United States Industrial Environmental Research EPA-600/7-80-013
Environmental Protection Laboratory January 1980
Agency Research Triangle Park NC 27711
Miniplant and Bench
Studies of Pressurized
Fluidized-bed Coal
Combustion: Final Report
Interagency.
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-80-013
January 1980
Miniplant and Bench Studies
of Pressurized Fluidized-bed Coal
Combustion: Final Report
by
R. C. Hoke, E. S. Matulevicius, M. Ernst, J. L. Goodwin,
A. R. Garabrant, I. B. Radovsky, A. S. Lescarret, R. R. Bertrand,
L. A. Ruth, V. J. Siminski, M. S. Nutkis, M. D. Loughnane,
H. R. Silakowski, M. W. Gregory, and A. Ichel
Exxon Research and Engineering Co.
P. 0. Box 8
Linden, New Jersey 07036
Contract No. 68-02-1312
Program Element No. INE825
EPA Project Officer: D. Bruce Henschel
Industrial Environmental Research Laboratory
Office of Environmental Engineering and Technology
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
ABSTRACT
The pressurized fluidized bed combustion (PFBC) of coal and regeneration
of spent sorbent were studied in the continuous 480 Ib coal/hr (220 kg/hr)
"miniplant" and the 28 Ib coal/hr (13 kg/hr) bench PFBC units. The ability
of the PFBC system to reduce SOg emissions by 90& or more was demonstrated
using both dolomite and limestone sorbents. The dynamic response of SO?
emissions to sudden changes in the sulfur content of the coal, and in the
dolomite to coal feed ratio, was also measured. NOX control tests carried
out in the bench scale unit indicated that either two stage combustion or
ammonia injection could reduce NOx emissions below the already low levels
normally measured in PFBC.
Particulate emissions were studied using a number of particulate control
devices. A three stage cyclone system was shown to be unexpectedly efficient.
However, it could not consistently reduce particulate emissions to the level
required by the current New Source Performance Standard for utility boiler.
A high temperature/pressure ceramic fiber filter was successfully tested,
giving collection efficiencies high enough to meet current emission standards.
Further testing of a granular bed filter was unsuccessful due to poor removal
efficiencies and various other operating problems. Conventional low tempera-
ture/low pressure electrostatic precipitator and bag house systems were also
tested. Both were shown to be capable of reducing particulate emissions,
following the three stages of cyclones, to levels meeting the current New
Source Performance Standard.
Regeneration of spent sorbent was studied in the miniplant and bench
units. In the miniplant, a series of extended tests were completed which con-
firmed the operability of the system and the large reduction in fresh lime-
stone resulting from the use of regeneration. An approximate four fold reduc-
tion in fresh limestone feed requirements results from the use of regeneration.
S02 retentions were in excess of 90% at all times. No loss in sorbent
activity due to regeneration was observed. Bench scale tests indicated that
coal could probably be used as the fuel for regeneration in place of natural
gas used in the miniplant program.
A series of sampling campaigns was completed which generated samples used
by another contractor for Level I and Level II comprehensive analyses of
emissions from PFBC.
This report 1s submitted in fulfillment of Contract Number 68-02-1312 by
Exxon Research and Engineering Company under sponsorship of the Environmental
Protection Agency. Work was completed in August 1979.
111
-------
TABLE OF CONTENTS
Page
Abstract ill-
List of Figures v
List of Tables ix
Acknowledgements xii
Sections
I Summary 1
II Introduction 13
III Miniplant Combustion Studies 17
Equipment, Materials, Procedures 17
Experimental Results and Discussion 47
IV Particulate Measurement and Control 86
Particulate Measurement 87
Cyclone Studies 95
Ceramic Fiber Filter Evaluation 105
Granular Bed Filtration Studies 123
Conventional Particulate Control 147
V Regeneration Studies 148
Equipment and Procedures 148
Experimental Results and Discussion 151
VI Comprehensive Analysis of Emissions 168
Level I and II Comprehensive Analysis Tests 168
Presence of Mg3(CaS0) in 3rd Cyclone Flyash 169
VII Bench Unit Studies 171
Combustion Studies 171
Bench Regeneration Studies 206
VIII References 221
IX List of Publications 223
X Appendices 225
1v
-------
LIST OF FIGURES
Page
II-l Pressurized Fluidlzed Bed Coal Combustion System 14
II-2 Exxon Fluidized Bed Combustion Miniplant 16
III-l Exxon Fluidized Bed Combustion Miniplant 18
II1-2 Coal and Limestone Feed System 19
III-3 Cross Section of Coal/Sorbent Feed Line 21
III-4 Combustor Vessel 23
III-5 Miniplant Combustor Cooling Coll 1A After 1200 Hours 24
Operation (Close Up)
III-6 Miniplant Combustor Cooling Coil 1A After 1200 Hours 25
Operation
III-7 Cyclone Design Dimensions 27
III-8 Variable Pressure Reducing Nozzle 30
II1-9 Controlled Condensation Coil Attached to a S03-S02 32
Impinger Train
111-10 Miniplant Flow Schematic 33
III-ll Balston Filter Particulate Sampling System 34
111-12 Original HTHP Particulate Sampling System 36
111-13 HTHP Particulate Sampling System as Modified 37
111-14 Southern Research Institute 5-Cyclone Train 38
III-l5 University of Washington 7-Stage Impactor 39
111-16 Schematic of Alkali Probe Train 41
111-17 Coal Particle Size Distribution 43
111-18 Dolomite Size Distribution 46
111-19 Selected Combustor Temperatures During Crashdown and Restart 49
111-20 Miniplant Flow Schematic Hot Corrosion/Erosion 50
Test Configuration
-------
LIST OF FIGURES (Continued)
Page
111-21 6E Turbine Blade Specimens 51
111-22 Westinghouse Boiler Tube Probes 53
111-23 Natural Gas Injection - Temperature Profiles 54
111-24 SOo Retention Adjusted to 2 s. Gas Residence Time (tg). 57
Dolomite Runs Having Actual tg Between 1.5 and 2.5 s.
