the fluidizing grid, is operated
at an air/fuel ratio slightly above stoichiometric, and supplementary fuel
is added just above the grid, in a quantity sufficient to produce a reducing
atmosphere of the desired concentration (typically 65-85 percent of the
stoichiometric air/fuel ratio). Higher in the bed supplementary air is
added to bring the overall air/fuel ratio up to about stoichiometric.
Supplementary fuel (natural gas) compressed to about 1300 kPa flows
through a measuring orifice, an automatic flow control valve, and 1.3 cm
(0.5 inch) stainless steel tube to the regenerator. Fuel enters the bed
through a 4.6 mm I.D. stainless steel probe that ends just at the inside
wall of refractory. The probe was made a smaller diameter than the tubing
which supplies fuel to the regenerator so as to impart a higher velocity to
the fuel as it enters the bed. This provides better cooling for both fuel
and probe, preventing overheating of the probe or decomposition of fuel in
the probe.
112
-------
Supplementary air is taken from the accumulator of the main air compres-
sor. Flow control is obtained with an automatic valve which is part of a flow
control loop. Flow rate is measured with an orifice and delta pressure trans-
mitter. Air enters the regenerator through a 1.1 cm I.D stainless steel tube
which terminates just at the inside refractory wall. Overheating of this tube
has not been observed.
Only one position for the supplementary fuel inlet has been used; this
is a point 12.7 cm above the fluidizing grid. Three positions for the sup-
plementary air inlet have been tried: these have been at locations 58, 74,
and 122 cm above the fluidizing grid. Height of the fluidized bed has
typically been about 120-130 cm. Changing the locations of the supplementary
air inlet changes the relative size of oxidizing and reducing zones. This
might be expected to affect the amount of CaS produced in the bed and possibly
the concentration of S02 that is generated; however, these effects were not
observed. It should be noted that only one run was made with supplementary
air admitted at each of the 58 and 122 cm locations, most work was done at
74 cm. No attempt was made to optimize the location as this would have required
many more runs than time permitted. Moreover as the level of sulfide in the
regenerated stone was already very low, not much benefit could result in this
respect.
Off-Gas Cooler—
The original off-gas cooler was a shell and tube unit approximately
1.8m long which contained nineteen 15.9 mm I.D. stainless steel tubes. Gas
flow was through the tubes. Plugging of the tubes with solids was an
occasional problem and the unit was replaced with the double pipe exchanger
described on p. . Performance has been satisfactory thus far, except
for overcooling the off-gas under some operating conditions where the flow
rate of off-gas was fairly low.
Fluidizing Grid—
The fluidizing grid (air distributor) for the regenerator is a water-
cooled stainless steel plate containing drilled holes for passage of the
fluidizing gas. Cooling is essential because hot gases from the burner,
located below, flow upwards through the grid.
The area available for flow of gases through the grid was chosen consis-
tent with conventional practice, i.e., to produce a pressure drop equal to
about 25% of the bed pressure drop. Reasonably uniform flow of gas over the
face of the grid was therefore to be expected.
Two grids were used in the regenerator which differed primarily in the
size and number of holes, but not very greatly in the total area for flow.
The first grid, which was used for the first nine (out of eleven) runs made
in the regenerator, contained 392 holes of 1.98 mm diameter. The second
grid, shown in Figure IV-4, had 96 holes of 3.18 mm diameter. This was
about the largest size hole through which solids woul not flow when the unit
was shut down. The new grid was cheaper to make because fewer holes had
to be drilled; but, there was another possible advantage.
The type of grid can strongly influence the quality of fluidization and
it is generally believed that a grid with many gas inlet openings is superior
to one with few gas inlet openings because bubbles are smaller and gas-solid
contacting is more intimate with many small openings. However, another
113
-------
school of thought contends that gas inlet openings should be larger than a
minimum size. It is known that the penetration of a jet into a fluid medium
increases as the diameter of the jet. Hence, larger gas inlet openings
produce jets which penetrate further into the bed, stabilizing the bed and
reducing the possibility of gross gas channeling. The new grid was designed
with these considerations in mind.
Run no. 10 was the first using the new grid. During preheating, the
temperature profile in the bed was flatter than had been seen before. After
regeneration was started, the entire reducing zone of the bed, 74 cm high,
was at the same temperature. The oxidizing zone of the bed, usually 10 to
SO^C hotter than the reducing zone, was at the same temperature as the
reducing zone.
During run no. 11, also made with the new grid but at far more severe
reducing conditions, the temperature spread across the reducing zone was
only 7°C. The temperature of the oxidizing zone was 11°C higher than the
reducing zone. The unusual uniformity of temperature implies that the
quality of fluidization was improved.
The S02 concentrations for runs 10 and 11, with the new grid, were 35%
and 70% higher respectively, than the highest concentrations obtained pre-
viously (in run no. 8). It should be noted that these concentrations were
69% and 52%, respectively, of the concentrations that would be expected if
chemical equilibrium had been obtained.
Changes in operating conditions are not suspected of causing the
increase in S02 concentration. It appears that the change in grid design
was responsible.
Thermocouple Protection Tubes—
The earliest thermocouple protection tubes used were silicon carbide
sheaths which were slipped over the thermocouples. The thermocouples were
Type K and were enclosed in a 3.2 mm O.D. stainless steel sheath. Castable
refractory was then used to fill the entire carbide sheath in order to seal the
thermocouple into the sheath. Breakage of these tubes occurred frequently,
usually they were snapped in two.
Another approach was to flame spray protective coatings on the thermo-
couples. Chromium carbide was placed over a nickel aluminate substrate.
However, these thermocouples failed quickly.
The third approach was to use again a silicon carbide sheath (6.4 mm
I.D.) and to insert into the sheath an inconel tube (6.4 mm O.D.) to provide
reinforcement. The thermocouple was then placed in the inconel tube.
Although the silicon carbide sheaths have cracked occasionally, this method
was by far the most satisfactory to date.
Procedures
Startup—
Checkout of the regenerator and its support equipment is carried out
prior to startup. This includes cleaning off-gas and sample filters, checking
cooling water flows and setting manual valves in the air, fuel, and off-gas
systems to proper positions. The nitrogen compressor is turned on so that
114
-------
flow is directed to all ports where pressure is measured, preventing these
ports from plugging when bed is added to the regenerator column. About 50 kg
of sulfated stone is usually added to the column, forming a settled bed of
1.0-1.3 m. Air and natural gas compressors are started and air is very slowly
admitted to the column until pressure has reached about 200 kPa. The burner
is usually ignited at pressures between 200 and 400 kPa. Heating of the bed
is done slowly without increasing pressure, so that superficial gas velocity
remains as high as possible (but always under 2 m/s). Bed temperatures
during heatup are more uniform at higher velocities. When temperature
approaches the desired level, generally 1040-1120°C, pressure is increased
to 910 kPa (the pressure at which all regeneration runs were made). Some-
times small mounts of supplementary fuel are added to increase further bed
temperature, but not enough to produce reducing conditions. Supplementary
fuel can only be added when bed temperature is above 650°C, the temperature
at which natural gas will ignite.
Normal Operating Procedures—
When the bed temperature is uniform and close to the temperature desired
under reducing conditions, the switchover to reducing conditions is made.
This is accomplished by increasing flow of supplementary air to the required
value and then increasing the flow of supplementary fuel. Supplementary air
flow is always increased before supplementary fuel so that air is not added
to a column already filled with a reducing gas. Temperature is continuously
monitored and flow rates of burner air, burner fuel, supplementary air, and
supplementary fuel are all adjusted to yield the desired bed temperature (see
section on Temperature Control). Oxygen and CO concentrations in the off-gas
are also monitored and the supplementary air flow is corrected to produce low
concentrations of CO (under 5000 ppm). The flow rate of supplementary air is
the minimum value which will just produce CO in the oxidizing zone.
Shutdown—
Normal shutdown can be accomplished several ways. One method is to press
the "emergency stop" button, which shuts all flows of air and fuel, fills the
column with nitrogen and slowly depressurizes the column. This method is
preferable if the final composition of solids is important, as nitrogen plus
the rapid temperature drop "freezes" the composition of the bed.
Another method of shutdown is to turn off supplementary fuel and air
flows, in that order. Burner fuel is shut next. Column pressure and burner
air flow must be brought down together so as not to create high superficial
velocities, which would blow solids out of the column.
Emergency Procedures—
During operation, shutdown is automatic if bed temperature exceeds
1260°C, if column pressure exceeds 1140 kPa, if burner air flow is lost, or
if a burner flameout occurs. The operator can shut down the unit at any time
by depressing a red emergency stop button on the control panel. An extensive
system of alarms alerts the operator to out-of-limits conditions on about 20
variables.
Temperature Control—
The source of heat for regeneration is combustion of a fuel, natural gas.
Heat sinks are (1) heating air and fuel entering the regenerator to tempera-
ture, typically about 1100°C; (2) endothermic heat of reaction for converting
CaS04 to CaO, equivalent to about 1300 kJ/kg CaS04 converted; (3) sensible
115
-------
heat required to bring solids entering from combustor to regenerator tem-
perature (applies to a continuous system) and (4) heat losses. Temperature
of the fluidized bed is controlled by regulating the energy or fuel input.
As shown in Figure TV-1, air and fuel are each added to both the burner and
bed (supplementary air and fuel) in order to create adjacent oxidizing and
reducing zones. Since it is usually desired to maintain fixed air/fuel ratios
in these zones, it is necessary to adjust air flows whenever changing fuel
flows. Thus, in order to change temperature, it is necessary to change the
flow rates of all four streams: burner and supplementary air and burner and
supplementary fuel. In practice, this is easy to do since all four flow
rates can be changed automatically from the control panel.
It is important to maintain a well-controlled and uniform temperature
in the regenerator because of the danger of agglomerating solids should
localized temperatures get too hot (e.g., greater than 1150°C). Also,
because the equilibrium concentration of S02 increases sharply with tempera-
ture, it is necessary to operate at as high a temperature as possible in
order to obtain the highest SC>2 level.
Uniformity of bed temperature in the miniplant regenerator has generally
been quite good. Variation of temperature within the reducing zone, which
extends from the fluidizing grid to 74 cm above it, has been less than 10°C.
The oxidizing zone is typically 10-50°C hotter than the reducing zone. The
temperature difference between zones depends on the air/fuel ratios: with
low air/fuel ratios in the reducing zone, more combustion occurs in the
oxidizing zone, which results in higher temperatures. A temperature profile
for a run made with the second fluidizing grid (larger holes) is given in
Figure IV-5.
A problem exists in controlling temperature when the regenerator is run
in "batch" fashion, i.e., when a batch of sulfated stone is charged and then
regenerated. The problem arises because the heat required by the highly
endothermic regeneration reaction varies throughout the run as the rate of
reaction. Soon after reducing conditions are established the energy require-
ment rapidly increases and then gradually falls off as the regeneration rate
decreases. The magnitude of the problem is illustrated in the following
example. The energy balance can be represented as:
E = MCp(T - Ta) + Qreaction + Qlosses U>
where E = total energy input to regenerator
M = mass flow rate of combustion gases
Cp = heat capacity of combustion gases
T = regenerator temperature
Ta = temperature of entering air
Q . = endothermic heat of reaction
^reaction . -
Q-. = heat losses
losses
The first term on the right hand side of equation (1) represents energy
required to heat gases to regeneration temperature. In a typical run,
Qreaction/MC (T-Ta) = 0.40, i.e., the average energy requirement for the
regeneration reaction is about 40 percent of the energy required to heat
116
-------
FIGURE IV-5
TYPICAL TEMPERATURE PROFILE IN MIN1PLANT REGENERATOR BED
o
o
uT
o:
n>
I—
ce:
LU
Q_
LU
I—
.1 1£»
1 1 1 1 1 1 1 1 1 1 1 1 1 1 Vv 1
•A
0 20 40 60 80100120140160180200220240260280300 600
HEIGHT ABOVE GRID, cm.
-------
gases entering the regenerator to temperature. If regeneration should cease
and the energy input remains constant, the temperature will increase to T1
according to:
1.40 MC (T - Ta) + Q,ooo - MC (T' - Ta) + Q,
P losses p ' xlosses
or T' = 1.40T - 0.40Ta
For example, if T = 1100°C and Ta = 65°C, then T1 = 1514°C. Therefore, bed
temperature can increase over 400°C if energy input remained constant when
regeneration ceased, as at the end of the run.
Practical consequences of the above problem are that the fuel input to
the regenerator must be frequently adjusted as the regeneration rate peaks
and then tapers off. In order to maintain constant bed temperature, total
input of fuel to the regenerator had to be decreased by about 40 percent over
the'course of a typical batch run.
In continuous, steady, operation, where solids are transferred to and
from the regenerator at constant rates, the above problem does not exist
because the heat load on the regenerator does not vary. Regeneration proceeds
at a fixed rate unless some external factor causes a change.
Bed Agglomeration—
Agglomeration of bed material was invariably associated with excessively
high temperatures; when temperature was well controlled agglomeration did not
occur. During the first few batch runs control problems caused air and fuel
flow rates to be erratic, at times, and control of temperature was made
especially difficult. Agglomeration did occur in moderate amounts, usually
towards the end of the runs, when bed temperature would tend to increase
because of a fall off in regeneration rates.
During a run at 1120°C, which was the highest temperature attempted,
problems developed in controlling pressure and flow rates of air and fuel
became unsteady, causing temperature at the supplementary fuel inlet to
rapidly increase to about 1350°C even though the flow of fuel was shut down
soon after the rise in temperature began. Examination of the bed revealed
some agglomeration but the limited amount suggested that high temperatures
were localized. Runs have been made at average bed temperatures(reducing
zone) up to 1092°C without any agglomeration.
It has been observed that when localized high bed temperatures cause
agglomeration to begin it is difficult to reverse the process and bring
temperature back under control. Whenever agglomeration occurs, temperatures
at different points in the bed quickly diverge and the portion of the bed
that is already too hot becomes even hotter. This is not surprising because
of the drop in the rate of heat transfer away from the agglomerated portion
of the bed. Hence, agglomeration tends to be "autocatalytic;" once initiated,
conditions become more favorable for further agglomeration.
Bed agglomeration appears to occur at temperatures considerably below
the melting points of either pure CaSO^ or CaO. It is known that the pre-
sence of even small amounts of coal ash reduce the temperature at which
agglomeration has been observed to occur. This had been noted in early work
118
-------
carried out with pure CaSO^ in the 8.3 cm diameter regenerator. Agglomera-
tion occurred in runs made with sulfated limestone prepared in a coal com-
bustor, but not in runs made with CaSO^ which had not been exposed to coal
(1).
S02 Levels—
The maximum S02 concentration that can be produced at a given tempera-
ture and pressure is the equilibrium concentration of S02 for the reaction:
3CaSO,
CaS = 4CaO + 4SO,
This concentration is inversely proportional to total pressures and increases
strongly with temperature. Table EHL gives partial pressures of S02 at_
several temperatures and the corresponding mole fractions at 910 kPa (nine
atmospheres), the pressure at which runs were conducted.
TABLE IV-1.
MAXIMUM PARTIAL PRESSURES OF SO FOR REDUCTION OF CaSO,
Temperature °C
1000
S02 Partial
Pressure, kPa
21
S02 Mole Fraction at
910 kPa Total Pressure
0.023
1025
24
0.026
1050
27
0.030
1075
35
0.038
1100
1125
1150
52
73
106
0.057
0.080
0.116
S02 levels from regenerator runs made with batches of sulfated stone have
been reported as average and peak values. Average values are somewhat
ambiguous because they are influenced by the time chosen to end the run.
Typically, S02 concentration increases rapidly at the outset, peaks, and then
gradually tails off. The longer a run continues, the lower will be the aver-
age concentration of S02- Usually, runs were terminated when S02 levels
declined to about ten percent of their peak values.
Because the off-gas is dried before analysis, measured concentrations
of S0£ and other components must be corrected. This correction is made as
follows:
[S09] = (1 - [HLO])
2. wet 2
where [H20] is the mole fraction of water present in the off-gas. Since this
quantity is not measured it must be calculated by material balance.
119
-------
Table IV-2. is a summary of runs made In the miniplant regenerator. SC>2
concentrations are reported as average and peak mole fractions and as the
ratio of the peak SC>2 concentration to the equilibrium concentration of SC>2
expected at the temperature and pressure of the run. The temperature used
in determining the equilibrium concentration is the temperature in the
oxidizing zone of the bed, just before the gas reaches the bed surface. In
some runs it was difficult to estimate average temperatures accurately;
hence in these cases the ratio of measured to equilibrium S02 concentration
is a rough approximation.
For ten runs, the mean of the average SC>2 concentration was 1.0 mole percent,
The mean of the peak concentration was 1.8 mole percent. The mean of the
ratio of peak to equilibrium S02 concentrations was 0.46, after eliminating
three runs (nos. 2, 4, and 9) with very low ratios.
Extent of Regeneration of CaSC>4 to CaO—
. Two figures of merit have been developed to express the degree to which
sulfated stone has been regenerated. The first, denoted as fractional regen-
eration, gives the fraction (0 -> 1) of sulfur removed from the stone. In
order for fractional regeneration to equal unity, all sulfate present in the
original stone must be converted to oxide. The second figure of merit is
called the fractional sulfide formation, and is the number of moles of sul-
fide produced per mole of sulfate decomposed. Sulfide formation must equal
zero if fractional regeneration is equal to one.
The fractional regeneration and sulfide formation are computed from
analyses of the initial and final stone (i.e., before and after regeneration)
according to:
regeneration =
sulfide formation =
/w/o S\ __ /w/£_S\
Iw/o Ca/ . ~ \w/o Ca/
N/o S \
\w/o Ca/ .
/w/o S04~2^
w/o Ca
where w/o denotes weight percent, S refers to total sulfur and S~2 to sulfide.
All quantities in the above expressions are taken per unit weight of calcium
in order to prevent the figures of merit from changing due to a reaction not
involving sulfate or sulfide. For example, if only calcination occurred,
then w/o S would increase because the stone loses weight; however, w/o S v
w/o Ca would remain the same.