111-25 S02 Retention Adjusted to 3 s. Gas Residence Time (tg). 58
Dolomite Runs Having Actual tg Between 2.5 and 3.5 s.
111-26 First Order Rate Constant Vs. Calcium Utilization 60
Miniplant Dolomite Runs (Approximate Only)
II1-27 Measured and Calculated S02 Retention Vs. Ca/S for 61
Miniplant Dolomite Runs
111-28 Calculated S02 Retention Vs. Ca/S Miniplant Dolomite Runs 62
111-29 SOo Retention Vs. Ca/S Ratio for Limestone No. 1359 at 63
Calcining Conditions
111-30 Change in Coal Sulfur/S02 Emissions (Run 99) 67
111-31 Instantaneous SO? Response (Run 99) (Champion to 68
Illinois No. 6 Coal)
111-32 Instantaneous S02 Response (Run 99) (Illinois No. 6 to 69
Champion Coal)
111-33 Run 100 S02 Emissions (10 Minute Averages) 70
111-34 Run 100 Instantaneous S02 Response (Champion Coal) 71
111-35 Run 100 Instantaneous S02 Response (Champion Coal) 72
II1-36 Correlation of NOX Emissions 74
111-37 Schematic of the Alkali Probe Train 76
111-38 Particle Concentration in Cyclone Gas Outlets 79
111-39 Primary Cyclone Dipleg Size Distribution 81
111-40 Magnetization Curve for the Miniplant Flyash Sample at 25°C 82
v1
-------
LIST OF FIGURES (Continued)
Page
111-41 Effect of Temperature on the Magnetic Induction Force 83
of the Miniplant Flyash Samples
IV-1 Comparison of Particle Size Distribution Via Bahco Vs. 88
Coulter Counter of Secondary Cyclone Captured Material
IV-2 Comparison of Particle Size Distribution Via Bahco Vs. 89
Coulter Counter of Tertiary Cyclone Captured Material
IV-3 Balston Filter/Coulter Counter Vs. APT Impactor 92
Particle Size Distribution
IV-4 Effect of Temperature on IKOR Monitor Reading 94
IV-5 Tertiary Cyclone Collection Efficiency 99
IV-6 Basic Cyclone Design 102
IV-7 Filter Housing Pressure Vessel Cross Section 106
IV-8 Acurex Test Filter Installation Schematic 108
IV-9 Acurex HTHP Ceramic Bag Filter Site 109
IV-10 Acurex HTHP Ceramic Bag Filter Pressure Drop and Flow 111
IV-11 Acurex High-Temperature Ceramic Bag Gas Inlet Schematic 114
IV-12 Acurex HTHP Bag Filter Outlet Loading Vs. Face Velocity 117
(Averaged Over First 6 Hours of Exposure)
IV-13 Acurex Ceramic Bag Filter - Bag No. 5 Particulate 118
Penetration History
IV-14 Ceramic Filter Bag No. 3 120
IV-15 Ceramic Filter Bag No. 3 Closeup of Vacuumed Strip 121
IV-16 Ceramic Filter Bag No. 9 After Run 96 122
IV-17 Modified Granular Bed Filter Element (Without Shroud) 125
IV-18 Modified Filter Bed 126
IV-19 Granular Bed Filter Schematic 127
IV-20 Filter Bed with Internal Baffle 129
vii
-------
LIST OF FIGURES (Continued)
Page
IV-21 Plexiglas Model GBF 130
IV-22 Increase in Outlet Particulate Concentration with Time 135
IV-23 Exxon Mark IV Granular Bed Filter (Single Bed) 141
IV-24 Granular Bed Filter Installation Schematic 142
IV-25 Granular Bed Filter Flow and Pressure Drop Vs. Time 146
V-l Mlniplant Sol Ids Transfer System (Schematic) 150
V-2 S02 Retention Adjusted to 2 s Residence Time Vs. 156
Ca/S Ratio
V-3 Combustor Bed Age and S02 Emissions Vs. Time (Run 102) 158
V-4 Combustor Bed Age and S02 Emissions Vs. Time (Run 103) 159
V-5 Combustor Bed Age and S02 Emissions Vs. Time (Run 105) 160
V-6 Regenerator and Combustor S02 Emissions Run 105 165
VII-1 Schematic of Batch Combustor Unit 172
VII-2 Bench Combustor Showing Locations for Ammonia Injection 186
VII-3 Emission Indices for Ammonia Injection Program Showing 189
95% Confidence Intervals
VI1-4 Ammonia Injection Program: Effect of Excess A1r Level 195
on NOX Emissions
VI1-5 Response Time 196
VII-6 Emission Indices for Simulated Flue Gas Redrculation 199
(SFGR) Program Showing 95% Confidence Intervals
VII-7 Comparison of the Simulated Flue Gas Rec1rculat1on 202
(SFGR) Program with the Ammonia Injection Program
VII-8 Reduction 1n N0« Emissions by Using Two Control 208
Techniques Simultaneously
VII-9 Bench Regeneration Unit 211
VII-10 Fuel Injection Mode 213
-------
LIST OF TABLES
Page
III-l Coal Composition 44
II1-2 Sorbent Composition 45
II1-3 Run Summary 90% S02 Retention Study 56
111-4 S02 Dynamic Response Summary 65
III-5 Emission of Sodium, Potassium, Chlorine, Vanadium 77
in Flue Gas
III-6 Particulate Emission Particle Size Distribution 78
III-7 Particulate Concentration and Size Ranges 78
Representing a Number of Runs
IV-1 Particulate Concentration and Size Miniplant 95
Cyclone System
IV-2 Tertiary Cyclone Total Efficiency Summary 97
IV-3 Tertiary Cyclone Fractional Efficiency Summary 98
IV-4 Third Cyclone Operating Conditions 100
IV-5 Tertiary Cyclone Design Dimensions 101
IV-6 Summary of Cyclone Test Program Results 104
IV-7 Acurex HTHP Bag Filtration Summary 112
IV-8 Acurex HTHP Bag Filter Inlet Particulate Loadings 113
IV-9 Acurex HTHP Bag Filter Collection Efficiency 116
IV-10 Granular Bed Filter Run Summary 132
IV-11 Particle Size Distributions 137
IV-12 Granular Bed Filter Cleaning Program 143
IV-13 Comparison of Particulate Size Distribution of Material 143
Before and After Filter Test Element
IV-14 Run 115 Filter Test Summary 145
ix
-------
LIST OF TABLES (Continued)
Page
V-l Mini pi ant Regenerator Run Summary 153
V-2 Minlplant Fluidlzed Bed Coal Combustion Run Summary 154
V-3 Sorbent Elutr1at1on Losses 162
V-4 Composition of Regenerator Off Gas 163
V-5 Minlplant Regenerator Mass Balances 167
VII-1 Location of Gas Injection Points Bench Unit NO 174
Control Studies
VII-2 Designed Experiment for Study of Staged Combustion 175
in Bench Unit
VI1-3 Effects of Staged Combustion on NO Levels at 178
8 Atmospheres
VI1-4 Effects of Staged Combustion on NOX Emissions at 179
5 Atmospheres
VII-5 Effect of the Bench Unit's Operating Parameters 180
upon SOp Emissions at 5 Atmospheres
VII-6 CO Emissions for Staged and Unstaged Combustion at 182
5 Atmospheres
VU-7 Ammonia Injection Run Conditions 185
VII-8 Ammonia Injection Location 187
VI1-9 Results of Program to Reduce NOX by Ammonia Injection 191
VII-10 Results of Ammonia Injection Runs 5 and 7 192
VII-11 Comparison of the Results of This Work and an Earlier 193
Work (21) on Ammonia Injection to Reduce NOX
VI1-12 Response Time for Ammonia Injection Run 5 197
VI1-13 Simulated Flue Gas Recirculation Program - Summary 200
of Results
VII-14 Significance of Changes 1n Emissions - Simulated Flue 201
Gas Recirculation Program
-------
LIST OF TABLES (Continued)
Page
VI1-15 Comparison of Ammonia Injection and Simulated Flue 203
Gas Recirculation Results
VII-16 Variation in Ca/S Mole Ratio 205
VII-17 Combined NO -Control Techniques Program. Changes in 207
NOX Levels x
VII-18 S09 and CO Emissions for Combined NO -Control Techniques 209
RuR 2 x
VI1-19 Bench Regenerator Run Summaries 216
VI1-20 Coal Fueled Bench Regenerator Run Summaries 219
xi
-------
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. We wish to acknowledge the efforts of the operating
and mechanical crews, Tony AHobelli, Tom Barresi, Jeff Bond, Jim Bond, Jack
Fowlks, Ted Gaydcs, Phi Hi pa Givens, Bob Groth, Ed Hellwege, Frank Huber,
Pete Madorma, Tom Morrison, Sal Pampinto, Bob Robinson, Warren Spend, Ted
Sutowski, Luther Tucker, and George Walsh. We also wish to thank Bill Dravis,
a designer of the Research Technology Services Division and our math clerks
Beth Fell and Sue Gregory. A special acknowledgement goes to Nancy Malinowsky
who typed this report.
We also wish to acknowledge the cooperation we received from the other
EPA contractors who worked with us on the mini pi ant on a number of joint
programs. This includes Chris Chaney, Mike Shackelton, Steve Schliesser and
Clyde Stanley of Acurex Corporation, Rick Parker of Air Pollution Technology,
Bob Hall, Willy Piispanen and Paul Fennelly of GCA Company, Ken Murphy and
Clem Thoennes of General Electric Company, Joe McCain of Southern Research
Institute and many of their co-workers.
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 Bruce Henschel, the EPA project officer, Pic
Turner and Bob Hangebrauck.
xii
-------
SECTION I
SUMMARY
The pressurized fluidized bed combustion of coal (PFBC) and regeneration
of sulfated S02 sorbent were studied in the continuous miniplant unit and the
semi-continuous bench unit. In the miniplant combustion program, a series of
runs were made to verify that PFBC could attain the 90% S02 retention level
using dolomite or limestone sorbents. The dynamic response in 502 emissions
to a sudden change in the sulfur content of the coal, and in the dolomite to
coal feed ratio, was also measured. Emission of NOX» other gases and parti-
culates was measured. Combustion efficiency results were updated.
The measurement and control of particulate emissions was studied. Various
particulate measurement methods were employed and cross checked. Particulate
emission control using cyclones, and using cyclones in series with a high
temperature/pressure ceramic fiber filter and with a granular bed filter was
studied. Particulate removal from the flue gas at low temperature/pressure was
also studied, using a conventional electrostatic precipitator and bag house.
Regeneration of sulfated limestone was studied in a series of extended
tests. A series of sampling campaigns was also completed in which samples
were taken which were used to conduct a comprehensive analysis of emissions
with and without regeneration.
Combustion and regeneration studies were also carried out in the bench
unit. NOX control methods were evaluated in a series of combustion tests.