To calculate the figures of merit expressed above, one requires concen-
tration of Ca and S04~2 in the initial stone, and Ca, S04~2, S~2, and total S
in the regenerated stone. Initial stone contains sulfur only as sulfate so
total S can be obtained-from 804-2. Unfortunately, all of this information
120
-------
TABLE IV-2. REGENERATOR RUN SUMMARY
Avg. Bed Sup. Gas Air/Fuel SC>2 Cone.
Temp., Vel.
Ratio in in Effluent Peak S02 Cone.
jx.uii settled aea rressure K.ea./ux.
No. Charge Height, m kPa Zones, °C
1
2
3
4
5
6
7
8
9
10
11
SG (1)
32 kg
SG, 29.6 wt.
% S04, 32 kg
Same as run 2 ,
48 kg
Same as run 2,
48 kg
ST, 19% SO,,
48 kg
SP, 17.2% SO,
48 kg
SP, 17.2% SO,
48 kg
SG, 15.0%
S04, 48 kg
SG, 25.8% S04,
48 kg
SP, 21.5% S04,
48 kg
SG, 37.5%
0.6
0.6
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
910
910
910
910
910
910
910
910
930
910
910
1120/1116
1165/1165
1050/1054
1070/1124
1027/1088
1077/1094
1077/1088
1071/1071
1087/1112
1083/1083
1092/1103
Kea./ux. Keaucing
Zones, m/s Zone (2)
0.82/1.
0.79/1.
0.72/0.
0.68/0.
0.75/0.
0.64/0.
0.73/0.
1.0/1.
0.93/0.
0.90/0.
0.84/1.
16
07
91
93
85
82
77
0
96
90
07
6.5
6.6
6.0
6.0
6.6
6.8
8.3
8.2
7.8
7.8
6.7
Avg. / reals., u; Equli.
Mole %
S0~ Cone.
not measured
0.7/1
0.6/1
0.6/1
0.5/1
1.1/1
0.8/1
1.4/2
0.4/1
1.8/2
2.0/3
.2
.2
.3
.4
.9
.7
.1
.3
.9
.0
0
0
0
0
0
0
0
0
0
0
.09
.39
.16
.30
.37
.36
.58
.19
.69
.52
804, 48 kg
Notes: (1) Codes for Charges: SG = sulfated Grove limestone
ST = sulfated Tymochtee dolomite
SP = sulfated Pfizer dolomite
(2) Stoichiometric air/fuel
ratio =9.9
(3) Wet basis, corrected for water
condensed prior to analysis
-------
is available only for the last six of eleven runs made in the regenerator.
Furthermore, for many runs the total sulfur analysis was not consistent with
the sulfate and sulfide analyses (1/3 w/o 804-2 + w/0 S~2 should equal w/o S)
In such cases, sulfide was calculated as the difference of total sulfur and
1/3 of sulfate since it was believed that analyses for these components
(barium precipitation for sulfate and Dietert technique for total sulfur)
were more accurate than that for sulfide (iodometric titration).
Table IV-3 gives fractional regeneration and sulfide formation for six
runs. These results indicate that: (1) most sulfur was removed from the
stone during regeneration and (2) formation of sulfide was near zero for all
runs. The lower regeneration for runs 10 and 11 may be a result of stopping
these runs earlier than the others.
TABLE IV-3. FRACTIONAL REGENERATION AND SULFIDE FORMATION
Run No.
6
7
8
9
10
11
Regeneration
0.99
0.99
0.99
0.97
0.84
0.88
Sulfide Formation
0.01
0
0
0.01
0.01
0
Note: Definitions of "Regeneration" and "Sulfide
Formation" are given in text.
Material Balances—
Material balances for sulfur were calculated for the last six runs made
in the regenerator. Results are given in Table IV-4.
Sulfur enters the system only in the sulfated stone that is charged
prior to a run. Sulfur leaves in the off-gas as S02, in bed that is dis-
charged from the regenerator after a run, and in fines that are entrained from
the bed. A portion of the entrained fines are collected by a cyclone; however,
fines that are not collected represent a loss of sulfur that is not accounted
for in the material balance.
With the exception of two runs in Table IV-4, sulfur recoveries are quite
low. There are a number of factors that could account for this; the most
likely are (1) loss of bed when it is being removed from the regenerator
after a run, (2) loss of sulfur in particles entrained from the bed and not
collected by the cyclone, (3) errors in analyzing solids. The most serious
potential cause of low sulfur recoveries would be erroneous determinatxons
of S02 concentration in the off-gas, caused either by the gas sampling system
or a malfunction of the analyzer. These latter possibilities appear unlikely
because both the sampling equipment and analyzer had been thoroughly checked.
122
-------
TABLE IV-4. SULFUR BALANCES FOR REGENERATION RUNS
Sulfur In, kg Sulfur Out, kg
1-0
CO
Run No.
11
10
9
8
7
6
5
Bed Charged
5.96
3.41
4.10
2.38
2.73
2.73
2.87
Bed Recovered
0.93
0.35
0.06
0.01
0.02
0.03
0.04
Dlpleg Solids
0.36
0.36
1.00
0.61
0
0.01
0.57
Flue Gas
2.01
2.67
0.55
2.08
1.66
1.87
0.93
% S Recovered
55
99
39
113
62
70
55
-------
On the other hand, loss of solids collected after a run are known to be at
least 10-20 percent, and possibly quite more for some runs. Calcium balances
will be made in future runs in order to provide evidence that loss of solids
is a major cause of the low sulfur recoveries.
C02 concentrations were calculated from air and fuel flow rates for each
run and compared to measured values. For the last six runs, measured COo
concentrations averaged 86 percent of the calculated values, with a range of
76-105 percent. A reason why measured C02 concentrations are low may be that
nitrogen is added at several places to the regenerator (e.g., for purges at
pressure taps). This additional flow is not figured when C0£ concentrations
are calculated. An error in the air flow rate could also account for the dif-
ference between measured and calculated concentrations. These discrepancies
will be further examined because they may also affect S02 concentration.
Discussion
The batch runs made in the regenerator were extremely useful from the
standpoint of establishing operating procedures and providing operating
experience necessary to improve the performance and reliability of equip-
ment. It was clearly shown that .a batch of sulfated limestone or dolomite
could be regenerated under well controlled conditions to produce a gaseous
effluent containing about three mole percent S02- Furthermore, calcium sul-
fate could be converted almost completely to calcium oxide with very little
formation of sulfide, using the technique of adjacent reducing and oxidizing
zones.
A cause for optimism is that agglomeration of the bed could be avoided
by careful control of temperature. In experiments made at Exxon several years
ago in an 8.3 cm diameter regenerator (1) (the miniplant regenerator is
21.6 cm in diameter) most runs were made with pure calcium sulfate. A run in
which real sulfated limestone was used resulted in a badly agglomerated bed.
Also even though pure calcium sulfate was used in the smaller regenerator,
only about two percent S02 was produced at temperatures under 1100°C. Three
percent S02 was obtained in the last two runs made in the miniplant regen-
erator, using sulfated limestone. The improved performance of the miniplant
regenerator, compared to the smaller unit, is probably largely due to the
improved quality of fluidization in the miniplant. The best means available
to determine how well solids are fluidized is the closeness of the tempera-
ture distribution in the bed. Temperatures in the miniplant were much more
uniform, indicating that quality of fluidization was much better than in the
smaller unit.
Peak values of measured S02 concentration were as high as 69 percent
of the equilibrium concentration at the temperature and pressure of the run.
The average approach to equilibrium was 46 percent. Temperature of the run
refers to the temperature in the oxidizing zone, where gas contacts the last
solids before leaving the bed.
Equilibrium concentrations were calculated based on the available free
energy data for CaS, CaSO^ CaO, and S02, assuming perfect gas behavior and
no formation of solid solutions. However, Curran, et. al. (10), determined
experimentally the equilibrium for the reaction:
3CaS04 + CaS = 4CaO + 4S02
124
-------
and found that the equilibrium pressures of S02 were considerably lower than
the calculated values. For example, at 1100°C, Curran's value is 32 kPa
compared to 52 kPa for the calculated value.
Table IV-5 gives the ratio of the peak SC>2 concentrations for run Nos.
2-11 to the equilibrium concentrations determined by Curran. The ratios
based on calculated S02 concentrations (given in Table IV-2) are repeated for
comparison. In three runs (e.g., nos. 2, 4, 9) measured SC^ concentrations
were much less than equilibrium concentrations; however, abnormal conditions
could account for this. In run nos. 2 and 4 the bed agglomerated and in run
no. 9, a large portion of the bed was blown out of the regenerator.
Because of the considerable differences between the equilibrium SC>2 con-
centrations calculated from available free energy data and Curran's experi-
mental values, additional experimental studies to find the true equilibrium
would be worthwhile. If Curran's data are correct, then equilibrium was
essentially established in several regeneration runs. If the calculated
values are correct, then further improvement is possible toward increasing
S02 concentrations.
COUPLING OF REGENERATOR TO COMBUSTOR
The overall objective of this program was to demonstrate combined opera-
tion of the combustor and regenerator with circulation of solids between the
two vessels. Characterization of the regenerative system will be made by
operating the system over a range of test conditions. It is planned to
determine what operating condition in the combustor and regenerator provide
low emissions of S02 from the combustor, high concentrations of S02 in the
off-gas from the regenerator, low makeup rates of fresh sorbent, and
moderate recirculation rates between combustor and regenerator.
Problems were anticipated in developing a system to continuously transfer
hot solids in the temperature range of 850-1125°C. Work proceeded in
several stages: (1) cold-test alternative transfer systems to find that sys-
tem which best deserved additional development, (2) make brief hot runs to
further test the system and make improvements, (3) conclude shakedown of the
combustor-regenerator by operating the units continuously for 24 hours at
elevated temperature and pressure, (4) demonstrate long term operability by
completing a run lasting 4-5 days, and (5) operate the combustor-regenerator
over a range of test conditions to begin to characterize the system. This
report describes stages (l)-(3), which were completed as of July, 1976.
Equipment Development
Early Problems—
The first system tested depended on pulses of nitrogen gas to transfer
solids. Figure IV-6 shows the combustor-to-regenerator transfer line. The
regenerator-to-combustor line was similar. Solids from the combustor spill
into the take-off port, fill the transfer line, and are moved into the
regenerator by pulsing the bottom of the transfer line with nitrogen. The
slide valve is used to correct upset conditions, for example to allow the
line to be refilled with solids if they get blown out.
125
-------
t-o
Temperature
TABLE IV-5. COMPARISON OF MEASURED AND EQUILIBRIUM SO- CONCENTRATIONS
Measured SO,.
Peak Measured Equil. S02 Cone.
Equil. S02
Cone, based on
Measured SO,.
K.un
No.
2
3
4
5
6
7
8
9
10
11
in ux. z.one, ;
°C
1165
1054
1124
1088
1094
1088
1071
1112
1083
1103
^2 concentration,
Mole %
1.2
1.2
1.3
1.4
1.9
1.7
2.1
1.3
2.9
3.0
calculated v,i;
Mole %
14
3U
8.0
4.7
5.1
4.7
3.6
6.7
4.2
5.8
curran, \,±j
Mole %
9.3
1.7
5.1
2.9
3.2
2.9
2.2
4.2
2.7
3.7
Calculated S02
0.09
0.39
0.16
0.30
0.37
0.36
0.58
0.19
0.69
0.52
Curran SO
0.13
0.71
0.25
0.48
0.59
0.59
0.95
0.31
1.07
0.81
(1) For 9 atm total pressure and average temperature in oxidizing zone.
-------
COMBUSTOR-REGENERATOR SOLIDS TRANSFER LINE
COM BUSTOR
EXPANSION
JOINT
NITROGEN PULSE
REGENERATOR
FIGURE IV-6
ORIGINAL SOLIDS TRANSFER SYSTEM:
COMBUSTOR-REGENERATOR SOLIDS TRANSFER LINE
127
-------
Early cold tests showed that solids could be easily transferred from
combustor to regenerator provided that the level of the fluidized bed in
the regenerator was low. As the bed level in the regenerator increased,
pressure increased at the bottom of the bed, i.e., at the exit of the trans-
fer line. This resulted in an increased back flow of gas up the transfer
line. As the regenerator bed level became still higher, gas backflow
increased sufficiently to blow solids up the transfer line.
Transfer in the opposite direction, from regenerator to combustor, fol-
lowed a similar pattern. Only at low bed levels in the combustor could transfer
be accomplished. This problem was explained by a simple pressure balance,
shown in Figure IV-7. The pressure drop across the combustor-to-regenerator
transfer line, AP]_2> becomes less favorable as bed level in the regenerator
increases. The value of APa (regenerator pressure-combustor pressure) is
the set point on the AP controller, and this value can be adjusted to pro-
vide a satisfactory value of AP-j^ for transfers from combustor to regenerator.
However, AP^^ would then be highly unfavorable for transfer from regenerator
to combustor. Unfavorable values of APi2 °r A?34 mean large negative values,
large enough to permit excessive gas backflow.
It was concluded that by adjusting APa solids could be easily trans-
ferred but only one way at a time.
Modifications—
Several alternative approaches for transferring solids were proposed
and cold tested. These would be used on only one transfer line; the second
line would remain unchanged. The pressure balance would be adjusted so as
to permit good transfer through the unmodified line. Summarized below are
the alternatives that were investigated.
(1) Two Valves in Transfer Line - Solids would be trapped between
the two valves by opening and then closing the upper valve.
Transfer would be accomplished by pulsing or blowing out
the solids after the lower valve was opened.
(2) Solids Reject Lockhopper/Pneumatic Conveying. Solids would
be rejected and collected into a lockhopper in much the same
manner as solids are rejected from the combustor. Solids
would then be pneumatically conveyed from the lockhopper
into the receiving column.
(3) Vary Combustor-Regenerator AP. Solids can be transferred one
way at a time by maintaining a favorable AP across the active
transfer line. In this approach solids would be transferred
from combustor to regenerator by setting the AP so as to
facilitate transfer in this direction. This transfer line
would then be closed with a slide valve and the AP adjusted
to permit transfer through the regenerator to combustor line.
(4) U-tube - This technique is similar to that used in catalytic
cracking to transfer catalyst between cracker and regenerator.
Solids would spill into a long vertical standpipe where they
would form an effective gas seal. Air (or nitrogen) would be
added at the bottom of the standpipe to convey solids into
the receiving column.
128
-------
A PCB
FLUIDIZING GRID
RB
COMBUSTOR
REGENERATOR
A P34 - APR _
APRB- APa
A > r> D A r 3
WHERE A P
A P
A P
= P _D AD = D _ D
12 Kl K2/AI34 *3 K4
,, =AP FOR COMBUSTOR BED ABOVE SOLIDS TAKEOFF
R
= AP FOR REGENERATOR BED ABOVE SOLIDS TAKEOFF
A PRB = AP FOR ENTIRE REGENERATOR BED
A PCB = AP FOR ENTIRE COMBUSTOR BED
APa = REGENERATOR PRESSURE-COMBUSTOR PRESSURE
(MEASURED AT TOPS OF COLUMNS)
FIGURE IV-7
TRANSFER SYSTEM PRESSURE BALANCE
129
-------
(5) Other Approaches. Two other possibilities include
transferring solids up (rather than down) the transfer
lines, or using a mechanical screw rather than pulsing
gas to move solids into the receiving column.
Approaches (l)-(3) were cold tested and the transfer line lockhopper
technique (2 valves in transfer line) was judged to offer the greatest
potential for success. Figure IV-8 is a diagram of the piping configuration
used in the transfer line lockhopper technique. Two automatic slide valves
in the combustor-to-regenerator transfer line trap solids in the piping
between them. Solids are discharged into the regenerator when the bottom
valve is opened. A stream of purge nitrogen that is supplied to the valves
assists in "pushing" solids out of the transfer line into the regenerator.
Nitrogen pulses are supplied to the bottom of the combustor-to-regenerator
line only as a backup method of keeping solids moving in the line. Pressure
in the regenerator is maintained above the combustor so that the regenerator-
to-combustor transfer line does not need two valves. The single manual slide
valve in this line is closed only during startup and during upset conditions.
Solids are blown into the combustor from the lower end of the regenerator-to-
combustor line by applying a fairly strong pulse of nitrogen (producing a
superficial velocity of approximately 2 m/s in the transfer line). The
rate of solids transfer is controlled by the frequency of the pulse.
Two plugs in the solids takeoff ports are shown in Figure IV-8 in the
open position. These plugs are inserted into the take-off ports during
startup in order to prevent solids from entering the transfer lines. In the
past, solids entering the transfer line during startup resulted in plugging
of the transfer line when condensation of water caused the solids to
agglomerate.
A six-hour test of the solids transfer system, during which coal was
burned in the combustor, was completed successfully. Solids were trans-
ferred between combustor and regenerator at rates of about 70 kg/hr. No
problems of any kind developed with the transfer system during the test.
Combustor and regenerator operating conditions are summarized below.
Illinois coal was fed to the combustor for about 1.3 hours; Champion coal
was fed during the remainder. It should be noted that the regenerator was
operated under oxidizing rather than reducing conditions. Reducing con-
ditions could have been established by increasing the input of supplementary
fuel. Operating the regenerator under reducing conditions would not have
been expected to affect transfer of solids.
Operating Conditions During Six Hours Solids Transfer Test
Variable
Avg. bed Temperature, °C
Pressure, kPa
Superficial velocity, m/s
Expanded Bed Depth, m
Coal Feed Rate, kg/hr
Excess Air, %
Combustor
862
870
1.8
3.6
93
49%
Regenerator
954
870
0.73
2.5
130
-------
REGENERATOR
COMBUSTOR
SOLIDS TAKE
OFF PLUG
7 AUTO
SLIDE
VALVE
MANUAL
SLIDE
VALVE
NITROGEN
PULSE
FIGURE IV-8
MINIPLANT SOLIDS TRANSFER SYSTEM
-------
Components of Transfer System—
Slide Valves—The slide valves are a critical part of the solids trans-
fer system. They must be able to withstand hot solids at temperatures up to
about 1000°C, working pressures up to 1000 kPa, and potentially corrosive
gas atmospheres. Furthermore, the valves must be repeatedly open and shut.