The use of coal as the regenerator fuel was also studied.
COMBUSTION STUDIES
The miniplant combustor consists of a refractory lined vessel 10 m (33
ft) high with an inside diameter of 32 cm (12.5 in). A number of vertical
water-cooled tubes are mounted in the combustor to remove the heat of com-
bustion.
Premixed coal and sorbent are injected into the combustor a single point
28 cm (11 in) above the fluidized bed support grid. The combustor normally
operates at pressures up to 950 kPa (9 atm), at temperatures up to the ash
agglomeration temperature of the coal (usually less than 960°C), at super-
ficial velocities of up to 2 m/s (7 ft/sec) and with expanded beds of up to
3.3 m (12 ft). The coal feed rate is normally less than 160 kg/hr (350
Ib/hr). Flue gas leaving the combustor passes through three cyclones in
series to remove most of the particulate matter. Particulates captured in the
first cyclone are recycled to the combustor to improve combustion efficiency.
Particulates captured in the second and third stage cyclones are rejected
through lock hoppers. Spent sorbent is also rejected from the combustor
through a lock hopper system to maintain a constant bed level in the combustor.
-------
Runs were made with: an Eastern bituminous Pittsburgh seam coal
(Champion) containing about 1.5% sulfur with a particle size distribution of
300 to 2400 microns; an Illinois No. 6 coal containing about 3.5% sulfur
screened to a particle size range of 400 to 3400 microns; and an Ohio coal
(Valley Camp) containing 2.5% sulfur screened to a size range of 300 to 3400
microns. A Virginia limestone (Grove No. 1359) and an Ohio dolomite (Pfizer
No. 1337) were used. Both were screened to a size range of 700 to 2400
microns.
Operational performance of the combustor was very good. At the comple-
tion of the program in August 1979, over 3700 hours of coal combustion time
was accumulated. Twelve runs of 100 to 250 hours duration were included in
the total. The ability of the mini plant to conduct short daily runs was also
demonstrated. In one period, fourteen 8 to 10 hour runs were completed in
fifteen working days. Multiple purpose runs were also conducted. In some
cases, as many as three EPA contractors, in addition to Exxon, were present at
the same time, conducting sampling and equipment evaluation tests during
extended combustion runs. During this reporting period, a total of 1117 hours
of test time was dedicated to a materials evaluation program sponsored by the
Department of Energy (DOE), under a cooperative agreement between DOE and EPA.
Gas turbine and boiler tube materials were evaluated under realistic PFBC
operating conditions. Evaluation of test results will be made by General
Electric and Westinghouse, the DOE contractors involved in the materials test
program.
S02 retention studies for EPA were conducted to confirm the sorbent
requirements, indicated in earlier tests, needed to assure 90% or higher S02
retention. The work was done in support of the most recent New Source Perfor-
mance Standards for large coal fired boilers. A number of runs were made with
Ohio and Illinois coals, dolomite and limestone sorbents at various calcium to
sulfur (Ca/S) molar ratios. It was found that 90% and higher S02 retention
levels could be reached with either sorbent. In fact, retention levels as high
as 99*% were measured. As expected, dolomite was shown to be more reactive.
A Ca/S ratio of 1.5 will assure 90% S02 retention at a gas superficial residence
time of 2 s while limestone use will require a Ca/S ratio of between 3.5 and 4.0,
The recent S02 retention results obtained with dolomite sorbent were
analyzed using the simple first order kinetic expression developed previously
(2). The recent data did not show the expected effect of gas residence time
on S02 retention. However, the lack of a gas residence time effect may have
been due to the fairly narrow range of residence time variations. It may also
indicate that the reaction rates may be affected in larger FBC units such as
the miniplant, by other operating parameters such as solids recycle and
rejection. These factors may make it difficult to obtain satisfactory results
using simple rate expressions, the rate expression must be refined to account
for these factors. However, this rate expression can be used as an approx-
imate guide until a better understanding of the FBC system 1s developed.
The dynamic response of the S02 emissions to sudden changes in the sulfur
content of the coal, and in the Ca/S ratio, was also studied. A new auxiliary
coal/sorbent feed system was built and installed to permit rapid switching
-------
from the current feed system to the new one to cause the rapid change in coals
and coal/sorfaent ratios. The time to switch from one system to the other was
about 40 s. The work was done in cooperation with General Electric, another
EPA contractor, who is developing a dynamic SOg response model for assessing
automatic control options. It was found that the S02 emissions were stabilized
in 7 to 8 minutes following a step change in the sulfur content of the coal
from 1.9 to 3.4%. However, when the Ca/S ratio was changed suddenly from 1.4
to 0.4, the time to stabilize was 110 min. The change back to the higher Ca/S
ratio required 300 min to reach a stable emission level. The reason for the
difference in the two response times (110 and 300 min) is not apparent.
However, it is clear that the response to a change in the Ca/S ratio has a time
constant much greater than that for a change in the coal sulfur content. These
results will be used by General Electric to develop the emission response model
model. The model will be described in a separate report by General Electric.
NOX emissions measured recently follow the same trend line with excess air
developed in earlier studies. At an excess level of 20%, NOX emissions
averaged 0.2 +0.1 Ib/MBTU. Even at excess levels in the range of 60 to 100%,
emissions are no greater than 0.4 Ib/MBTU. Therefore, NOX emissions are still
well within the New Source Performance Standard even with the recent reduction
in the NSPS to 0.6 Ib/MBTU.
emissions were measured using the controlled condensation method in
place of the older Method 8 method. Method 8 results averaged 12 +.12 ppm,
controlled condensation results averaged 6 +_ 9 ppm. In both cases~the range
of values was 0 to 30 ppm. Although the controlled condensation method gave
an average value one half that of the Method 8 average, the uncertainty ranges
were so wide that the difference may not be significant. No positive con-
clusions could be drawn from the results regarding the cause or the factors
affecting the degree of $03 formation. The reduced sulfur compounds, HgS, COS
and CS2, were found to be less than the detectability level of 1 ppm in the
flue gas. Methane averaged 7 + 5 ppm, ethane 4 + 4 ppm, C3 through C^ hydro-
carbons were generally less than 1 ppm, the detectability limit. Sodium and
potassium present in the flue gas were measured by quenching and filtering a
flue gas sample, then extracting the collected particulates and sample system
with hot water. Sodium concentration in the extract was equivalent to 2 to
3 wppm in the flue gas. Potassium was found to be between 0.3 and 0.5 wppm
in the flue gas. No vanadium was detected, Chloride levels of about 50 wppm
in the flue gas were also measured.
Particulate emissions in the flue gas, after passage through three con-
ventional cyclones in series, generally ranged from 0.03 to 0.15 g/Nm^, com-
pared to the NSPS of about 0.035 g/Nm3 (13 ng/J) for particulates. The par-
ticulates generally had a median particle size of 1 to 2 ym, with 80 to 90%
smaller than 5 ym.
Carbon combustion efficiency measurements averaged 99.3% for the runs
made during this period, about 0.8% higher than expected based on the correla-
tion published in the previous report (1).
-------
PARTICULATE MEASUREMENT AND CONTROL
An important technical issue to be resolved before pressurized fluidized
bed combustion can be applied commercially is the degree of particulate
removal needed to protect the gas turbine. A related issue is which tech-
nology to use to achieve the needed degree of the particulate control. In
addition to meeting the particulate removal requirements set by the gas tur-
bine, the environmental New Source Performance Standards published by the
Environmental Protection Agency must also be met. Although the EPA is not
directly concerned with studying particulate control to protect gas turbines,
it 1s responsible for the evaluation of such particulate control devices to
the extent that these devices could determine the level, size and composition
of particulate matter emitted from a PFBC system. Therefore a particulate
control program was begun in which pre-turbine devices such as granular bed
filters, a ceramic fiber filter and high efficiency cyclones were evaluated.
In addition, post-turbine devices were also tested, based on the realization
that the degree of pre-turbine cleanup may not be sufficient to meet the more
stringent New Source Performance Standards for particulates. In this case, a
trailer mounted electrostatic precipitator (ESP) and a trailer mounted bag
house were connected to the mini plant flue gas system and tested under typical
low pressure, low temperature conditions.
In addition to evaluating the particulate control devices, It was also
necessary to develop and improve particulate measurement systems. Such sys-
tems were used to determine particulate concentrations in the flue gas before
and after the control devices and to measure particulate size distributions.
For most of the current program, particulate concentrations were measured by
passing isokinetlc samples of the flue gas through high temperature total fil-
ters. Particle size distribution was measured using a Coulter Counter on a
sample of particulate captured on the filter. A question arose as to whether
the size distribution measured by the Coulter Counter was the same as that
occurring in the flue gas. The degree of particulate agglomeration on the
filter and redlspersion during the Coulter Counter measurement were unknown,
and could affect the measured particle size. The primary concern was that
redlspersion in the Coulter Counter resulted in a particle size distribution
significantly finer than that actually existing in the flue gas. To answer
the question, a series of samples were obtained by Southern Research Institute
and Air Pollution Technology, Inc. using a high temperature, pressure cascade
impactor developed by Air Pollution Technology. This is a different parti-
culate sampling system which would not cause possible agglomeration and redis-
persion effects. The impactor results indicated a larger concentration of fine
particles and a mass median size of about 0.8 m in the flue gas following the
third cyclone, compared to 1.6 m measured by the Coulter Counter. Electron
micrographs of the material captured on the impactor stages also showed very
little sign of agglomeration. These findings indicated that the results of
the filter/Coulter Counter method differed somewhat from cascade impactor
results, but to the degree expected, based on measurements made in other par-
ticulate systems. The Coulter Counter is definitely not biased toward the
finer particles by breakdown of agglomerates in the Coulter Counter during
sample preparation. A cross check was also made between the Coulter Counter
and a Bahco inertia! system. Both devices gave very similar particle size
distributions. An attempt was also made to use a miniature five stage cyclone
-------
train to obtain participate concentration and size distributions. Satisfactory
particle size distributions could not be obtained with this system due to
overlapping size differentiation on the cyclone stages. A continuous particle
monitor was also tested but did not perform satisfactorily. This device
measured the rate of transfer of electrical charge from particles in the duct
to a stationary rod as the charged particles impinged on the rod. The system
was very sensitive to temperature changes and was not reliable. It was con-
cluded that either the filter/Coulter Counter system or a cascade impactor
system was satisfactory. The filter/Coulter Counter system is easier to use
and probably gives a better measure of particulate concentration, but gives a
coarser size distribution.
Particulate control studies included the testing of the three-cyclone
system installed on the mi nip!ant in addition to a number of other systems.
The second cyclone was found to be 90 to 95% efficient, reducing the concen-
tration of particulates entering the cyclone with a mass median diameter of
20 to 25 um to 0.4 to 1.2 g/Nm3 in the cyclone outlet. The outlet particulate
stream had a median size of 3 to 5 ym, The third cyclone was found to be
about 90% efficient, reducing the concentration of particulates entering it
with a mass median diameter of 3 to 5 ym to 0.03 to 0.15 g/Nm3 in the outlet.