Because of these severe operating conditions, Exxon Engineering Technology
Department (EETD) was asked to evaluate our design and recommend improvements.
As a result, a modified design was developed which included (a) enlarging the
bore of the valve from 5.1 to 7.6 cm, eliminating the stainless steel liner
and replacing it with reinforced refractory, using a wider knife with a blunt
edge and providing better cooling for the knife, and using a more rugged stem
packing system.
Expansion Joints—Each transfer line contains an expansion joint to
take up thermal expansion. These joints were quite bulky and made it dif-
ficult to perform mechanical work on the transfer lines. EETD was therefore
asked to determine if expansion joints were really needed. Their analysis
showed that the joint on the combustor-to-regenerator line could be removed
if skin (shell) temperatures were under 177°C; the joint on the regenerator-
to-combustor line could be removed if temperatures were under 135°C. The
cumbersome expansion compensators attached to the joints were judged not
needed and were removed.
Measurements of transfer line skin temperatures were made during hot
tests. Not surprisingly, skin temperatures depended on solids' transfer
rates. For rates up to about 145 kg/hr maximum average skin temperatures
for the combustor-to-regenerator and regenerator-to-combustor transfer lines
were 146°C and 200°C, respectively. During these tests, the regenerator was
at 980°C, a somewhat lower than normal temperature. The combustor tempera-
ture was about 900°C. Hence, it appears that the expansion joint on the
regenerator-to-combustor line will be required, but that the joint on the combus-
tor-to-regenerator line could be removed.
Transfer Line Pipe and "Pulse Pots"—The transfer lines are fabricated
from 15.2 cm (6 inch) Schedule 40 pipe refractory lined to an inside diameter
of 7.6 cm. The sloping portions of the transfer lines are sleeved with
6.4 cm (2-1/2 inch) Sch. 10 316 stainless steel pipe with an inside diameter
of 6.7 cm. Because the slide valves have a smaller diameter bore (5.1 cm)
than the lines, transition pieces are used to reduce smoothly the diameter
from 6.7 to 5.1 cm.
Pulses of nitrogen which are admitted to the lower section of the
regenerator-to-combustor transfer line, called a "pulse-pot," blow solids
lying between the pulse-pot and regenerator vessel into the vessel. Nitrogen
enters through a 1.3 cm (1/2 inch) stainless steel tube whose outlet is
positioned approximately at the center of the pulse-pot. Superficial
velocity of nitrogen in the 7.6 cm diameter line connecting the pulse pot
and regenerator is about 2 m/s. The solids transfer rate is controlled by
adjusting the duration and frequency of the pulse. Typically, a pulse
lasting 1-2 seconds occurs every 30-60 seconds.
132
-------
Procedures
Startup of Transfer System—
The solids transfer system is not operated during startup to prevent
moisture from entering the transfer lines. Plugs are inserted in the solids
takeoff ports so that solids cannot spill into the lines from the combustor
and regenerator vessels. Nitrogen is pulsed into the transfer line pulse
pots in order to prevent solids from backing up into the lines. The slide
valves are also cycled occasionally in order to dislodge any solids that^may
have passed by the takeoff plugs. Pressure in the regenerator is set slightly
above that in the combustor (usually 2.5-10 kPa) and the beds are heated to
near operating temperatures. Transfer of solids is started by pulling the
plugs out of the take-off ports.
Operation—
Transfer rate is controlled by setting the cycle time of the slide valves
in the combustor-to-regenerator line. The valves are opened and closed auto-
matically by a series of timers on the miniplant control panel. Action fol-
lows the following sequence: top valve opens, line fills with solids, top
valve closes, bottom valve opens, line empties, bottom valve closes. The
volume between the slide valves holds about 6.0 kg of solids; hence, setting
the cycle time gives the transfer rate of solids from combustor to regenerator.
If bed levels are kept constant, then the transfer rate from regenerator to
combustor must be the same.
Shutdown—
It is important, during shutdown, to empty the transfer lines of solids;
otherwise plugging of the lines may occur when the unit is restarted. Hence,
the first step is to shut the plugs in the solids takeoff ports, thereby
preventing solids from entering the lines. Cycling of the slide valves in
the combustor-to-regenerator line and pulsing nitrogen in the regenerator-to-
combustor line is continued until the miniplant is shutdown so that the lines
are emptied.
Performance of Transfer System -
24 Hour Shakedown Run
The miniplant was operated continuously for 24 hours in order to test
the system for transferring solids between the combustor and regenerator.
Operation of the system is described and process data are given in the fol-
lowing sections.
Operation of the Solids Transfer System During Shakedown Run—
The miniplant regenerator and combustor were operated simultaneously for
24 hours on July 26-27, 1976. Solids were continuously transferred between
the two vessels for the entire period. Operation of the solids transfer sys-
tem went extremely well and not a single problem developed once flow of solids
was begun. The regenerator was operated under reducing conditions for about
23 and 1/2 hours. However, problems developed in feeding Illinois coal and
kerosene had to be burned in the combustor for about six of the 24 hours.
However, the basic intent of the run was to demonstrate continuous operation
of the combustor and regenerator for a 24-hour period and this was done.
Nominal operating conditions are given in Table IV-6. A log of noteworthy
events during the run is given in Table IV-7.
133
-------
TABLE IV-6. NOMINAL OPERATING CONDITIONS FOR
COMBUSTOR AND REGENERATOR DURING SHAKEDOWN RUN
Pressure, kPa
Temperature, °C
Fluidized bed height, m
Superficial gas velocity, m/s
Combustor coal feed rate, kg/hr
Combustor excess air, %
Air/fuel ratio in regenerator
reducing zone*
Limestone feed rate to combustor,
equiv- Ca/S
Solids recirculation rate, kg/hr
Combustor
910
900
4.5
1.8
80.3
44
0.8
Regenerator
915
1040
2.5
0.7
8.2
45
* Stoichometric - 9.9
TABLE IV-7. LOG OF EVENTS FOR 24 HOUR SHAKEDOWN RUN
Hours After
Run Start
0
1/2
1
1 1/2
14 1/2
20 1/2
25
Event
Coal feed started to combustor (Champion, 2% S)
Solids transfer begun from combustor to regenerator
Solids transfer begun from regenerator to combustor
Regenerator in reducing conditions
Plugging in coal feeding system due to Illinois coal,
switched to kerosene fuel for combustor
Arkwright coal (2.6% S) feed started to combustor
Failure of regenerator fluidizing grid, combustor
and regenerator shut down
Transfer rate of solids was 45 kg/hr during the entire run. The slide
valves in the combustor-to-regenerator line were cycled every eight minutes.
In the regenerator-to-combustor line nitrogen was pulsed for 1 second every
60 seconds.
Process Data - Shakedown Run—
It was intended to operate the combustor and regenerator at constant con-
ditions, varying only the ratio of Ca/S in the feed in order to keep emissions
of S02 from the combustor steady at about 300-400 ppm. However, steady
conditions were not reached until about seven hours into the run. Over this
134
-------
period, sulfation levels in the combustor and regenerator reached approx-
imately steady values. The material charged into the combustor and regener-
ator prior to the run was sulfated, not fresh, limestone. During a second
seven-hour period after steady conditions were reached, SC>2 levels of 300-
400 ppm were maintained with a Ca/S feed rate of 0.8. The level of S02
expected for once-through (non-regenerative) operation of the combustor at
this Ca/S ratio was about 1000 ppm. This result showed that the regenerator
was effectively reducing the emissions of S02 from the combustor or, put
another way, decreasing the feed rate requirement of limestone into the
combustor.
Emissions from both combustor and regenerator are summarized in
Table IV-8. The regenerator may have also acted as a calciner. Conditions
in the combustor favored carbonate whereas higher temperatures in the regen-
erator favored oxide. Hence, in a regenerative system it would be possible
to use limestone in the combustor even at conditions where limestone would
not calcine in the combustor.
TABLE IV-8. SHAKEDOWN RUN NOMINAL EMISSION LEVELS
FROM COMBUSTOR AND REGENERATOR
Combustor Regenerator
S02 350 8750
C02, % 14.0 13.5
CO, ppm 120 1825
02, % 6.3 0.3
NO , ppm 110 3
X
(For steady period, 7-14 hours from start of run.
Operating conditions given in Table IV-6)
The concentration of S02 in the regenerator off-gas was typically
0.7-1.0%. This level of S02 is close to that which is predicted by a sulfur
material balance on the regenerative system. The equilibrium level of S02
at regenerator conditions (1040°C, 915 kPa) is about 2.8%. Hence, the S02
level was limited by the rate of input of sulfur to the system (material
balance) rather than by thermodynamics. Increasing the coal feed rate or
burning coal with a higher sulfur content could have increased the S02 con-
centration.
After 14-1/2 hours into the run (see Table IV-7), kerosene (which contains
no sulfur) replaced the Champion coal feed. Regeneration continued during the
six hour period during which kerosene was burned in the combustor. When coal
combustion was resumed the measured level of S02 emissions in the combustor
was zero, perhaps indicating that a fully regenerated and a highly calcined
stone with a strong affinity to capture S02 was produced in the regenerator
during this period.
135
-------
Another optimistic result was that bed levels remained nearly constant
for about the last ten hours of the run even though no makeup limestone was
added. This may be due to a high degree of attrition resistance. More data
will be needed, however, in order to determine the effect of cycling between
combustion and regeneration conditions on the attrition resistance of the
sorbent.
136
-------
SECTION V
DISCUSSION OF RESULTS
COMPARISON OF BATCH UNIT AND MINIPLANT RESULTS
The operation of the batch unit and miniplnat are quite different, and
it is interesting to compare the more significant results from the two units.
The miniplant is a continuous unit. A mixture of coal and sorbent is fed
continuously and solids are removed directly from the bed to keep the bed
height at a desired level. Some solids are also removed overhead as
entrained flyash. The miniplant can be operated under steady state condi-
tions. In the batch unit, on the other hand, only coal is fed continuously
into a bed of sorbent whose composition changes continuously.
The sulfation level of the bed in the batch unit increases continuously,
whereas, in the miniplant combustor, the average sulfation level of the bed
reaches a constant value after a sufficient length of time on stream. There
is always a certain amount of fresh sorbent in the miniplant bed giving a
distribution of particles with different levels of sulfation. In the batch
unit, all particles have the same "age" and approximately the same sulfation
level.
Figure V-l is a comparison of the sulfur dioxide retention in the mini-
plant and batch unit for runs with dolomite sorbent. The solid lines repre-
sent data from the miniplant for residence times of 1 second and 0.5 second.
The points are retentions in the batch unit for residence times between 0.5
and 1 second. The retention in the batch unit is lower than would be pre-
dicted from miniplant data.
This result is not at all surprising when the difference between the
two units is considered. It is known that a sorbent particle loses activity
rapidly as its level of sulfation increases. Since there is a continuous
stream of fresh sorbent into the miniplant, the activity of the bed is
expected to be higher, for the same average level of sulfation, than that of
the batch unit.
Figure V-2 shows a comparison of the combustion efficiencies of the two
units. The solid lines represent two of the levels of combustion efficiency
found in the miniplant. The points in Figure V-2 represent runs in the
batch unit»
At lower temperatures the miniplant has a higher combustion efficiency
than does the batch unit. However, data from both units seem to converge to
greater than 99% at a temperature of about 975°C. The higher efficiency of
the miniplant is to be expected since the unit is equipped to recycle over-
head solids. This recycle of unburned carbon should increase the combustion
efficiency.
Figure V-3 compares the NOX emissions of the two units. In general,
the batch unit has higher emissions than does the miniplant. As pointed out
earlier, the batch unit seems to give the same type curve of NOX vs. excess
137
-------
FIGURE V-l
OJ
00
COMPARISON OF SULFUR DIOXIDE RETENTION MEASURED IN MINIPLANT
AND BATCH UNITS - DOLOMITE SORBENT
100
80
2 60
LU
I-
LU
CM
o 40
CO
0
I
Miniplant
Batch
CALCIUM TO SULFUR MOLAR RATIO
-------
FIGURE V-2
Lo
VO
>
o
o
LL_
u.
LU
Z
O
h-
(/)
13
CQ
^
O
O
o
CQ
O
100
98
96
94-
92
90
88
86
84
82
80l—
500
COMPARISON OF COMBUSTION EFFICIENCIES MEASURED IN
MINIPLANT AND BATCH UNITS
— Miniplant
• Batch Unit
600 700 800
AVERAGE TEMPERATURE (C°)
900
1000
-------
FIGURE V-3
1.0
COMPARISON OF NOX EMISSIONS MEASURED IN
MINIPLANT AND BATCH UNITS
I
I
0.8
CQ
x
O
0.6
0.4
0.2
Batch With
Western Coal
Data
Batch Without .Western Coal Data
Miniplant
1
0
20
40
60 80 100
EXCESS AIR (%)
120
140
160
-------
air as does the miniplant when runs with Western coal are not included. If
Western coal data are included, the NOX emissions from the batch unit con-
tinue to increase even to 160% excess air instead of leveling out as had been
observed from miniplant results.
The difference in the NOX emissions from the two units is not easily
explained. The nature of NOX formation and disappearance is not known. At
the low temperatures of fluid bed combustion, it seems certain that the NOX
is produced from the nitrogen in the coal. Various suggestions have been
made as to how it disappears (e.g., reacts with CO, C, SC>2 , etc.) but since
these mechanisms are not very well substantiated, explaining differences
between two units is very difficult.
141
-------
SECTION VI
CONTINUING STUDIES
The future program in the miniplant and batch units will be centered on
four major tasks; combustion studies, flue gas particulate removal, a
comprehensive analysis of emissions and a study of sorbent regeneration.
COMBUSTION STUDIES
In the miniplant, the current combustion program will be continued to
include tests using a high sulfur Illinois No. 6 coal with dolomite sorbent
in a series of runs in which combustor temperature and Ca/S molar ratio are
varied.
A study will also be made of the effect of using precalcined limestone
sorbent under combustion conditions which do not favor extensive calcination
of the limestone. If successful, this would increase the desulfurization
activity of limestone at intermediate combustion temperature conditions where
the stone is only partially calcined and the limestone activity is low. It
could also make it possible to use limestone at very low temperature con-
ditions, such as those occurring during "turndown" operation, where limestone
does not calcine and is virtually inert.
A combustion program will also be carried out in the batch unit. The
batch combustor will first be modified to convert it to a continuous unit.
This will require the addition of continuous sorbent feeding and spent sor-
bent removal systems. The combustor will then be used in a program to pro-
vide technical support for the miniplant and also to investigate specific
technical problem areas.
FLUE GAS PARTICULATE REMOVAL
Particulates must be removed from the combustor flue gas to low levels
to meet both environmental and gas turbine requirements. The current EPA
particulate emission standard for a new coal fired boiler is 0.1 Ib/M BTU
of coal fired (0.043 g/MJ). Currently, particulate size and/or composition
are not specified in the emission control regulations. For a typical coal,
this standard translates to a particulate concentration in the flue gas of
about 100 mg/m^ (0.05 gr/SCF). This level cannot be reached by the use of
conventional cyclones to reduce the particulate loading. A realistic level
that can be attained by the use of two conventional cyclones operated in
series is about 300 mg/m3 (0.15 gr/SCF). Therefore if a two stage cyclone
system is used to remove a portion of the particulates from the flue gas, a
third stage device would still be required,to reduce particulate loadings
from the vicinity of 300 mg/m to 100 mg/m to meet current environmental
standards. This requires a removal efficiency of about 67% in the third
stage cleanup device.
Allowable particulate loadings to minimize gas turbine erosion are pre-
sently indicated to be even lower than those required to meet the environ-
mental regulations. At the present time, turbine erosion limits are not well
142
-------
defined. Westinghouse Research Laboratory recently estimated allowable levels
based on limited data and model studies (5). The estimates covered a range
of 50 to 0.9 mg/m3 (0.02 to 0.0004 gr/SCF). Based on these estimates,
Westinghouse suggested a tentative allowable level of 5 mg/m^ (0.002 gr/SCF).
To meet this very low particulate concentration, a third stage particulate
removal device would be required to operate at an efficiency of 98.7%. These
allowable particulate concentrations and required removal efficiencies are
summarized in Table VI-1.
TABLE VI-1. PARTICULATE EMISSION CONTROL REQUIREMENTS
Required Efficiency
Requirement Allowable Level of Third Stage Device
(mg/m3)(gr/SCF) (%)
Environmental 100 0.05 67
Turbine Erosion
Range of estimates 50-0.9 0.02-0.0004 87-99.7
Tentative level 5 0.002 98.7
In addition to the above, an even tighter limitation on allowable par-
ticulate concentrations may be imposed by turbine corrosion considerations.
However, at the present, insufficient data are available to estimate the
allowable level required to prevent corrosion.
The objective of the flue gas particulate removal program is to evaluate
two removal devices which have the potential for reducing the particulate
loadings to the required levels. Since gas turbine considerations now appear
to set the required particulate removal efficiencies, these more stringent
efficiencies will form the target levels for the study. Although the primary
intent of this EPA sponsored program is to measure and characterize the par-
ticulates escaping the third stage device for their potential environmental
impact, the measurements, to be realistic, must be made under conditions which
are aimed at protecting the gas turbines. The effect of the particulates on
gas turbines is of secondary importance to this program and will be measured
using an erosion test passage desinged and fabricated by Westinghouse Research
Laboratory under contract to the EPA. In addition, EPA and the Energy
Research and Development Administration (ERDA) are cooperating on a second
program to evaluate in the miniplant the resistance to erosion and corrosion
of a number of turbine test specimens to be provided by the General Electric
Company under contract to ERDA. Therefore, the particulate control program
will ultimately be concerned with both environmental and gas turbine con-
siderations.
The first particulate control device to be evaluated will be a granular
bed filter purchased from the Ducon Company. This filter was chosen after
surveying the type of devices currently available. Use was made of a series
of earlier surveys made by Stone and Webster (11) and Westinghouse (5).
Discussions were also held with EPA and Exxon personnel active in the area of
particulate control. Other control devices mentioned in the surveys are
143
-------
high temperature metal or ceramic filters and a low temperature scrubber com-
bined with an efficient heat exchanger. However, the consensus favored the
granular bed filters as the type of device which currently offers the best
chance of meeting the high removal efficiency targets.
A number of groups are currently developing granular bed filter systems
which are described in the above referenced Westinghouse and Stone and
Webster reports. The Ducon Company has tested granular bed filters on
refinery and other waste gas streams. A type of filter, called the panel bed
filter is currently being studied at the City College of New York under
sponsorship of the Electric Power Research Institute (EPRI). The Rexnord
Company has installed a number of "gravel bed" filters on cement kiln and
other industrial off gases. Combustion Power Company has also installed
a number of "dry scrubber" filters on flue gas from wood waste boilers.
Combustion Power is also developing a similar system for application on FBC.
The Ducon filter was chosen for testing on the Exxon/EPA miniplant after
considering all the available systems. The selection was based on previous
experience with the Ducon system which indicated that it had the potential of
providing high removal efficiency. Another desirable feature of the Ducon
system is the retention of the granular filter medium in the filter vessel.
In all other systems, the medium is removed, cleaned externally and recycled
back to the filter vessel.
A sketch of the conventional Ducon filter is shown in Figure VI-1. The
filter consists of a series of beds containing the filter medium, stacked
vertically to form a filter element. A number of elements are contained in
a pressure shell. Dirty gas passes through an inlet screen, through the
filter medium and out through an outlet screen into a central collecting tube.
Clean gas exits the vessel at the bottom. Each element is periodically
cleaned by a short pulse of high pressure clean air flowing in reverse flow
through each element. The dust is blown out through the inlet screens and
collects in the bottom cone of the pressure vessel. The filter medium is
retained in the beds.
The filter system to be used on the miniplant will be a modified version
of the conventional Ducon system described above. Four filter elements will
be used and will be designed to test three different filter medium cleaning
(blow back) techniques. Each element will be encased in a shroud to allow
segregation of the particulates removed from it during the blow back.
Separate feed ducts will supply dirty gas to each element and each duct will
contain an orifice to permit measurement of flow to each element. One of the
elements will use the high pressure pulse blow back described above. Another
element will use a "positive blow back" technique. It will be equipped with
shutoff valves to allow it to be completely isolated from the feed and pro-
duct streams during blow back. The element can then be depressured and blown
back with a larger volume of low pressure air. Two of the elements will use
a modified "positive blow back" technique. In this case only one end of the
element will be shut off by a valve during blow back to permit blow back by
a larger volume of air slightly above the filtration pressure. The positive
blow back systems should provide more effective regeneration of the filter
beds and a better chance of meeting the high removal efficiency targets.
144
-------
Flue Gas
Inlet
Clean Gas -^
Outlet
Filter
Element
Collected
Fines
Blowback
Gas Ports
FILTER ELEMENT INTERNALS
Outer
Screen
Granular
Sand Bed
Inner Screen
FIGURE VI -1
DUCON GRANULAR BED FILTER
145
-------
The positive blow back element consist of 10 beds, each 0.3 m (1 ft) in
diameter. The element is 1.8 m (6 ft) long. The remaining three elements
contain 12 beds, each 0.3 m (1 ft) in diameter. These elements are also 1.8 m
(6 ft) long. The elements are contained in a refractory lined pressure vessel
designed for 1000 kPa (10 atm) operating pressure. The system will be capable
of filtering 345 m^/min (1200 SCFM) of hot flue gas. Installation is sched-
uled for completion in late 1976.
The filter will be evaluated and performance optimized in an experimen-
tal program in which the particulate removal efficiency and efficiency main-
tenance will be measured as a function of operating parameters. These para-
meters include the method of blow back, blow back frequency, duration, pres-
sure and blow back gas volume. The nature and particle size of the filter
medium will be varied as will the flow rate and particulate loading of the
inlet gas. Provisions are included to permit major modifications to this
filter system as required to improve and/or maintain particulate removal
efficiency.
In addition to the evaluation of the Ducon filter, a second particulate
removal device will be chosen, procurred, and tested. The choice of this
device will be made in cooperation with the EPA.
COMPREHENSIVE ANALYSIS OF EMISSIONS
A comprehensive analysis of gaseous and solid streams leaving the mini-
plant and batch units will be made to obtain a complete inventory of poten-
tially harmful emissions. Sampling and analytical systems will be chosen
for this program in cooperation with the EPA and Battelle Columbus Laboratory,
the EPA Environmental Assessment Contractor for fluidized bed combustion.
The comprehensive analysis program will be carried out in three phases.
In the first phase, "Level 1" analyses will be made to obtain an indication
of the type of compounds present in the combustor flue gas before and after
the third stage particulate removal device, in the regenerator off gas and
in the solids streams removed from the combustor, i.e., the spent sorbent
removed from the combustor and the particulates removed by the cyclones and
third stage device. The particulates may be divided into two size fractions,
respirable and non-respirable. The general type of chemical compounds to be
sought in the sampled streams are various sulfur compounds, nitrogen com-
pounds, organic compounds, inorganic carbon compounds, chlorine and fluorine
compounds and trace elements. Particulates will also be tested for size,
size distribution and physical structure. Biological testing of some of
the sampled materials will also be carried out.
After Level 1 analyses have been conducted, Level 2 analyses will be
made in which potentially harmful species identified in the Level 1 series
will be measured quantitatively. This will be followed by Level 3 analyses
which will determine periodically during operation of the fluidized bed
units the concentration of those species identified in Level 2 as being
present in significant concentrations.
146
-------
REGENERATION STUDIES
Regeneration studies will be carried out in miniplant and modified
batch units. In the miniplnat, an extended demonstration run of 4 to 5
days duration will be made in which the combustor and regenerator will be
operated continuously.
A series of regeneration tests will then be made in both the miniplant
and batch units. In the batch unit, a test program will be run to identify
the operating conditions which provide high levels of sulfur removal from
the sorbent and high levels of SC>2 in the regenerator off gas. Cyclic studies
involving the batch combustor will also be made to measure activity main-
tenance of the sorbent.
In the miniplant, a series of tests will be made which will be aimed at
characterizing and optimizing the continuous combustion/regeneration system.
The test series will determine which combustor and regenerator operating con-
ditions provide, simultaneously, low emissions of S02 and other pollutants
from the combustor, high concentrations of S02 in the regenerator off-gas,
low makeup rates of fresh sorbent and moderate recirculation rates between
combustor and regenerator. An additional objective will be to minimize the
attrition rate of sorbents, since this will reduce the required amount of
makeup sorbent and reduce the load on the flue gas particulate removal
-devices.
147
-------
SECTION VII
REFERENCES
1. Hoke, R. C., et al, "A Regenerative Limestone Process for Fluidized-Bed
Coal Combustion and Desulfurization," EPA-650/2-74-001, January 1974.
2. Hoke, R. C., et al, "Studies of the Pressurized Fluidized-Bed Coal
Combustion Process," EPA-600/7-76-011, September 1976.
3. Skopp, A., et al, "Studies of the Fluidized Lime-Bed Coal Combustion
Desulfurization System," Esso Research and Engineering Company for EPA,
1971.
4. Jonke, A. A., et al, "Reduction of Atmospheric Pollution by the Applica-
tion of Fluidized Bed Combustion," EPA-Argonne National Laboratory,
ANL/ES-CEN-1002, June 1970.
5. Keairns, D. L., et al, "Fluidized Bed Combustion Process Evaluation -
Phase II Pressurized Fluidized Bed Coal Combustion Development,"
Westinghouse Research Laboratories, EPA-650/2-75-027c, September 1975.
6. Vogel, G. J., et al, "Annual Report on a Development Program on
Pressurized Fluidized Bed Combustion," ANL/ES-CEN/1011, July 1975.
7. National Research Development Corp., "Pressurized Fluidized Bed
Combustion," R&D Report No. 85, Interim No. 1 for Office of Coal Research.
8. Cox, D. G., et al, National Coal Board Final Report, "Reduction of
Atmospheric Pollution," Vol. 2, for EPA, September 1971.
9. Vogel, G. J., et al, "A Development Program on Pressurized Fluidized
Bed Combustion," Quarterly Report to U.S. Energy Research and Development
Administration, Contract No. 14-32-0001-1780, January 1976.
10. Curran, G. P., Fink, C. E., and Gorin, E., "C02 Acceptor Gasification
Process," Fuel Gasification, Advances in Chemistry Series, 69, 1967.
11. Zabolotny, E. R., et al, Stone and Webster Engineering Corp., Report to
EPRI, November 1974.
148
-------
SECTION VIII
LIST OF PUBLICATIONS
1. Hoke, R. C., Bertrand, R. R., "Pressurized Fluidized Bed Combustion of
Coal," Institute of Fuel Symposium, Series No. 1: Fluidised Combustion
Vol. 1. London, September 1975.
2. Hoke, R. C., "Emissions from Pressurized Fluidized Bed Coal Combustion,"
Proceedings of the Fourth International Conference on Fluidized Bed
Combustion, Washington, DC, December 9-11, 1975.
3. Hoke R. C., "FBC Particulate Control Practice and Future Needs: Exxon
Miniplant," Symposium on Particulate Control in Energy Processes,
San Francisco, May 11-13, 1976.
4. Nutkis, M. S., "Operation and Performance of the Pressurized FBC
Miniplant," Proceedings of the Fourth International Conference on
Fluidized Bed Combustion, McLean, Virginia, December 9-11, 1975;
pp. 221-238.
5. Ruth, L. A., "Combustion and Desulfurization of Coal in a Fluidized Bed
of Limestone," Fluidization Technology, v. II, D. L. Keairns, ed.,
Hemisphere Publ. Corp., Washington, DC 1976, pp. 321-27.
6. Ruth, L. A., "Regeneration of CaSO^ in Fluidized Bed Combustion,"
Proceedings of the Fourth International Conference on Fluidized Bed
Combustion, McLean, Virigina, December 9-11, 1975; pp. 425-38.
7. Nutkis, M. S., "Pressurized Fluidized Bed Coal Combustion,"
Fluidization Technology, V. II, D. L. Keairns, ed., Hemisphere
Publ. Corp., Washington, DC 1976, pp. 329-337.
149
-------
PATENT MEMORANDA SUBMITTED
1. Ruth, L. A., Design of Gas Distributors for Small Fluidized Beds.
150
-------
SECTION IX
APPENDICES
TABLES PAGE
A. ANALYTICAL TECHNIQUES 152
B. MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY 153
C. DETERMINATION OF S02 AND S03 BY WET CHEMISTRY 168
D. ENTRAINMENT RATES FOR GROVE NO. 1359 LIMESTONE
WITH LIMITED CALCINATION 170
E. ENTRAINMENT RATES FOR GROVE NO. 1359 LIMESTONE
WITH EXTENSIVE CALCINATION 171
F. ENTRAINMENT RATES FOR PFIZER NO. 1337 DOLOMITE 172
G. PARTICLE SIZE DISTRIBUTION - MINIPLANT USED
LIMESTONE NO. 1359 SORBENT 173
H. PARTICLE SIZE DISTRIBUTION - MINIPLANT USED
DOLOMITE NO. 1337 SORBENT 174
I. PARTICLE SIZE DISTRIBUTION - MINIPLANT SECONDARY
CYCLONE CAPTURE 175
J. PARTICLE SIZE DISTRIBUTION - MINIPLANT FLUE GAS
PARTICULATES 177
K. MINIPLANT SOLIDS ANALYSES 178
L. MINIPLANT SOLIDS COMPOSITION 186
M. SUMMARY OF BATCH COMBUSTOR OPERATING CONDITIONS 190
N. SUMMARY OF BATCH COMBUSTOR EMISSIONS DATA 193
0. BATCH FLUIDIZED BED COMBUSTOR CO EMISSIONS 195
P. BATCH COMBUSTOR PARTICLE SIZE DISTRIBUTION -
OVERHEAD SAMPLES 196
Q. BATCH COMBUSTOR BED AND OVERHEAD SOLIDS ANALYSIS 197
R. SULFUR BALANCES FOR BATCH COMBUSTOR 199
S. CALCIUM BALANCES FOR BATCH COMBUSTOR 200
151
-------
Analysis of Solids
TABLE A. ANALYTICAL TECHNIQUES
-2
+2
u +2 u
Mg carbon
Solids from combustion runs were analyzed for 804 ", C0^~^t Ca
and total sulfur. The analytical techniques that were used are described
below.
SO
CO
-2
-2
Ca
+2
+2
Mg
Total Sulfur -
The sample was treated with acidic BaCl2 solution.
The BaS04 precipitate was weighed.
HC1 was added to an acidified sample. The solution
was stripped with N£ and the gas passed through
drierite, CuS04 and ascarite. C03~2 was determined
from the,weight gain of the ascarite.
The sample was digested by heating vigorously in a
medium of perchloric acid/nitric acid. The determination
of Ca and Mg was made by atomic absorption.
The sample was mixed with sodium peroxide and a catalyst.
The sulfur was converted to the sodium sulfate. The
sample was heated above the melting point and the melt
was extracted with water. The sulfur was converted to
barium sulfate, precipitated and weighed.
Samples were combusted within a packed tube in an oxygen
atmosphere. Helium was used to sweep the combustion gases
into the analytical system. Carbon dioxide was determined
by differences in thermal conductivity.
Analysis of Flue Gas by Wet Chemical Methods
SO- — The amount absorbed by an 80% isopropanol solution was
determined titrimetrically using 0.01N barium perchlorate
as the titrant and thorin as the indicator.
S02 - 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.
Carbon
152
-------
TABLE B. MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Ul
U)
Operating Conditions:
Run Length, hrs.
Pressure, kPa
Lower Bed Temperature, °C
Avg. Bed Temperature, °C
Superficial Velocity, m/sec
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-Calculated
Excess Air, %
Sorbent
Coal
19.2 (7/31/75) 19.3 (8/4/75) 19.4 (8/20/75) 19.5 (8/21/75)
10.5
930
889
860-885
1.9
0.71
1.45
83-130
1.45
—
4.4-8.9
GL
CH
6
930
897
865-878
1.9
1.45
1.58
136-140
1.45
2.48
5-7
GL
CH
10
920-930
889
870-885
1.98-2.0
1.58
1.98
132-144
2.5
2.65
~15
GL
CH
6
930-940
901
885-900
2
1.98
1.75
136-144
2.5
2.68
~15
GL
CH
Flue Gas Emission:
so2,
NO ,
CO?
2,
ppm
ppm
ppm
C0
02,
Results:
S02 Retention,
Ca Sulfation,
Lb S02/M BTU
Lb NOX/M BTU
GL = Grove Limestone (BCR No, 1359)
CH = Champion Coal
160-815
208-87.5
50-150
7.2-10.5
6.4-1.7
715-780
78
—
12.75
1-1.4
510
125
75-100
12-13
2.5-3.0
537
91
50
11.4-12
2.7-3.5
20.4
56.4
22.7
1.26
0.10
69
26
0.92
0.16
67
25
0.97
0.12
-------
TABLE B (Continued). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Operating Conditions:
Run Length, hrs.
Pressure, kPa
Lower Bed Temperature, °C
Avg. Bed Temperature, °C
Superficial Velocity, m/sec
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-Calculated
Excess Air, %
Sorbent
Coal
19.6 Test 1
(8/26/75)
7.5
930
—
880-888
2.01-2.04
1.58
—
131-143
2.5
3.2
15
GL
CH
19.6 Test 2
(8/26/75)
6
930
—
820-830
1.93-1.95
__
1.60
120-131
2.5
3.2
22
GL
CH
19.7 Test 1
(8/28/75)
3.5
930
892
877-882
1.98
1.96
—
131-139
2.5
4
9.1
GL
CH
19.7 Test 2
(8/28/75)
6
970
967
935-940
2.16-2.20
—
1.27
139-147
2.5
4
9.6
GL
CH
19.9 Test 1
(9/11/75)
6.5
930
894
870
1.68-1.9
1.09
—
106-125
2.5
2.56
10
GL
CH
19.9 Test 2
(9/11/75)
6
930
943
926
2.2-2.26
—
1.88
131
2.5
2.56
20
GL
CH
Flue Gas Emission;
S02> ppm
NOX, ppm
CO, ppm
C02, %
00, %
500
104
—
11.7-12.3
2.5-3.0
620
127
— .
11.1
3.5-4
450
102.5
63
10.2
1.7
180-340
141
132
10.2
1.8
560
120-150
100
14.5-15.2
2
300
90-120
100
13. '6-14
4
Results:
S02 Retention, %
Ca Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
GL = Grove Limestone (BCR No. 1359)
CH = Champion Coal
68
21
0.95
0.14
58
18
1.23
0.81
73
—
0.85
0.14
83.5
20.9
0.19
66
1.31
0.18
81.9
32
0.85
0.16
-------
TABLE B (Continued). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Ln
Operating Conditions:
Run Length, hrs.
Pressure, kPa
Lower Bed Temperature, °C
Avg. Bed Temperature, °C
Superficial Velocity, m/sec
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-Calculated
Excess Air, %
Sorbent
Coal
20.1 Test 1
(9/17)
6
930
—
870
2.06
1.85
—
127
3.3
—
11
GL
CH
20.1 Test 2
(9/17)
2
930
—
827
1.83
—
—
116
3.3
—
12
GL
CH
20.1 Test 3
(9/17)
2.5
930
—
927
2.32
—
1.32
136
3.3
—
18.5
GL
CH
20.2 Test 1
(9/22)
9
930
947
927
2.3
1.22
—
127 \
1.58|
— ;
18
GL
CH
20.2 Test 2
(9/22)
2.3
930
947
927
2.3
—
1.65
127
3.3
3.75
13.7
GL
CH
Flue Gas Emission:
S02,
NOX,
CO,
co,
ppm
ppm
ppm
720
90-145
14.5-15.1
2.1
725
95
2.25
200-350
130-185
-100
—
3.0
420
200
130
14.5-14.75
2.5
200
195
130
14.5-14.75
2.5
Results:
SO2 Retention,
Ca Sulfation, ;
Lb S02/M BTU
Lb NO /M BTU
X
GL = Grove Limestone (BCR No. 1359)
CH = Champion Coal
54
1.45
0.17
53
1.5
0.14
71.2
1.16
0.33
86.3
23
0.71
0.32
-------
TABLE B (Continued). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Operating Conditions:
Run Length, hrs.