Particulate in the flue gas leaving the third cyclone had a mass median par-
ticle size of 1 to 3 ym. The efficiency of third stage cyclone was indepen-
dently verified by a team from Southern Research Institute and Air Pollution
Technology, Inc., using the same high temperature/pressure cascade Impactor
system mentioned previously. The impactor results verified those measured
using total filters and Coulter Counter particle size measurements. Both
methods indicated the cyclone cut diameter (particle size captured with 50%
collection efficiency) to be about 0.7 to 0.9 ym. The measured efficiency
is significantly greater than that expected from theoretical efficiency models.
The reason for the difference is not known.
Although the cyclone system, at times, did just meet the EPA New Source
Performance Standard for particulates (0.035 g/Nm3), it did not do so con-
sistently. Further studies were carried out with cyclones of differing designs
to see if the performance could be optimized and higher efficiencies obtained.
Two new cyclones were built similar to those currently being tested at the
National Coal Board PFBC facility in Leatherhead, U.K. They were tested on the
miniplant flue gas in place of the existing conventional third stage cyclone.
Contrary to expectations, the new cyclones were found to be no more efficient
than the existing third stage cyclone. Also, no significant effects of pres-
sure, inlet velocity, temperature or coal type on cyclone performance were
found. An increase in inlet particulate concentration did increase overall
efficiency significantly. Comparisons will be made with the performance of
the similar cyclones at the National Coal Board PFBC unit as the data become
available.
Tests were also conducted on the performance of a high temperature/pressure
ceramic fiber filter provided by Acurex Corporation. A single cylindrical
filter element consisting of a mat of Saffil alumina, was mounted in a heated
pressure vessel and used to filter a slipstream of flue gas extracted between
the second and third cyclones. The filter was cleaned periodically by a com-
bination of reverse flow of compressed air and short, high pressure pulses.
-------
Pressure drops were continuously recorded and were found to be 0.2 to 2 kPa
after cleaning, rising to a maximum of 14 kPa before cleaning. Filtration
efficiencies were high, ranging from 96 to 99.5%. Outlet particulate loadings
averaged 0.013 g/Nm3, well under the EPA emission standard for particulates
and significantly lower than that obtained with the three stage cyclone system.
An extended test was attempted but was terminated after 19 hours due to failure
of the filter. The failure was caused by a combination of corrosion of the
very fine filter support screen and excessively high reverse pressure drops
during bag cleaning. Corrosion of the support screen was probably accelerated
by a failure in the temperature control system during this test. The support
screen used in these tests was the only type available at the time and it was
recognized that it would have a short lifetime at test conditions. The tests
did indicate the ceramic filter was operable under PFBC conditions. High
collection efficiency at high face velocities was measured. In general, the
test filter exhibited performance similar to that of a conventional bag
filter. However, additional work is needed to develop a better mechanical
support system to prevent bag failure.
A granular bed filter was also evaluated as a high temperature, pressure
particulate control device. This is a system in which particulates are
removed by passage through a bed of granular material. A number of small beds
operated in parallel are used to reduce the pressure drop. The beds are
periodically cleaned by the reverse flow of clean gas. The "blow back" gas
fluidizes the granular filter media at a velocity sufficient to blow the
collected dust off, but low enough to prevent blowing the filter media itself
out of the filter vessel. The removed dust settles outside the filter and is
collected at the bottom of a containment vessel. Previous studies with this
system were beset with operating problems, primarily caused by plugging of the
filter inlet sections with particulates, poor bed cleaning and loss of filter
media during cleaning. Filtration efficiency was low and furthermore, decreased
with time. Modifications were made to the system in an attempt to improve
performance. This included testing a number of inlet screen sizes, using
baffles to prevent the loss of filter media and testing filter media of various
sizes and densities, also to prevent loss during the cleaning step. However,
the test program was again unsuccessful due to poor filtration efficiency,
loss of filter media and other serious operating problems which persisted to
the end of the program. The lowest outlet particulate concentrations measured
were 0.07 to 0.11 g/Nm3, but the concentrations increased with time during the
tests by as much as a factor of three. At no time did the filter performance
equal that required by the particulate emission standards. Since the filter
was installed after the second stage cyclone, the outlet loadings could be
compared to those from the third stage cyclone. The comparison indicated that
the cyclone was more efficient and moreover, the efficiency was maintained.
The poor filter performance was believed to be due to poor cleaning. Filter
beds examined after a test program were generally found to contain high concen-
trations of dust intimately mixed with the filter media. Tests made with
transparent models at ambient conditions showed that the dust was adhesive and
only partly removed during blow back. Some of the dust adhered to the filter
media and was mixed into the filter bed by the motion of the fluidized filter
media. Attempts to improve bed cleaning by modifying the blow back conditions
were partly successful. However, outlet concentrations were still high and
-------
increased with time. Attempts to prevent filter media Toss were also unsuc-
cessful . Screens were found to plug, even when screen opening sizes as large
as 10 mesh were used. Larger filter media and very dense filter media were
tested but did not prevent the loss. Baffles were also unsuccessful. Trans-
parent model tests were made and indicated that the loss was probably caused
by a surge in the blow back gas flow caused by the sudden opening of the blow
back valve. It was also suggested that the bed height was too low* not
allowing enough space for disengagement of the fluidized filter media. In
addition to these problems, the operation of the filter caused periodic upsets
in the miniplant pressure control system, which in turn caused coal feeding
problems. Pressure drops across the filter were also high, probably due to
poor bed cleaning.
A final attempt was made to determine if an operable granular bed system
could be developed. A small single bed filter was installed and tested on a
flue gas slipstream after the second stage cyclone. The blow back system was
modified to bring in some of the reverse flow gas at the interface between
the filter media and the dust layer. This blow back gas was directed
horizontally across the top of the bed and was intended to shear off the dust
layer without disturbing the filter media enough to cause its loss. A single
test was made with the modified filter but was unsuccessful. Pressure drops
were very high after the first few minutes of operation. Increasing the
severity of the cleanup step again resulted in loss of filter media.