Pressure, kPa
Lower Bed Temperature, °C
Avg. Bed Temperature, °C
Superficial Velocity, m/sec
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-Calculated
Excess Air, %
Sorbent
Coal
Flue Gas Emission:
SOo, Ppm
NOX, ppm
CO, ppm
co2, %
02, %
Results:
S02 Retention, %
Ca Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
21 Test 1
(9/22)
8.5
930
870
2.0
1.6
—
127-130
3.3
—
14
GL
CH
600
195
100
14.2
2.5
60
—
1.18
0.28
21 Test 2
(9/22)
4.5
930
930
2.2
—
2.11
127-130
3.3
_t_
18
GL
CH
0-230
210
100
16.3
3.2
—
—
—
0.31
22 (10/2)
10
930
870
1.90
2.11
1.19
113
0
—
28
GL
CH
600-1040
—
30
13
4.25-4.75
—
—
—
— —
23 (10/3)
7
930
885
1.72-1.92
1.19
1.85
-120
1.3
—
12
TD
CH
810-270
100-115
100
12.75
2.25
—
—
—
0.15
25 (10/9)
5.5
930
894
970
1.75
1.37
1.78
-120
1.3
—
8
PD
CH
245
125
200
14.5
1.0-1.75
—
—
—
0.17
26 Test 1
(10/14)
11.25
930
949
927
2.1
1.12
—
130
3.7
2.75
9.5
GL
CH
140
185
130
12.75
1.8
90.8
33
0.28
0.26
26 Test 2
(10/14)
4.25
930
—
885
1.90
—
2.28
130
3.7
2.75
11.5
GL
CH
300
180
130
13
2.15
81.2
29
0,56
0.24
GL = Grove Limestone (BCR No. 1359)
TD « Tymochtee Dolomite
PD = Pfizer Dolomite (BCR No. 1337)
CH = Champion Coal
-------
TABLE B (Continued). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
i-1
Ln
-•J
Operating Conditions:
Run Length, hrs
Pressure, kPa
Lower Bed Temperature, °C
Avg. Bed Temperature, °C
Superficial Velocity, m/sec
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-Calculated
Excess Air, %
Sorbent
Coal
Flue Gas Emission;
S02» PPm
NOX, ppm
CO, ppm
C02, %
02, %
Results;
SO2 Retention, %
Ca Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
27.1 (10/27-28)
13.5
930
902
880
2.01
135
0.5
0.8
18.9
PD
CH
650
207
83
14.5
3.3
55
69
1.17
0.27
27.2 (10/28)
14
930
956
927
2.15
27.3 (10/28-29)
10
930
902
881
1.98
139
0.5
0.8
15.4
PD
CH
612
199
88
15.8
2.7
57.1
70
1.12
0.27
133
0.5
0.84
15.3
PD
CH
630
176
60
11
2.8
55.6
66
1.16
0.24
27.4 (10/29)
13
930
843
829
.1.07
123
0.5
0.91
20.5
PD
CH
532
200
64
14.3
3.6
62
68
1.05
0.28
PD = Pfizer Dolomite (BCR No. 1337)
CH = Champion Coal
-------
TABLE B (Continued). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
00
Operating Conditions:
Run Length, hrs.
Pressure, kPa
Lower Bed Temperature, °C
Avg. Bed Temperature, °C
Superficial Velocity, m/sec
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-Calculated
Excess Air, %
Sorbent
Coal
27.5 (10/29-30)
27
930
844
829
1.72
122
2.5
2.2
12.1
PD
CH
27.6 (10/30-31)
8.5
930
843
829
1.72
123
0.8
14.5
PD
CH
27.7 (10/31)
11
930
847
835
1.72
132
1.5
1.85
12.4
PD
CH
27.8 (10/31)
7
930
898
878
1.83
134
1.5
1.85
12.6
PD
CH
Flue Gas Emission;
NOX, ppm
CO, ppm
C02» %
02, %
Results:
S02 Retention, %
Ca Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
PD = Pfizer Dolomite (BCR No. 1337)
CH = Champion Coal
34
189
81
13.8
2.3
97.8
44
0.06
0.27
306
143
77
16.3
2.6
80
0.56
0.19
47
135
106
16.6
2.3
97.2
53
0.08
0.17
28
128
81
16.6
2.3
98
53
0.05
0.16
-------
TABLE B (Continued). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Operating Conditions; 27.9 (10/31) 27.10 (11/1) 27.11 (11/1-2) 27.12 (11/2) 27.13 (11/2)
Run Length, hrs. 3 17 11 10.5 8
Pressure, kPa 930 930 930 930 930
Lower Bed Temperature, °C 948 900 952 842 898
Avg. Bed Temperature, °C 922 883 931 834 888
Superficial Velocity, m/sec 2.09 1.87 2.23 1.94 2.02
Settled Bed Height, m
Initial
Final
Expanded Bed Height, m -W5 ~5 ~7 ~7 -v/6
Coal Feed Rate, kg/hr 142 135 149 133 134
Ca/S Molar Feed Ratio-Set 1.5 0.75 0.75 0.75 0.35
Ca/S Molar Feed Ratio-Calculated 2.1 1.03 1.0 1.03 0.6
Excess Air, % 13.1 7.8 11 23 13.5
Sorbent PD PD PD PD PD
Coal CH CH CH CH CH
Flue Gas Emission:
S02, ppm 24 455 444 400 822
NOX, ppm 164 93 109 130 110
CO, ppm 75 109 74 74 64
C02, % 17.4 15.1 16.4 15.1 16
02, % 2.4 1.5 2 3.9 2.5
Results:
S02 Retention, % 98.5 72 70 71 46
Ca Sulfation, % 47 70 70 70 76
Lb S02/M BTU 0.04 0.79 0.79 0,76 1.54
Lb NOX/M BTU 0.21 0.11 0.14 0.17 0.15
PD = Pfizer Dolomite (BCR No. 1337)
CH - Champion Coal
-------
TABLE B (Continued). MINIPLANT FLUIDIZED BED GOAL COMBUSTION RUN SUMMARY
_ Operating Conditions: _
Run Length, hrs.
Pressure, kPa
Lower Bed Temperature, °C
Avg. Bed Temperature, °C
Superficial Velocity, m/sec
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-Calculated
Excess Air, %
Sorbent
Coal
Flue Gas Emission:
27.14 (11/2-3)
11
930
898
892
2.04
NO
pptn
X, ppm
CO , ppm
C02, %
Results;
S02 Retention, %
Ca Sulfation, %
Lb S02/M BTU
Lb NO../M BTU
137
0.75
0.
14.
PD
CH
568
100
70
16
2.7
64
92
1.05
0.13
27.15 (11/3-4)
24.5
930
896
889
1.88
-v5
136
1
0.97^
8.2
PD
CH
465
103
107
14.7
1.6
71
73
0.86
0.13
27.16 (H/4)
6.5
930
844
841
1.74
-5
129
0.75
1.12
10.4
PD
CH
204
94
29
11
2
87
—
0.36
0.12
27.17 (11/4)
8
930
899
891
1.97
-6
136
0.75
1.12
13.8
PD
CH
454
95
34
14
2.5
72
64
0.81
0.12
27.18 (11/5)
5
930
919
909
2.09
-7
143
1
—
8.2
PD
CH
249
97
37
16.9
1.7
84.5
—
0.45
0.12
PD = Pfizer Dolomite (BCR No. 1337)
CH = Champion Coal
-------
TABLE B (Continued). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Operating Conditions;
Run Length, hrs.
Pressure, kPa
Lower Bed Temperature, °C
Avg. Bed Temperature, °C
Superficial Velocity, m/sec
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-Calculated
Excess Air, %
Sorbetit
Coal
Flue Gas Emissions;
S02, ppm
NOX, ppm
CO, ppm
COo, %
o2, %
Results;
S02 Retention, %
Ca Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
PD = Pfizer Dolomite (BCR No. 1337)
CH = Champion Coal
27.19 (11/5)
7
930
915
907
2.08
137
1.5
1.5
12.3
PD
CH
150
101
45
15.3
2.3
90
0,28
0.13
27.20 (11/5)
10.5
930
897
895
2.03
138
0
—
12.1
PD
CH
1272
91
34
12.9
2.3
11
2.33
0.12
27.21 (11/6) 28.1 (12/15) 28.2 (12/15)
2.5
930
10.5
930
897
895
2.03
13.5
930
844
846
1.67
4.25
930
—
840
2.13
1290
65
56
15.8
2
14
2.24
0.08
1.02
975
134
45
10
6.2
10
2.7
0.27
870
2.16
"5
125
0
10.9
PD
CH
105
0
40
Alumina
CH
109
0
35
Alumina
CH
1013
145
45
11
5.4
13.5
2.68
0.27
-------
TABLE B (Continued). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
ON
NJ
Operating Conditions;
Run Length, hrs.
Pressure, kPa
Lower Bed Temperature, °C
Avg. Bed Temperature, °C
Superficial Velocity, m/sec
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-Calculated
Excess Air, %
Sorbent
Coal
Flue Gas Emission:
S02, ppm
NOX, ppm
CO, ppm
C02, %
Results:
S02 Retention, %
Ca Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
28.3 (12/15) 28.4 (12/16)
2.5
930
920
2.13
114
0
18
Alumina
CH
1163
125
45
11
3.1
2.76
0.22
2
930
920
1.40
91
0
s—~
4.5
Alumina
CH
1575
41
11
16
0.9
1.6
3.08
0.06
28.5 (12/16/75)
2.5
930
—
930
2.65
—— „
9.7
123
0
—
21
Alumina
CH
1088
146
11
12
3.6
3.6
—
2.95
0.28
29 (1/15/76)
7.3
930
887
875
2.16
0.8
1.6
138
3.7
2.8
19
GL
CH
623
137
167
14.6
3.4
55.7
19.9
1.28
0.2
30.1 (1/27/76)
7
920
889
885
2.15
1.6
—
133
3.7
1.5
13.7
GL
CH
750
50
175-200
14.9
2.5
49.7
32.9
1.57
0.08
GL = Grove Limestone (BCR No. 1359)
CH = Champion Coal
-------
TABLE B (Continued). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Operating Conditions;
Run Length, hrs.
Pressure, kPa
Lower Bed Temperature, °C
Avg. Bed Temperature, °C
Superficial Velocity, m/sec
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-Calculated
Excess Air, %
Sorbent
Coal
Flue Gas Emission:
S02, ppm
NO , ppm
CO, ppm
h- *
30.2 (1/27-28) 30.3 (1/28) 30.4 (1/28) 31 (2/4-5) 32.1 (2/10)
C00,
%
Results;
SO 2 Retention, I
Ca Sulfation, %
Lb S02/M BTU
Lb NCL./M BTU
8.5
920
945
929
2.5
137
3.7
3.3
17.2
GL
CH
137
15.1
3.1
89.1
27
0.31
10
920
886
885
2.1
130
3.7
1.65
13.9
GL
CH
480
15.3
2.5
64.5
39.1
1.0
6
920
833
835
1.9
2.26
120
3.7
1.1
16.1
GL
CH
32.1
30
1.92
0.10
12
520
838
1.96
1.6
1.35
82
2.5
3
23
GL
CH
894
70
—
14.4
2.8
525
55
200
12.7
4.4
66
22
0.96
0.07
8.5
600
950
2.2
1.4
87
2.5
3.2
13
GL
CH
330
80
175
13.5
2.5
76.4
24
0.67
0.12
GL = Grove Limestone (BCR No. 1359)
CH = Champion Coal
-------
TABLE B (Continued). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Operating Conditions;
Run Length, hrs.
Pressure, kPa
Lower Bed Temperature, °C
Avg. Bed Temperature, °C
Superficial Velocity, m/sec
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-Calculated
Excess Air, %
Sorbent
Coal
32.2 (2/10-11)
11.3
600
947
2.1
32.3 (2/11)
33 (3/8)
34 (3/11)
35 (3/17)
89
0.75
1.5
16
PD
CH
8.25
600
— -
836
1.5
——
0.9
68
0.75
1.5
20
PD
CH
14
920
890
918
1.56
0.76
2.29
4.6
95
2.5
—
18.1-28.7
PD
CH
13.25
932
868
900
1.5
2.29
2.29
4.6
90
0.75
0.9
20.9
PD
CH
6.5
930
904
905
1.54
2.26
2.26
4.4
80
0.75
0.9
45.8
PD
CH
Flue Gas Emission:
SO~, ppm
NOX, ppm
CO, ppm
co2, %
°2» %
Results;
S02 Retention, %
Ca Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
600
106
200
14.3
2.9
60
40
1.14
1.44
CH = Champion Coal
PD = Pfizer Dolomite (BCR No. 1337)
*Water on Sampling Line Affected Readings.
490
89
200
14.3
3.4
-0
65-230
200
13.1-15
3.1-4.6
*100-300
52
200
15.5
3.5
442
72
225
13
6.5
66
44
0.96
0.12
61.6
69
1.1
0.13
-------
TABLE B (Continued). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Operating Conditions:
Run Length, hrs.
Pressure, kPa
Lower Bed Temperature, °C
Avg. Bed Temperature, °C
Superficial Velocity, m/sec
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-Calculated
Excess Air, %
Sorbent
Coal
36.1 (3/24)
6.5
930
906
900
1.6
1.4
2.
80
0,
1.
42
PD
CH
75
3
36.2 (3/24)
5
930
902
913
3
37 (4/7)
13.5
930
899
902
2.96
1.68
1.7
3
110
0.75
1.3
96
PD
CH
2.13
4
110
0.75
0.5
92
PD
CH
38.1 (4/14)
7.3
930
898
891
2.08
1.14
2.9
90
0.75
0.61
47 (1)
PDU'
CH
38.2 (4/14)
6
930
899
894
2.09
3.6
90
0.75
0.63
44
CH
Flue Gas Emission;
S02, ppm
N0x, ppm
CO,
C02,
02,
ppm
380
94
200
11.8
6
Results:
S02 Retention, %
Ca Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
CH = Champion Coal
PD = Pfizer Dolomite (No. 1337)
(1) 14 X 25 Mesh Sorbent Particle Size
(2) 8 X 14 Mesh Sorbent Particle Size
65.6
0.98
0.17
202
159
180
8.8
10
440
121
50
9.75
10
761
135
175
8.5
6.6
75.6
58
0.71
0.39
46
90
1.53
0.3
40
66
2.29
0.29
690
127
225
8
6.3
45
72
2.08
0.27
-------
TABLE B (Continued). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Operating Conditions;
Run Length, hrs.
Pressure, kPa
Lower Bed Temperature, °C
Avg. Bed Temperature, °C
Superficial Velocity, m/sec
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-Calculated
Excess Air, %
Sorbent
Coal
Flue Gas Emission:
so2,
NOX,
CO,
co2,
°2»
ppm
ppm
ppm
Results;
S02 Retention, %
Ca Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
CH = Champion Coal
PD = Pfizer Dolomite (BCR No. 1337)
GL = Grove Limestone (BCR No. 1359)
38.3 (4/14-15)
6
930
758
762
1.85
— —
3.76
76
0.75
0.94
94
PD
CH
398
125
325
6.47
7.68
48
51
1.42
0.32
38.4 (4/15)
8.25
930
758
762
1.84
—
4.75
78
1.5
1.14
94
PD
CH
223
149
375
6.6
10
71
62
0.78
0.37
38.5 (4/15)
1.5
930
681
690
1.7
—
4.75
75
1.5
1.49
112
PD
CH
180
114
800
5.91
10.9
76
51
0.64
0.29
38.6 (4/15)
2
930
678
684
1.3
2.26
4.56
67
1.5
1.43
78
PD
CH
163
102
400
7.08
9.1
82
57
0.5
0.23
39.1 (5/19)
11
902
756
750
1.52
1.22
2.25
74
2.5
68
GL
CH
912
120
350
9.56
8.37
0
2.69
0.26
-------
TABLE B (Continued). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Operating Conditions; 39.2 (5/19-20) 39.3 (5/20) 39.4 (5/20-21)
Run Length, hrs. 6.6 15 5
Pressure, kPa 902 932 932
Lower Bed Temperature, °C 673 946 944
Avg. Bed Temperature, °C 674 938 938
Superficial Velocity, m/sec 1.4 1.56 1.58
Settled Bed Height, m
Initial
Final — — 1.88
Expanded Bed Height, m 2.4 3.55 3.76
Coal Feed Rate, kg/hr 73 80 81
Ca/S Molar Feed Ratio-Set 2.5 2.5 2.5
Ca/S Molar Feed Ratio-Calculated — 3.5 3.8
Excess Air, % 60 32m 30(2)
Sorbent GL GL GL
Coal CH CH CH
Flue Gas Emission;
S02, ppm 900 329 345
NOX, ppm 174 125 115
CO, ppm — 100 100
C02, % 9.37 12.86 12.93
02, % 7.7 4.9 4.68
Results;
S02 Retention, % 0.6 70.3 68.7
Ca Sulfation, % — — 20
Lb S02/M BTU 2.64 0.80 0.84
Lb NOX/M BTU 0.37 0.22 0.2
CH = Champion Coal
GL = Grove Limestone (BCR No. 1359)
(1) 8 X 14 Mesh Sorbent Particle Size
(2) 14 X 25 Mesh Sorbent Partical Size
-------
ON
oo
Run No.
.2
,2
,2
,3
19.
19.
19.
19.