Unfortunately, the testing of this filter was terminated due to lack of time
before all operability questions could be answered. An unresolved issue was
whether the loss of filter media could have been prevented by modifying the
blow back procedure, possibly by introducing the blow back gas more gradually.
The filter face velocity was also very high in the test and the effect of
lowering the velocity on filter performance was also unresolved.
A conventional low temperature, low pressure electrostatic precipitator
(ESP) and bag house were also tested on the flue gas from the miniplant in a
series of long term tests. The purpose of the tests was to determine if an
ESP or bag house could be used after expansion of the flue gas through the gas
turbine, to meet particulate emission standards. This assumed that cyclones
may be sufficient to protect the gas turbine from excessive wear but would not
be sufficient to meet environmental standards. The tests were performed using
mobile, trailer mounted systems operated by Acurex Corporation for the EPA.
Both systems appeared to be applicable in this service. The ESP overall
efficiency was 87%, corresponding to an emission level of about 0.02 g/Nm3.
The bag house overall filtration efficiency was 99.3%, corresponding to an
emission level of about 0.001 g/Nm3.
REGENERATION STUDIES
A series of extended runs was made in which sulfated limestone was con-
tinuously regenerated and recirculated to the combustor. The primary objective
of the runs was to determine the reduction in fresh limestone requirement
resulting from regeneration of the sul fated limestone.
-------
The regenerator consists of a refractory lined vessel with an inside
diameter of 22 cm (8.5 1n) 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 reduc-
tion reaction occurs. Supplementary air is injected directly into the bed
at a higher elevation to create an oxidizing zone. The oxidizing environment
at the top of the bed assures high selectivity to CaO, the desired product of
the regeneration reaction, by minimizing the formation of CaS, an undesired
byproduct.
Three extended runs totalling 370 hours were completed 1n the test series.
Operations were generally good. Fresh limestone, at a Ca/S ratio ranging
from 0.7 to 1.5 was fed to the combustor. SOg emissions from the combustor
were low, the SO? retention exceeded 90% at all times and was normally above
97%. In a once through system, a Ca/S ratio of 3 to 4 would have been
necessary to achieve the same level of S02 retention. The recirculating
regenerated sorbent rates were much higher than the fresh limestone rates,
giving total Ca/S ratios entering the combustor usually between 6 and 12 (one
was 54). The regenerated sorbent appeared to have the same activity of fresh
limestone with no sign of diminished activity due to regeneration. There was
also no evidence of loss in activity during the runs. Loss of sorbent by
attrition and entrainment from the combustor and regenerator was somewhat
greater than the corresponding loss from a once through operation. In fact,
the fresh limestone Ca/S ratios fed to the combustor were usually those
required to maintain constant sorbent levels in the combustor and regenerator.
It had been planned to reduce fresh limestone feed rates to the point where
the S02 retention in the combustor was about 90%. This could not be done,
since sorbent feed would have had to be reduced to a point where it would have
been inadequate to make up for attrition/entrainment losses, and bed depth
would have dropped. In a once through operation, the entrainment losses are
about 1% of the bed/hr. In the regeneration runs, the entrainment losses
averaged about 1.8%. S02 concentration in the regenerator off gas was low,
ranging from 0.2 to 0.5% equivalent to 6 to 16% of the calculated equilibrium
concentration. These levels are lower than what might reasonably be expected
in a commercial system. As pointed out in previous reports (1,2), due to the
size of the miniplant regenerator, the S02 level in the miniplant regenerator
off gas is determined by energy and mass balance requirements rather than
chemical equilibria. It had been planned to use a higher sulfur coal in these
tests to maximize the S02 concentrations. Unfortunately, a prepared high
sulfur coal could not be obtained and a lower sulfur coal was used, contri-
buting to the low S02 concentrations.
However, the runs did establish the minimum degree to which the sorbent
requirement would be decreased by regeneration. This is a factor of 3 to 4
The runs also indicated that regenerated sorbent activity was high and
activity loss due to regeneration was low.
8
-------
COMPREHENSIVE ANALYSIS OF EMISSIONS
The comprehensive analysis of emissions program initiated previously was
concluded. In the earlier tests, a Level I sampling and analytical program
was conducted in cooperation with Battelle Columbus Laboratory, the EPA con-
tractor responsible for the program at that time. In the current series of
tests, a set of samples was taken during a run with Illinois No. 6 coal and
analyzed for inorganic elements by spark source mass spectroscopy (SSMS).
The results were forwarded to Battelle to be included with the Level I results
obtained by them in the earlier tests which used Champion (Eastern) coal.
Also, in the current series of tests, sampling campaigns were conducted in
cooperation with GCA Corp., the current EPA contractor. In this case, Exxon
assisted 1n sampling but essentially all analytical work will be done, inter-
preted and reported by GCA. In the recent tests, a Level I sampling campaign
was conducted during an extended run in which the regenerator was in operation.
A Level II sampling campaign was also conducted with the combustor operating
in a once- through fashion.
A brief investigation into the formation of the double salt MgsCaCSO^H or
MgS04 *n tne PFBC solid waste products was conducted. These materials, if
present, could cause environmental problems upon disposal since they are water
soluble. No MgS04 was found in any of the samples. MgsCa(S04)4 was found in
solid samples collected in the third stage cyclone and the gas turbine test
section downstream of the cyclone. However, S02 concentrations in the run from
which samples were taken was considerably higher than normal. Also, the resi-
dence time in the mi nip! ant ducting is probably greater than that which will
occur in a commercial plant and the average temperature somewhat lower. All of
these factors promoted the formation of the double salt. Additional tests
should be conducted in newer PFBC units to determine the extent of the double
salt formation under more realistic conditions.