19.4
19.4
19.5
19.6
19.6
19.7
19.7
19.9
20.1
21
21
21
21
21
22
22
23
26
26
26
26
26
27,
27,
27,
27,
27.4
27.4
27.5
27.5
TABLE C. DETERMINATION OF S02
Sampling Wet Chemistry Analysis
Location S02 (ppm)
Third Deck 138
>
445
679
692
257
445
606
445
475
346
173
751
712
346
415
0
50
356
495
850
326
188
198
59
129
208
340
470
550
390
450
70
f 40
SO 3 (ppm)
24
95
64
11
2
6
164
5
10
6
18
9
4
4
6
0
12
0
262
5
1
0
0
0
0
0
59
94
31
19
18
2
0
AND S03
BY WET CHEMISTRY
U.V. Monitored
Value S02 (ppm)
180
505
732
750
500
500
500
500
500
450
200
680
730
460
460
0
105
250
1000
1040
305
250
162
140
220
300
390
660
630
450
510
75
30
Ratio of Wet Chemistry
and U.V. Analysis
for S02 (xlQO)
77
88
93
92
51
89
121
89
95
77
87
110
98
75
90
100
48
142
50
82
107
75
122
42
59
69
87
71
87
87
88
93
133
-------
TABLE C (Continued).
DETERMINATION OF SO AND SO BY WET CHEMISTRY
Run No.
27.6
- 27.7
27.10
27.11
27.11
27.12
27.13
27.14
27.15
27.16
27.17
27.17
27.19
27.20
27.20
27.21
34
34
35
35
38.1
38.1
38.2
38.2
38.3
38.3
39.1
39.2
39.2
39.4
Calibration Gas
Calibration Gas
Calibration Gas
Calibration Gas
Sampling v '
Location
Third Deck
Control Room
Third Deck
Control Room
Third Deck
Control Room
Third Deck
Control Room
Control Room
Control Room
Control Room
Control Room
Wet Chemistry Analysis
SO 2 (ppm)
50
20
180
0
190
210
710
410
350
90
320
400
70
790
1240
1150
297
129
119
150
534
623
227
563
365
395
890
910
327
405
306
270
1318
1532
S03 (ppm)
8
7
52
30
7
30
29
30
—
39
23
23
10
14
35
45
0
0
—
20
31
16
0
0
0
0
0
0
0
0
—
—
_._
—
U.V. Monitored
Value SQ2 (ppm)
240
39
90
360
330
720
480
450
60
375
420
180
900
1280
1200
266
266
349
407
755
825
675
623
420
400
912
900
450
290
290
1320
1490
Ratio of Wet Chemistry
and U.V. Analyses
for SO? (xlOO)
21
51
53
64
99
85
78
150
85
95
39
88
97
96
112
48
34
37
71
76
34
90
87
99
96
101
90
105
93
100
103
(1) Third Deck = At the flue gas sampling port on the off-gas line prior to the sample preparation system
Control Room - Prior to the continuous analyzers and after sample preparation
-------
TABLE D. ENTRAINMENT RATES FOR GROVE NO.
1359 LIMESTONE WITH LIMITED CALCINATION
Entrainment
Superficial Volume % Losses Wt. Percent
Minlplant
Run No.
19.2
19.3
19.4
19.5
26
29.1
30.1
30.3
31
39.1
39.2
Ca/S
(Mole/Mole)
1.45
1.45
2.65
2.7
2.75
2.8
1.5
1.6
3.0
2.5
2.5
Velocity
(m/sec)
1.9
1.9
2.0
2.0
2.0
2.16
2.15
2.1
1.96
1.52
1.4
Bed Entrained
Per Hr.
2.8
0.9
0.8
2.3
1.6
1.5
1.2
1.2
1.3
1.0
1.1
1.0
1.6
1.0
0.5
0.8
0.8
0.4
0.4
0.4
Ca/S Equlv.
(Mole/Mole)
0.40
0.18
0.20
0.50
0.40
0.25
0.21
0.22
0.24
0.21
0.22
0.20
0.38
0.24
0.15
0.25
0.25
0.13
0.12
0.11
Input Ca Lost
by Entralnment
27 •)
12 I
8 f
19 (
14 J
9 -.
14
15
16
13
14
13
13
8
6
10
10
5
5
4 J
Remarks
Based on flyash
sample represen-
tative of the
entire run.
Based on flyash
samples taken at
one-hour intervals
Avg. 1.1 + 0.6
0.24 + 0.1
12 + 5
Pressure - 930 kPa except Run 31 which was at 520 kPa
Temperature - 835°C - 900°C except Run 39.1 at 760°C and Run 39.2 at 680°C.
-------
TABLE E. ENTRAINMENT RATES FOR GROVE NO. 1359
LIMESTONE WITH EXTENSIVE CALCINATION
Miniplant
Run No.
19.7
20.1
21
30.2
32.1
39.3
39.4
(Mole/Mole)
4.0
3.3
3.3
3.3
3.2
2.5
2.5
Superficial
Velocity
(m/sec)
2.1
2.06
2.1
2.5
2.2
1.56
1.58
Volume %
Bed Entrained
Per Hr.
3.3
5.8
3.6
1.3
1.2
0.9
1.6
1.3
0.5
0.4
0.4
0.4
0.4
0.3
Entraintnent
Losses
Ca/S Equiv.
(Mole/Mole)
0.60
1.1
0.80
0.32
0.30
0.24
0.46
0.38
0.20
0.15
0.17
0.15
0.14
0.12
Wt. Percent
Input Ca Lost
by Entrainment
15 1
32
25 J
10 -x
9
7
14
12
8
6
7
6
6
5
J
Remarks
Based on a flyash
sample representative
of the entire run
Avg. 0.8 + 0.5
0.24 + 0.11
8 + 3
Based on flyash
samples taken at
one-hour intervals
Runs 19.7, 20.1 and
21 are omitted from
average
Pressure - 930 kPa except Run 32.1 which was 600 kPa
Temperature - 950°C
-------
I—
TABLE F. ENTRAINMENT RATES FOR PFIZER NO. 1337 DOLOMITE
Volume Percent Entrainment Losses
Miniplant
Run No.
27.8
27.10
27.13
27.14
27.16
27.17
32.2
32.3
Ca/S
(Mole/Mole)
1.8
1.03
.6
.7
1.1
1.1
1.5
1.5
Bed Entrained
Per Hr.
5.5
1.95
2.4
1.5
3.1
5.5
5.6
3.6
3.6
3.9
Ca/S Equiv.
(Mole/Mole)
0.75
0.24
0.34
0.21
0.46
0.82
1.1
0.7
0.65
0.71
Wt. Percent Input
Ca Lost by Entrainment Remarks
40
23
57
30 Based on flyash
samples taken at one
to three hour
73 intervals
73
47
58
47
Avg.
3.7
0.6
80
Pressure - 930 kPa except runs 32.2 and 32.2, which were at 600 kPa
Superficial Velocity - 1.5-2.1 m/sec
Temperature - 835-950°C
-------
TABLE G. PARTICLE SIZE DISTRIBUTION -
MINIPLANT USED LIMESTONE NO. 1359 SORBENT
Particle Size (ym),
At Indicated Points in Distribution
CO
Run No.
19.3
19.3
19.4
19.4
19.5
19.6
19.7
19.7
19.9
19.9
20.1
20.1
20.2
20.2
21
21
22
26
26
29
30.1
30.2
30.3
30.4
39
Material
Final Bed (1)
Rejected Solids (2)
Final Bed
Rejected Solids
Rejected Solids
Rejected Solids
Final Bed
Rejected Solids
Final Bed
Rejected Solids
Final Bed
Rejected Solids
Final Bed
Rejected Solids
Final Bed
Rejected Solids
Final Bed
Final Bed
Rejected Solids
Final Bed
Rejected Solids
Rejected Solids
Rejected Solids
Final Bed
Final Bed
10%
650
1000
700
880
910
940
770
800
290
880
670
740
710
780
690
750
790
280
410
510
680
750
950
730
630
25%
1100
1340
1000
1080
1110
1050
950
980
480
1150
800
1000
880
970
820
930
1010
440
730
890
960
980
1160
950
850
50%
1490
1820
1340
1420
1490
1330
1220
1250
800
1630
1020
1360
1270
1310
1140
1280
1320
860
1100
1200
1220
1220
1480
1180
1140
75%
2100
2200
1820
1900
1920
1750
1630
1700
1330
2040
1330
1840
1560
1780
1650
1650
1770
1240
1620
1730
1620
1530
1960
1580
1580
90%
2460
2460
2180
2270
2190
2120
2010
2080
1860
2300
1800
2200
1960
2140
2040
2180
2110
1710
2040
2170
2080
1920
2260
2000
1980
(1) Sorbent in Combustor at End of Run
(2) Sorbent Removed from Combustor During Run
-------
TABLE H. PARTICLE SIZE DISTRIBUTION -
MINIPLANT USED DOLOMITE NO. 1337 SORBENT
Particle Size (ym),
Run No. Material
25 Final Bed (1)
27.3 Rejected Solids(2)
27.5 Rejected Solids
27.8 Rejected Solids
27.10 Rejected Solids
32.3 Final Bed
32.3 Rejected Solids
34 Bed 1st Analysis
34 Bed 2nd Analysis
35 Final Bed
36.2 Final Bed
37 Final Bed
38.6 Final Bed
At Indicated
10%
820
640
340
170
260
530
440
160
170
250
400
240
170
25%
1100
960
960
360
570
720
590
230
230
430
760
500
250
Points in Distribution
50%
1560
1430
1580
1120
1220
960
800
450
460
830
1420
980
480
75%
1960
1920
1920
1760
1830
1280
1000
960
1000
1380
1930
1480
1070
90%
2200
2180
2140
2020
2120
1780
1260
1400
1480
1920
2260
1860
1700
(1) Sorbent in Combustor at End of Run
(2) Sorbent Removed from Combustor During Run
-------
TABLE I. PARTICLE SIZE DISTRIBUTION -
MINIPLANT SECONDARY CYCLONE CAPTURE
Particle Size (ym) At Indicated Points in Size Distribution
Run No.
19.2
19.3
19.4
19.5
19.6
19.7
19.9
20.1
20.2
21
22
23
25
26
27.1
27.2
27.3
27.4
27.5
27.8
27.9
27.10
27.11
27.13
27.14
27.15
27.16
27.17
10%
7
7
4
6
6
8
5
8
7
8
6
5
7
5
6
5
6
5
6
6
17
5
5
5
6
6
5
5
20%
13
11
6
9
11
13
7
15
10
15
9
7
12
8
8
7
8
7
9
9
30
7
9
7
10
9
7
7
30%
20
16
9
13
15
20
10
44
13
25
12
9
17
11
11
9
11
10
12
12
46
10
12
9
15
12
10
10
40%
25
22
13
18
20
32
12
46
17
30
17
11
25
14
14
11
14
13
16
16
72
12
16
12
19
16
12
12
50%
30
33
19
25
25
48
14
50
22
37
24
15
38
17
18
14
18
17
21
21
110
16
20
15
24
20
16
16
60%
48
50
28
36
35
71
17
52
30
44
33
22
56
22
22
19
22
21
28
28
160
22
24
19
28
25
21
21
70%
61
64
45
55
50
100
23
54
46
54
50
35
80
31
27
25
27
27
37
37
216
31
31
27
35
30
27
27
80%
91
95
76
90
80
159
33
57
90
80
75
59
120
46
37
35
37
37
51
51
290
44
43
40
46
40
48
48
90%
150
160
140
180
145
300
52
60
290
180
150
110
190
88
56
56
56
58
62
62
320
70
70
70
68
58
54
54
-------
TABLE I (Continued). PARTICLE SIZE DISTRIBUTION -
MINIPLANT SECONDARY CYCLONE CAPTURE
Particle Size (\M) At Indicated Points in Size Distribution
Run No.
28.1
28.2
28.3
28.4
28.5
29.1
30.1
30.2
30.3
30.4
31
32.1
32.2
32.3
10%
4
4
5
5
5
5
5
5
6
5
5
7
7
5
20%
7
7
7
8
8
9
8
9
10
7
8
9
10
7
30%
11
10
10
12
11
10
11
11
14
8
10
11
12
9
40%
13
13
13
15
14
16
14
14
19
10
12
13
14
11
50%
17
17
17
20
18
21
18
18
26
12
14
16
16
13
60%
21
22
22
26
24
29
25
21
33
16
18
20
20
16
70%
30
30
31
35
32
40
36
29
42
21
23
26
28
22
80%
42
41
45
50
49
53
52
43
54
34
36
43
43
35
90%
66
65
72
75
75
77
78
74
74
64
64
78
72
64
-------
••J
-g
TABLE J. PARTICLE SIZE DISTRIBUTION -
MINIPLANT FLUE GAS PARTICULATES
Run
Number
31
32.2
32.3
33
36.2
37
Particle Size (ym) , At Indicated Points in Distribution
5%
1.43
1.48
1.75
2.23
1.7
1.5
10%
1.83
1.95
2.3
3.3
2.5
2
25%
3. '05
3.1
3.3
4.75
4.1
3.25
50%
8.0
5.8
5.9
7.55
7.5
6.5
75%
23.5
12
13
13.5
14.3
14.5
90%
-50
24
32.5
25.5
30
30
95%
-60
37.2
44
41
43
43
-------
TABLE K. MINIPLANT SOLIDS ANALYSES
oa
Run
No._ Source
19.2 Final Bed (1)
Second Cyclone (2)
Rejected Solids(3)
19.3 Final Bed
Second Cyclone
Rejected Solids
19.4 Final Bed
Second Cyclone
Rejected Solids
19.5 Final Bed
Second Cyclone
Rejected Solids
19.6 Final Bed
Second Cyclone
Rejected Solids
19.7 Final Bed
Second Cyclone
Rejected Solids
Weight Percent
24
9 3
~ • J >
32.2
25.6
4.3,
23.3
33.6
4.77
34.8
33.8
9.86
33.4
36.4
10.1
35.0
45.2
16.8
40.1
Ca
9.65, 10.0
, 35.9
8.16
, 32.2
, 32.5, 34.5
, 7.34, 7.64
, 33.7
, 8.01, 9.76
, 47.5
, 17.1
S
4.88
1.40
5.92
7.77
2.89
7.7
4.3,
4.17, 4
3.6, 3.
5.3, 5.
3.7, 4.
2.2, 2.
3.78
3.7, 3.
2.3, 2.
4.8
7.3, 5.
2.8, 2.
5.97
.7
09
89
46
38
64
26
72
91
S04
22.6, 22.7,
22.97, 23.0
5.96
18.24
20.8
8.4
20.4
18.65
8.97
22.08
19.74
6.08
18.88
15.16
7.34.
19.58
21.79
8.27
20.45
0)3
20.89
1.72
6.37
29.8
.66
30.2
35.38
1.69
32.22
31.98
1.40
33.62
40.45
2.37
33.87
0.68
1.38
12.18
Total C
N.A. (4)
22.52, 22.61
N.A.
6.25
27.9
5.67
7.73
24.90
6.59
6.46
22.38
6.92
8.0
25.85
6.86
.37
17.30
2.87
(1) Spent Sorbent Removed from Combustor After End of Run
(2) Solids Captured in Secondary Cyclone
(3) Spent Sorbent Removed from Combustor During Run
(4) Not Analyzed
-------
TABLE K (Continued). MINIPLANT SOLIDS ANALYSES
Weight Percent
•vl
VO
Source
19.9 Final Bed
Secondary Cyclone
Rejected Solids
20.1 Final Bed
Secondary Cyclone
Rejected Solids
20.2 Final Bed
Secondary Cyclone
Rejected Solids
21 Final Bed
Secondary Cyclone
Rejected Solids
22 Final Bed
Secondary Cyclone
23 Final Bed
Secondary Cyclone
25 Final Bed
Secondary Cyclone
26 Final Bed
Secondary Cyclone
Rejected Solids
27.1 Final Bed
Sec. Cyclone #1
Sec Cyclone '#2
37.8,
13.2,
36
47.5,
11.9,
33.2
40.2,
13.8,
33.2,
45.2,
17.7,
37
Ca
39.5
12.8
45.7
13.8
37.8
15.4
59.3
46.7
17.4
30, 31.7, 39.5
14.7,
21.7
10.2
25.7
17.1
26.6,
13.9,
32.9
19.4
7.1,
6.1
12.5, 13.9
35.4, 32.5
11.9
6.04, 6.71
S
8.0, 10.5
3.74
9.09
8.11
4.72
5.98
7.41
7.61
7.98
7.48
3.77
6.87
N.A.
4.11
N.A.
3.93
5.95
2.76
9.3
4.82
11.6
13.2
3.2
3.42
S04
C03
29.4
11.46
25.5
23.68
9.98
17.9
22.4
12.28
18
20.95
10.91
11.08
30.7
11.46
29.57
10.76
17.18
7.24
27.34
8.4
24.6
40.05
9.17
9.28
2.37
1.19
35.7
.51
2.51
13.1
.78
1.98
13.1
23.1
4.3
16
18.8
2.74
16.78
1.22
28.5
14.9
19.02
4.2
5.65
1.44
0.79
0.58
Total C
1.17
17.44
4.64
N.A.
11.5
N.A.
N.A.
7.07
N.A.
N.A.
12
N.A.
N.A.
18.93, 18.45
N.A.
15.24
N.A.
16.2
N.A.
13.06
2.03
.76
19.70
5.75
-------
TABLE K (Continued), MINIPLANT SOLIDS ANALYSIS
Weight Percent
Run
No.
27.2
27.3
27.4
27.5
H-'
00
0 27.8
27.9
27.10
27.11
27.13
27.14
27.15
Source
Secondary Cyclone
Rejected Solids
Secondary Cyclone
Rejected Solids
Secondary Cyclone
Rejected Solids
Secondary Cyclone
Rejected Solids
Secondary Cyclone
Rejected Solids
Secondary Cyclone
Rejected Solids
Secondary Cyclone
Rejected Solids
Secondary Cyclone
Secondary Cyclone
Rejected Solids
Secondary Cyclone
Rejected Solids
Rejected Solids
Ca
8.7, 8.07
29.5, 28.9
4.8, 4.7
28.6, 24.2, 31.9
3.5, 3.58
29.9, 24.3
12.0, 5.65, 8.65
19.8, 21.9
14.1, 26.6
22.8, 20.7
16.1, 17.8
23.8, 21.7
6.4, 5.48, 6.72
20.9, 24.6
6.1, 9.08, 8.65
7.7, 7.44
19.6, 12.8
6.2, 4.44, 5.7
21.6, 22
23.4, 20.2
S
3.7
10.5
2.2
8.7
1.98
10.3
5.83
8.4
7.0
8.9
7.4
8.3
2.9
10.31
3.9
1.3
9.44
7.28
12.2
11.0
S04
14.63
30.26
6.71
29.83
4.95
30.34
15.48
26.11
9.94
28.94
22.10
25.41
16.47
32.54
13.66
10.57
32.54
15.43
38.13
31.76
C03
.82
11.83
.03
21.55
.63
13.28
5.48
16.08
4.13
14.45
2.69
16.66
0.99
4.46
0.93
.07
3.70
0.41
6.56
12.62
Total C
16.20
N.A.
16.55
N.A.
20.94
2.78
12.06
2.26
9.23
.99
9.26
4.61
18.00
2.17
9.15
14.74
2.08
12.22
2.42
N.A.
-------
TABLE K (Continued). MINIPLANT SOLIDS ANALYSIS
Run
No.
27.16
27.17
28.1
28.2
28.3
28.4
29
30.1
Source
30.2
Secondary Cyclone
Final Bed
Secondary Cyclone
Rejected Solids
Final Bed
Secondary Cyclone
Secondary Cyclone
Secondary Cyclone
Secondary Cyclone
Final Bed
Secondary Cyclone
Rejected Solids
Sec. Cyclone #1
Sec. Cyclone #2
Sec. Cyclone #3
Rejected Solids
Sec. Cyclone #1
Sec. Cyclone #2
Sec. Cyclone #3
Rejected Solids
Weight Percent
Ca
8.33, 8.87
19.4
15.7, 15.7, 14.9
19.5
N.A.
N.A.
N.A.
N.A.
N.A.
32.4
6.74
N.A.
5.40
5.68
6.05
30.0
9.70
10.0
9.31
42.3
S
2.38
13.2
6.4
10.0
3.1
0.53
1.15
0.55
0.50
5.3
3.9
N.A.
2.73
2.80
3.0
7.89
2.90
3.26
3.2
8.8
S04
14.90
40.05
19.05
30.10
N.A.
N.A.
N.A.
N.A.
N.A.
15.90
8.47
18.13
6.25
7.26
8.64
23.9
8.78
9.4
9.28
27.25
C03
N.A.
1.44
1.18
1.95
N.A.
N.A.
N.A.
N.A.
N.A.
34.59
0.97
N.A.
.78
.81
.55
28.55
.61
.94
.65
.94
Total C
12.81
.76
5.17
2.13
N.A.
9.28
13.22
7.16
6.47
7.46
11.89
4.43
16.98
16.57
17.04
5.30
8.90
11.44
7.72
.34
-------
TABLE K (Continued). MINIPLANT SOLIDS ANALYSIS
Weight Percent
00
ho
Run
No.
30.3
30.4
32.1
32.2
32.3
Source
Ca
Sec. Cyclone #1
Sec. Cyclone //2
Sec. Cyclone #3
Rejected Solids
6.20
5.63
6.27
30.2
3.02
2.5
2.33
9.06
7.08
8.46
6.35
28.19
.85
.28
.55
23.49
12.88
15.86
16.16
4.14
Final Bed 33.8
Sec. Cyclone #1 5.97
Sec. Cyclone #2 5.27
Sec. Cyclone #3 5.34
Rejected Solids 31.9
Final Bed 38.4
Sec. Cyclone #1 7.41
Sec. Cyclone #2 6.22
Rejected Solids 33.2
Sec. Cyclone #1 12.5
Sec. Cyclone #2 12.0
Sec. Cyclone #3 11.9
Rejected Solids 40.1
Sec. Cyclone #1 18.5
Sec. Cyclone #2 16.7
Sec. Cyclone #3 16
Final Bed 28.8
Sec. Cyclone #1 10.3
Sec. Cyclone #2 11.6
Sec. Cyclone #3 7.63
3.43
2.6
2.27
2.4
7.54
6.30
2.7
2.2
6.83
2.3
1.72
1.8
9.7
4.8
6.67
6.8
23.05
20.24
7.62
6.88
21.81
7.14
5.96
7.00
7.6
23.1
6.2
5.93
4.76
8.5
3.48
4.08
2.98
17.4
17.2
14.3
29.37
10.3
13.5
8.5
44.75
.71
.75
.45
31.89
27.90
1.34
1.27
12.11
1.46
.69
.63
2.80
3.34
1.72
1.76
14.78
2.25
1.12
1.80
7.22
17.68
18.40
21.17
N.A.
5.63
17.69
19.19
2.93
7,
5.
5,
13
56
68
.96
3.50
3.42
3.58
3.04
13.54
10.50
13.65
33
Secondary Cyclone
22.5
5.3
15.14
5.10
1.53
-------
TABLE K (Continued). MINIPLANT SOLIDS ANALYSIS
Weight Percent
oo
OJ
Run
No.
34
35
36.1
36.2
37
38.1
38.2
Source
Ca
Final Bed 24.6
Sec. Cyclone #1 7.69
Sec. Cyclone #2 7.54
Sec. Cyclone #3 8.15
Sec. Cyclone #1 7.28
Sec. Cyclone #2 7.77
Sec. Cyclone #3 7.29
Final Bed 26.3
Sec. Cyclone #1 17.0
Sec. Cyclone #3 15.0
Final Bed 21.5
Sec. Cyclone #1 12.9
Sec. Cyclone #2 11.0
Sec. Cyclone #3 11.7
Sec. Cyclone #1 6.56
Sec. Cyclone #2 6.69
Sec. Cyclone #3 5.51
Sec. Cyclone #1 5.92
Sec. Cyclone #2 7.01
Sec. Cyclone #3 7.49
S04
C03
13.5
3.7
3.8
3.6
3.0
3.9
3.7
12.0
8.0
6.7
14.3
8.02
7.07
7.40
3.66
3.4
3.0
3.22
3.45
3.78
40.81
12.68
10.64
11.10
10.06
10.50
9.70
36.54
22.28
20.17
46.34
24.12
20.70
24.88
10.75
9.68
8.97
9.53
12.29
12.84
7.12
.38
.96
.18
.89
.70
.74
11.55
1.67
1.80
2.30
.81
1.01
.71
1.72
1.24
1.15
.93
.94
1.02
Total C
Final Bed
Sec. Cyclone #1
Sec. Cyclone #2
Sec. Cyclone #3
22.2
13.6
12.7
13.1
10.2
7.0
5.7
5.9
33.52
20.21
18.03
17.70
.52, .70
.46
.54
.57
N.A.
2.20
4.06
3.11
N.A.
6.09
,25
39
6.90
6.85
6.27
N.A.
2.10
3.68
N.A.
2.35
2.69
3.78
5.10
4.52
5.34
.19
.48
4.89
-------
TABLE K (Continued). MINIPLANT SOLIDS ANALYSIS
Run
No.
38.3
38.4
38.5
2 38.6
39.1
39.2
39.3
Source
Sec. Cyclone #1
Sec. Cyclone #2
Sec. Cyclone #3
Sec. Cyclone #1
Sec. Cyclone #2
Sec. Cyclone #3
Secondary Cyclone
Final Bed
Secondary Cyclone
Sec. Cyclone #1
Sec. Cyclone $1
Sec. Cyclone #3
Rejected Solids
Sec. Cyclone #1
Sec. Cyclone #2
Sec. Cyclone #3
Rejected Solids
Sec. Cyclone #1
Sec. Cyclone #2
Sec. Cyclone #3
Rejected Solids
Ca
9.01
9.34
8.68
10.1
11.1
12.4
12.8
22.7
12.5
7.84
7.85
7.94
32.9
6.25
6.11
6.19
31.9
8.71
8.39
8.17
40.7
Weight Percent
S
4.63
4.94
5.9
5.46
5.60
5.74
5.70
10.7
5.64
3.19
2.95
2.97
2.10
2.65
2.37
2.47
2.81
2.83
2.74
2.50
6.67
504
CO 3
Total C
14.48
15.39
14.74
16.34
16.70
16.52
15.69
2.98
2.88
2.66
6.10
6.85
7.09
12.82
17.90
12.01
12.55
11.22
12.11
12.83
14.76
10.7
5.64
3.19
2.95
2.97
2.10
2.65
2.37
2.47
2.81
2.83
2.74
2.50
6.67
35.05
17.23
8.22,
8.09,
8.47,
5.82,
7.81,
7.40,
7.26,
8.27,
8.42,
8.17,
6.97,
19.52
8.62
7.60
8.13
2.86
7.19
7.66
7.14
4.94
8.37
8.12
7.58
, 17.30
14.87
4.40
2.12,
2.28,
2.56,
45.52
1.07,
1.03,
1.07,
31.15
.72,
• 54,
.53,
2.49
2.48
2.49
2.71
, 51.91
1.20
1.14
1.14
.66
.64
.58
N.A.
12.49
12.15
15.58
15.37
10.23,
5.27
6.36
6.40
9.54,
4.16
3.77
5.10
.97
9.44
9.52
-------
Run
No.
39.4
Source
Final Bed
Sec. Cyclone #1
Sec. Cyclone #2
Sec. Cyclone #3
Rejected Solids
TABLE K (Continued)-. MINIPLANT SOLIDS ANALYSIS
Weight Percent
504
Ca
47.0
8.47
7.69
7.74
49.6
S
7.05
2.46
2.45
2.60
6.87
C03
Total C
22.63
7.75, 7.59
7.53, 7.36
8.54, 7.36
21.72, 21.68
0.45
.30, .31
.19, .32
.49, .38
2.44
<0.3
4.53
4.88
4.42
0.67
00
-------
TABLE L. MINIPLANT SOLIDS COMPOSITION
Sorbent Portion
Run
Number
19.2
19.3
19.4
19.5
19.6
19.7
19.9
20.1
20.2
21
Composition
Source
Final Bed (1)
Reject. Solids (2)
Sec. Cyclone (3)
Final Bed
Reject. Solids
Sec. Cyclone
Final Bed
Reject. Solids
Sec . Cyclone
Final Bed
Reject. Solids
Sec . Cyclone
Final Bed
Reject. Solids
Sec. Cyclone
Final Bed
Reject. Solids
Sec. Cyclone
Final Bed
Reject. Solids
Sec. Cyclone
Final Bed
Reject. Solids
Sec. Cyclone
Final Bed
Reject. Solids
Sec. Cyclone
Reject. Solids
C
-4
-1
22
0.3
-0.4
28
1
0.2
25
0.1
0.2
22
-0.1
0.1
25
.2
.4
17
1
-3
17
.2
3
11
1
-2
7
-2
11
Ash
37
36
58
18
29
56
19
9
60
14
13
60
11
10
54
20
18
50
17
5
54
20
26
56
32
12
61
30
52
(Wt. %)
Sorbent
67
65
20
82
72
16
80
91
15
86
87
18
89
90
21
80
82
33
83
98
29
83
71
33
67
89
32
73
37
Composition (Mole %)
CaO
4
63
62
7
-23
37
-14
12
-2
11
9
58
-0.4
12
50
78
59
76
66
4
57
77
51
65
71
65
56
59
58
CaC03
58
13
12
65
87
7
86
62
24
64
67
11
81
65
17
1
20
5
4
66
6
1
26
10
2
19
9
29
16
CaSO^
38
24
26
28
37
56
28
27
78
25
24
31
19
23
33
21
21
19
30
30
37
22
23
25
27
16
35
13
26
(1) Spent Sorbent Removed from Combustor After Run
(2) Spent Sorbent Removed from Combustor During Run
(3) Solids Captured in Secondary Cyclone
186
-------
TABLE L (Continued). MINIPLANT SOLIDS COMPOSITION
Sorbent Portion
Run
Number
22
23
25
26
27.1
27.2
27.3
27.4
27.5
27.8
27.9
27.10
27.11
27.13
27.14
27.15
27.17
Source
Final Bed
Sec. Cyclone
Final Bed
Sec . Cyclone
Sec. Cyclone
Final Bed
Reject. Solids
Sec. Cyclone
Sec. Cyclone
Reject. Solids
Sec. Cyclone
Reject. Solids
Sec. Cyclone
Reject. Solids
Sec. Cyclone
Reject. Solids
Sec. Cyclone
Reject. Solids
Sec. Cyclone
Reject. Solids
Sec. Cyclone
Reject Solids
Sec. Cyclone
Sec. Cyclone
Reject. Solids
Sec. Cyclone
Reject. Solids
Sec. Cyclone
Reject. Solids
Reject. Solids
Sec. Cyclone
Composition
C
1
18
1
15
13
1
1
12
15
0,3
16
-0.2
17
0.1
21
-1
11
-2
8
1
9
1
18
9
1
15
1
12
1
2
5
Ash
13
51
20
61
34
19
28
60
68
17
59
13
71
25
70
20
60
14
52
11
47
16
59
68
27
66
15
67
16
26
57
(Wt. %)
Sorbent
87
31
79
24
53
81
71
28
17
82
25
89
12
75
10
81
29
88
40
88
44
83
23
23
71
20
84
21
83
72
38
Composition (Mole %)
CaO
25
52
-9
48
24
23
57
51
33
30
21
5
41
5
30
3
-15
0.1
66
5
35
27
-22
21
1
41
-1
-23
1
29
43
CaCOs
37
13
52
8
58
40
12
22
8
27
7
51
0.4
39
12
48
42
44
14
49
11
13
11
8
15
1
22
5
39
7
5
38
35
57
44
18
36
31
27
60
43
73
44
59
56
58
49
74
56
20
47
54
60
111
72
84
58
79
118
61
64
52
187
-------
TABLE L (Continued). MINIPLANT SOLIDS COMPOSITION
Sorbent Portion
Run
Number
29
30.1
30.2
30.3
30.4
31
32.1
32.2
32.3
33
34
35
36.1
Composition
Source
Final Bed
Sec. Cyclone
Reject. Solids
Sec. Cyclone
Reject. Solids
Sec. Cyclone
Reject. Solids
Sec. Cyclone
Final Bed
Reject. Solids
Sec. Cyclone
Final Bed
Reject. Solids
Sec. Cyclone
Part. (1)
Reject. Solids
Sec. Cyclone
Sec. Cyclone
Part.
Final Bed
Sec. Cyclone
Part.
Sec. Cyclone
Part.
Sec. Cyclone
Sec. Cyclone
Sec. Cyclone
C
1
12
-0.4
17
0.2
9
-0.6
15
-2
2
19
0.1
1
18
7
0.4
6
3
5
0.1
12
2
1
3
3
7
7
Ash
16
71
18
69
17
69
18
71
14
11
68
9
26
65
79
22
71
45
75
10
57
72
28
67
50
65
68
(Wt. %)
Sorbent
84
17
83
15
83
22
83
15
88
87
13
91
74
17
14
77
23
52
20
90
31
26
72
30
47
28
26
Composition (Mole %)
CaO
8
38
3
38
72
56
9
43
-0.4
3
47
29
48
43
11
71
72
50
63
23
43
65
57
65
37
34
37
CaC03
71
10
64
8
2
5
52
7
88
67
8
49
24
13
14
5
5
10
3
34
11
4
15
10
2
4
7
CaSO^
21
52
33
54
27
39
39
50
12
30
49
22
27
44
76
24
23
40
33
43
45
32
28
25
61
62
57
(1) Flue Gas Particulate
188
-------
TABLE L (Continued). MINIPLANT SOLIDS COMPOSITION
Sorbent Portion
Run
Number
36.2
37
38.1
38.3
38.4
38.5
38.6
39.1
39.2
39.3
39.4
Composition
Source
Final Bed
Sec. Cyclone
Part.
Final Bed
Sec. Cyclone
Part.
Sec. Cyclone
Sec. Cyclone
Sec. Cyclone
Sec. Cyclone
Final Bed
Sec. Cyclone
Reject Solids
Sec. Cyclone
Reject. Solids
Sec. Cyclone
Part.
Reject. Solids
Sec. Cyclone
Final Bed
Reject. Solids
Sec. Cyclone
C
2
2
4
1
3
4
5
14
11
12
-0.2
12
0.1
14
3
6
4
0.5
4
0.9
0.2
5
Ash
4
44
73
13
48
79
68
50
43
34
7
41
12
66
18
79
80
20
77
14
16
78
(Wt. %)
Sorbent
98
53
23
87
49
17
27
36
46
54
93
48
88
20
79
16
16
79
19
85
84
18
Composition (Mole %)
CaO
13
38
30
3
14
35
20
10
-0.9
-18
-8
19
2
36
35
38
-40
79
56
79
77
57
CaC03
29
7
6
7
5
2
9
21
39
67
44
24
93
21
57
12
3
4
5
0.6
4
3
CaSO^
58
55
64
90
81
63
71
69
61
51
64
58
5
43
8
50
137
17
39
20
20
40
189
-------
TABLE M. SUMMARY OF BATCH COMBUSTOR OPERATING CONDITIONS
Run
No.
59
60
61
62
63
64
65
66
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
95
96
97
99
Sorbent
1R
Bed from
Bed from
1R
1R
Bed from
1R
1R
2R
Bed from
Bed from
2R
2R
Bed from
Bed from
Bed from
2R
Bed from
2R
2R
1C
1C
1C
1R
Bed from
Bed from
1R
59
60
63
71
72
75
76
77
79
95
96
Pressure,
kPa
805
805
805
395
805
805
805
310
815
821
821
821
811
815
811
811
811
811
811
426
821
821
811
815
822
821
820
Superficial
Velocity
1
1
1
1
1
1
1
1
1
0
0
1
1
1
1
0
1
0
1
1
1
1
1
1
1
1
1
m/s
.53
.44
.08
.43
.17
.15
.19
.85
.11
.86
.87
.20
.29
.09
.05
.96
.08
.88
.36
.64
.04
.04
.00
.16
.13
.18
.18
Settled Bed
Depth, m
0.77
—
—
0.46
0.77
—
0.77
0.48
0.77
—
—
0.77
0.77
0.40
0.38
0.32
0.77
—
0.61
0.61
0.61
—
—
0.77
—
—
0.61
Bed
Temperature °C
Avg.
870
865
894
875
914
925
887
903
870
855
875
810
820
820
830
940
930
935
960
925
875
880
850
850
830
850
902
Range
840-895
820-895
861-905
840-896
850-952
895-959
804-906
876-915
845-890
840-860
860-890
790-821
770-840
775-840
805-840
915-960
900-957
930-950
950-975
900-960
850-914
850-920
800-890
822-873
794-862
812-890
870-890
Excess
Air, %
21
37
62
61
12
16
17
33
22
41
53
88
48
73
109
86
37
53
61
50
1
0
6
57
55
60
61
Coali1'
kg/hr
10.44
9.08
5.58
3.77
9.42
8.17
8.39
5.90
7.95
5.44
4.99
5.95
7.95
5.72
4.54
4.22
6.54
4.77
6.81
4.77
9.08
9.08
8.49
7.491
7.581
7.491
7.131
Run
Length,
hrs
3.00
3.00
2.75
4.00
4.50
2.00
4.00
4.00
2.00
2.00
1.50
1.75
2.00
2.00
2.00
1.50
3.75
2.00
2.00
2.00
2.00
1.75
2.00
2.00
1.50
2.00
1.50
-------
TABLE M (Continued). SUMMARY OF BATCH COMBUSTOR OPERATING CONDITIONS
Run
No.
100
101
102
103
104
105
106
107
Sorbent
Bed from 99
Bed from 100
1R
Bed from 102
3
Bed from 103
Bed from 105
Bed from 106
Pressure,
kPa
821
821
821
821
821
821
821
821
Superficial
Velocity
m/s
1.17
1.16
1.15
1.12
1.13
1.15
1.14
1.15
Bed
Settled Bed Temperature °C
Depth, m Avg.
882
883
0.61 838
825
833
850
850
843
Range
830-896
875-902
800-860
770-845
800-873
810-868
825-855
825-855
Excess
Air, %
70
104
119
120
129
114
157
153
Coal,
kg/hr
6.811
5.681
6.97W
6.81W
6.54W
7.04W
5.83W
6.04W
Run
Length,
hrs
1.50
2.00
5.50
2.00
0.58
2.50
3.00
5.08
-------
TABLE M (Continued) SUMMARY OF BATCH COMBUSTOR OPERATING CONDITIONS
NOTES;
Coal: Arkwright Mine, W. Va., 2.6% S, -16 mesh used unless otherwise noted.
S = Arkwright coal screened to remove fines less than 70 mesh.
W = Wyoming coal, 0.7% S.
I = Illinois coal, 4.1% S.
Stone: (1) = Grove limestone, 8 x 25 mesh (BCR No. 1359).
(2) = Tymochtee dolomite, 8 x 25 mesh.
(3) = Alundum
C = Calcined
R = Uncalcined
-------
TABLE N. SUMMARY OF BATCH COMBUSTOR EMISSIONS DATA
LO
NO (Average)
Run
No.
59
60
61
62
63
64
65
66
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
95
96
97
SO 2 (Average)
Ib N02
ppm
rr*u
190
199
191
220
175
151
231
208
n.m.
n will*
nfn
• 111 .
337
286
221
193
183
226
181
287
382
140
112
122
201
210
230
106 BTU
0.30
0.33
0.38
0.43
0.22
0.21
0.34
0.34
0.77
0.51
0.47
0.48
0.41
0.37
0.34
0.56
0.69
0.17
0.14
0.16
0.39
0.40
0.46
ppm
rr"
550
646
402
431
350
566
266
647
123
381
535
173
267
384
560
558
389
764
230
229
374
502
1031
1214
1576
1856
Ib S02
10b BTU
1.22
1.50
1.10
1.17
0.61
1.12
0.54
1.45
0.25
0.91
1.45
0.55
0.67
1.13
1.96
1.74
0.90
1.97
0.62
0.58
0.64
0.86
1.86
3.3
4.2
5.2
% S02
Reduction
68.5
62.0
69.0
69.5
82.6
70.5
85.6
61.6
93.3
76.0
63.5
85.6
82.4
70.5
47.7
54.0
76.4
47.7
83.5
84.6
83.1
77.5
51.5
52.3
39.2
26.0
SO? (Final)
ppm
778
722
498
761
940
792
604
911
151
647
649
270
382
576
754
654
825
955
345
267
411
670
1181
1721
1815
1952
Ib SO 2
10° BTU
1.72
1.68
1.36
2.06
1.78
1.56
1.24
2.05
0.31
1.54
1.68
0.85
0.95
1.69
2.63
2.04
1.90
2.46
0.94
0.68
0.71
1.15
2.14
4.7
4.9
5.4
% S02
Reduction
55.5
57.5
61.7
46.1
53.2
58.8
67.4
45.9
91.7
59.3
55.7
77.5
74.8
55.7
29.6
46.1
49.8
34.7
75.3
82.1
81.4
70.0
44.5
32.4
30.0
22.2
Combustion
% Ca Efficiency
Sulfation Ca/S (%)
33
36
32
11
28
39
28
18
24
21
62
22
23
35
42
46
33
43
28
20
21
29
40
6
8
14
1.7
1.6
In
.9
4.2
1f\
.9
1.5
2.4
2.6
3.9 ^
2.8 >
0.9 J
3.6
3.3^|
1.6 >
0.7 J
1.0
1.5 X
0.8 •>
2.7
4.1
3.9 "\
2.4 1
1.1 J
5'5~\
3.8 >
1.6 J
93.3
95.5
no o
yo ./
A ~7 1
y 7.1
%-t
. 7
Cl £. 1
96.1
%f\
.9
C\ 1 1
91.1
97.5
93.0
93.5
98.7
98.0
91.9
98.8
-------
TABLE N (Continued). SUMMARY OF BATCH COMBUSTOR EMISSIONS DATA
NO (Average)
Run
No.
99
100
101
102
103
104.
105
106
107
ppm
239
240
251
241
299
174
261
291
272
Ib NO?
10& BTU
0.48
0.51
0.63
0.65
0.81
0.49
0.69
0.92
0.84
SO? (Average)
ppm
335
983
1086,
0
0
135
39
19
19
Ib S02
10b BTU
0.9
2.9
3.8
0.0
0.0
0.5
0.1
0.1
0.1
% S02
Reduction
86.6
58.6
45.6
100.0,
100.0
58.1
88.4
93.3
93.6
S02
ppm
933
1222
1496
0
0
174
39
19
19
(Final)
Ib.SO?
10° BTU
2.6
3.6
5.2
0.0
0.0
0.7
0.1
0.1
0.1
% S02
Reduction
62.6
48.5
25.1
100.0
100.0
46.0
88.4
93.3
93.6
% Ca
Sulfation
16
18
28
10
8
9
13
12
Ca/S
3.8 ^)
2.7 >
0.9 J
10.0
12.8
9.9")
7.3
7.7 J
Combustion
Efficiency
(%)
97.7
99.9
n.m. = not measured because of problems with equipment
-------
TABLE 0. BATCH FLUIDIZED BED COMBUSTOR CO EMISSIONS
Run No. 3675- CO (ppm) Temperature (°C) Excess Air (%)
59 332 870 21
60 NA 865 37
61 201 894 62
62 287 875 61
63 164 914 12
64 118 925 16
65 170 887 17
66 360 903 33
71 104 870 22
72 167 855 41
73 167 875 53
74 154 810 88
75 134 820 48
76 125 820 73
77 NA 830 109
78 87 940 86
79 79 930 37
80 134 935 53
81 72 960 61
82 143 925 50
83 187 875 1
84 163 880 0
85 305 850 6
95 287 850 57
96 287 830 55
97 287 850 60
99 96 902 61
100 144 882 70
101 145 883 104
102 97 838 119
103 121 825 120
104 242 833 129
105 121 850 114
106 121 850 157
107 97 843 153
N.A. = Not Analyzed
195
-------
TABLE P. BATCH COMBUSTOR PARTICLE SIZE
DISTRIBUTION - OVERHEAD SAMPLES
Run Particle Weight Percent Finer Than
No. Slze(ym)
61 600
300
150
106
75
44
65 600
300
150
106
75
44
73 600
300
150
106
75
44
81 600
300
150
106
75
44
Cyclone 1
99.3
80.7
48.6
37.2
22.9
14.3
94.3
88.6
74.3
68.6
60.0
45.7
99.0
88.4
74.0
66.8
57.7
43.3
99.3
82.9
63.6
56.5
47.2
34.3
Filter
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
87.5
96.3
92.6
88.9
87.0
85.1
79.5
100.0
100.0
98.7
98.1
79.2
75.4
196
-------
TABLE Q. BATCH COMBUSTOR BED AND OVERHEAD SOLIDS ANALYSIS
Run
No.
59
60
61
62
63
64
65
66
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
95
96
97
99
100
101
102
103
Ca^
35.30
36.60
43.50
52.30
43.00
39.70
43.80
50.00
29.20
44.60
17.30
25.50
25.00
25.00
22.90
24.90
27.40
25.40
27.60
28.90
45.20
42.90
39.60
36.90
36.10
35.10
46.40
43.10
41.30
34.90
39.20
Bed
S04=
27.86
32.60
33.16
14.05
28.84
38.37
29.76
21.05
16.32
22.52
27.17
13.12
13.59
20.56
23.66
28.16
22.35
27.25
18.18
14.03
22.99
29.47
37.41
5.19
6.74
11.60
18.33
18.51
29.13
8.47
22.09
(7.35)
C03
19.98
8.97
3.13
2.39
2.93
2.17
6.01
1.34
2.35
0.95
1.27
NA
24.60
15.53
16.68
1.76
1.97
2.33
3.50
3.71
2.21
1.94
3.66
50.75
41.11
44.60
0.32
3.54
0.76
38.64
47.49
Total Ca
35.46
34.02
11.32
20.00
27.69
34.81
25.11
52.51
NA
14.52
NA
12.83
NA
26.68
NA
8.91
15.97
NA
NA
NA
NA
50.74
NA
NA
8.11
NA
NA
9,55
NA
2.85
Cyclone
Total S
2.96
3.75
9.31
2.42
1.40
2.06
2.39
2.62
NA
3.60
NA
5.00
NA
5.30
NA
NA
3.00
NA
NA
NA
NA
3.10
NA
NA
3.44
NA
NA
NA
3.37
4.27
1
soiT
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
8.64
NA
NA
9.68
NA
12.34
co-T
2.25
0.66
4.07
2.21
1.07
0.50
0.58
2.52
NA
3.32
NA
NA
NA
4.02
NA
2.83
6.53
NA
NA
NA
NA
NA
NA
NA
0.96
NA
NA
3.87
NA
7.33
Filter
Total C
27.15
15.86
NA
NA
13.88
16.10
15.17
24.44
NA
11.13
NA
13.42
NA
30.31
NA
8.91
10.78
NA
NA
NA
NA
44.92
NA
NA
11.23
NA
NA
9.26
NA
Total S
3.95
4.60
NA
NA
3.50
4.73
4.41
3.43
NA
5.40
NA
4.10
NA
3.10
NA
NA
4.50
NA
NA
NA
NA
2.90
NA
NA
8.20
NA
NA
8.14
NA
SO&
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
21.85
NA
NA
25.69
NA
COq"
0.52
2.24
NA
NA
7.21
0.12
0.84
0.83
NA
5.70
NA
NA
NA
2.94
NA
NA
6.87
NA
NA
NA
NA
NA
NA
NA
0.07
NA
NA
0.53
NA
3.13
6.38
3.82
8.03
-------
TABLE Q (Continued). BATCH COMBUSTOR BED AND OVERHEAD SOLIDS ANALYSIS
ivun
No.
105
106
107
Bed
Ca"*" S04=
51.50 10.98
46.10 14.20
49.10 14.08
CO-}
8.35
12.12
3.60
Cyclone
Total Ca Total S
NA NA
1.48 3.98
NA NA
1
S04~ C03~ Total C
NA NA NA
12.35 3.39 3.47
NA NA NA
Filtt
Total S
NA
5.30
NA
ir
S04
NA
16.86
NA
COs
NA
10.44
NA
CO
All values are weight percent
3. =
Includes carbon present as CO,,
-------
TABLE R. SULFUR BALANCES FOR BATCH COMBUSTOR
(All weights in kg)
Run No.
59 "I
60 }
61 )
62
64 /
65
66
n]
73J
74
?f\
77 J
8QJ
3
S}
97 J
QQ"V
100 s
10U
105^
106 i
10 7 )
Sulfur
In Coal
1.840
0.392
1.479
0.847
0.595
0.863
0.922
0.854
1.288
1.751
1.419
0.403
Flue
Gas (1)
0.636
0.120
0.308
0.122
0.230
0.173
0.286
0.281
0.378
1.006
1.060
0.000
Sulfur
Overhead
Solids
0.539
0.049
0.143
0.116
0.120
0.232
0.386
0.191
0.259
0.189
0.302
0.242
Out
Bed
0.477
0.177
0.900
0.775
0.290
0.411
0.377
0.454
1.132
0.351
0.466
0.144
Total
1.652
0.346
1.351
1.013
0.640
0.816
1.049
0.926
1.769
1.546
1.828
0.385
% Sulfur
Balance
90
88
91
120
108
95
114
108
137
88
129
96
(1) Based on average S02 concentration.
Average - 105%
Standard deviation - 17%
199
-------
TABLE S. CALCIUM BALANCES FOR BATCH COMBUSTOR
Run Charge
No. (stone,
145 TD,
2C TD,
3C G,
4C TD,
5C TD,
17C G,
23C TD,
95C~)
96C G,
97C J
99C~
100C
101C-
102C~
103C
105C
106C
107C_
G,
G,
7
7
7
7
7
7
7
11
9
9
kg)
.54
.54
.72
.54
.54
.72
.72
.35
.08
.08
Input
Wt. Bed
Recovered,
(kgCa)
1
1
3
1
1
3
1
4
3
3
.58
.58
.02
.58
.58
.02
.61
.44
.55
.55
2
2
4
0
4
5
4
9
4
3
kg
.63
.63
.72
.82
.65
.08
.27
.08
.99
.06
Wt.
fr.
Ca in Bed
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
220
256
433
285
265
460
295
351
413
491
Wt.
in
Ca
Bed
kg
0.
0.
2.
0.
1.
2.
1.
3.
2.
1.
58
67
05
23
23
34
26
19
06
50
Wt. Ca Wt. Overhead Wt. fr.
7
to
TD - Tymochtee Dolomite, 20.9 Wt. % Ca
G - Grove Limestone, 39.1 Wt. % Ca
Solids
Ca in Wt. Ca in Output Calcium
Recovered, kg Overhead Overhead (kgCa) Balance
5.04
4.20
4.23
4.72
1.81
0.170
0.101
0.090
0.218
0.081
0.86
0.42
0.38
1.03
0.15
0.24
0.39
0.94
2.04
,44
.09
2.43
,26
,38
,58
1.65
3.00
3.54
91
69
80
80
87
85
102
85
100
-------
TECHNICAL REPORT DATA
(Please read hizmictions on the reverse before completing)
. REPORT NO,
EPA-600/7-77-107
3. RECIPIENT'S ACCESSION- NO.
4. T,TLE AND SUBTITLE studies of the pressurized Fluidized-
Bed Coal Combustion Process
5. REPORT DATE
September 1977
6. PERFORMING ORGANIZATION CODE
•AUTHORts)R.C.Hoke, R.R. Bertrand, M.S.Nutkis,
D. D.Kinzler, L.A. Ruth, M.W.Gregory, and
E.M. Magee
8. PERFORMING ORGANIZATION REPORT NO
EXXON/GRU.16GFGS.77
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Exxon Research and Engineering Co.
P. O. Box 8
Linden, New Jersey 07036
10. PROGRAM ELEMENT NO.
E HE 62 3 A
11. CONTRACT/GRANT NO.
68-02-1312 and -1451
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Phase; 8/75-7/76
14. SPONSORING AGENCY CODE
EPA/600/13
is.SUPPLEMENTARY NOTES IERL-RTP project officer for this report is D. Bruce Henschel,
Mail Drop 61, 919/541-2825. EPA-600/7-76-011 is the previous EPA report relating to
this workr
The report gives results of studies of the environmental aspects of the pres-
surized fluidized-bed coal combustion process, using two experimental facilities: a
218 kg coal/hr "miniplant" continuous combustion/sorbent regeneration system (0. 63
MW equivalent), and a 13 kg coal/hr "batch" combustion unit. Combustion studies
were conducted to characterize the emissions of SO2, SOS, NOx, particulates, and CO
from the combustors as a function of combustion conditions. Operating results from
these combustion runs defined the dolomite and limestone sorbent feed rate required
to keep SO2 emissions within the current EPA New Source Performance Standards for
coal-fired utility boilers, considering the effects of coal sulfur content, sorbent type,
gas residence time in the bed, bed temperature, bed pressure, and excess air. NOx
emissions remained within the range 0.1-0.4 Ib/million Btu, compared to the current
standard of 0. 70 Ib/million Btu. Particulate emissions from the miniplant combustor,
after two stages of cyclones, ranged from 0. 8 to 4.2 Ib/million Btu (mass median
particle size of 7 microns), compared to the EPA standard of 0.10 Ib/million Btu;
thus an additional stage of particle control would be required. Emissions of SO3 and
CO were generally low. Shakedown of the miniplant regenerator was completed, cul-
minating with a 24-hr combustion/regeneration run during which sorbent was trans -
ferred continuously between the combustor and the regenerator. __
16. ABSTRACT
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Air Pollution
Desulfurization
Flue Gases
Limestone
Dolomite (Rock)
Calcium Oxides
Fluidized-Bed
Processors
Combustion
Regeneration
(Engineering)
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Held/Group
Air Pollution Control
Stationary Sources
Fluidized-Bed Combus-
tion
Limes tone-Based Desul-
furization Process
13 B
07A,07D
21B
08G
07B
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS {This Report)
Unclassified
212
20. SECURITY CLASS (Thispage)
Unclassified
EPA Form 2220-1 (9-73)
201
------- |