BENCH UNIT STUDIES
Programs were carried out in both the bench combustor and regenerator sec-
tions. The combustor had been modified to permit continuous solids feeding and
removal. The combustor was then used to evaluate three NOx control methods:
two stage combustion, NHs injection and simulated flue gas recirculation. The
regenerator was used in a series of tests using sul fated sorbent produced in
the mini pi ant, studying the use of natural gas and coal to fuel the regenera-
tion section.
The bench combustion unit consists of a refractory lined combustor vessel
which normally operates at temperatures of 840 to 950°C, pressures of five to
eight atm, superficial velocities of 1 to 2 m/s and coal feed rates of 1 to 12
kg/hr. The inside diameter is 11.4 cm, the interior height is about 4.9 m.
Three sets of vertically mounted water cooled coils are located inside the com-
bustor to remove heat of combustion. Coal is fed by a pneumatic injection
system similar to that used on the miniplant. Sorbent is fed separately by
means of a transfer line lock hopper consisting of two cycling valves mounted
in the sorbent feed line. Provisions were also added to inject supplementary
air into the combustor at various locations to study two stage combustion, and
to inject NHa and N2 to study the effects of NHa injection and simulated flue
gas recirculation on NOX emissions.
-------
Two sets of runs were made studying the effect of two stage combustion.
The first was made at 8 atm pressure. The runs were subject to frequent
temperature excursions and bed agglomeration caused by low superficial veloc-
ities 1n the bed below the point where the supplementary air was Injected.
As a result, the test series was not completed. The pressure was reduced to
5 atm to Increase the velocity and the second set of tests was run. The
location of the supplementary air Injector was varied during the Initial tests
at 8 atm. Injection of the air Into the bed, 15 cm above the grid, rather
than above the bed, was necessary to promote CO burnout. Despite the problems,
some data were obtained during the 8 atm tests which Indicated that NOX emis-
sions could be reduced about 30% by using two stage combustion. The test
series at 5 atm pressure was completed with no serious operating problems.
The supplementary air Injection point was raised to 33 cm above the grid to
Increase the residence time 1n the lower reducing zone thereby promoting NO
reduction reactions. However, the runs at 5 atm resulted 1n only a 10 to 20%
reduction 1n NOX emissions. The Increase from 10 to 20% reduction occurred as
the amount of primary combustion air was decreased from 90 to 75% of the
stolchlometrlc amount at a fixed level of total combustion air (primary plus
supplementary). The reason for the apparent effect of pressure on the effect-
iveness of two stage combustion Is not known. NOX emissions were also shown
to Increase with overall excess air and temperature as had been noted in
previous studies (1,2).
Staged combustion increased S02 emission levels about 20%. The effect
was believed to be caused by low oxygen concentrations in the reducing zone,
depressing the calcium sulfation reaction which requires the presence of
oxygen. CO emissions were increased up to 20% by staging, but only at low
excess air levels. The increased S02 and CO emissions could be offset by
increasing the gas residence time in the oxidizing zone slightly. A more
serious concern is the effect of two stage combustion on boiler tube materials.
This potential problem must be addressed 1f two stage combustion is to be
studied further.
NH3 injection was also evaluated as a means of reducing NOX emissions
from PFBC. The use of NHo to reduce NOX emissions was developed by Exxon
Research and Engineering Company and successfully applied to conventional
oil, gas and coal fired furnaces. The current program was designed to test
the effectiveness of NHo injection under PFBC conditions. A series of tests
was conducted by Injecting NH3 at NH3/NO ratios ranging from 1 to 8 at three
locations, below the grid Into the combustion air, into the combustor at the
bed overflow port (168 cm above the grid) and into the combustor freeboard
region, 290 cm above the grid. Combustor pressure was approximately 7 atm
during the tests. The effect on NOX emissions was found to be a strong
function of location and the NHo/NOx ratio. Injection below the grid into the
combustion air resulted in a 50% Increase in NOX emissions. Injection at the
bed overflow port had no effect on emissions. Injection in the freeboard zone
decreased NOX emissions 30 to 50%. Increasing the NH3/NOX ratio Increased NOX
emissions at the below-grid location, but decreased emissions at the freeboard
zone location. The results are probably related to local oxygen concentra-
tions and temperatures. Injecting below the grid exposed NH3 to high 02 con-
centrations and high temperatures near the coal feed point. These conditions
10
-------
promoted NH3 oxidation rather than the NO/NH3 reaction. In the freeboard
zone, the lower temperatures (~700°C) and Og concentrations favored NO reduc-
tion. At the intermediate location, the temperature, and possibly 02 concen-
tration conditions, promoted each of the competing reactions equally, resulting
in no significant effect on NOX emissions. The optimum temperature for the
NH3/NO reaction found in this study (~700°C) is lower than expected based on
previous studies and further work may be needed to determine if it is the true
optimum. If such a temperature is required, NHs could be injected ahead of the
gas turbine. The desired temperature would occur within the turbine itself,
where the NOX removal reactions would take place.
The effect of flue gas recirculation on NOX emissions was simulated by
adding nitrogen to the combustion air inlet stream. The ratio of N? to air
was set at 0.10 vs 0.2. Excess air and temperature were also varied. The use
of N2 addition was found to cause an apparent decrease in NOX emissions.
However, the addition of N2 always caused a decrease in the excess air level
since more coal had to be fed to heat up the cold N? added to the inlet air.
When the results were plotted against excess air and compared to results
obtained without N2 addition, no significant effect of N£ addition was seen.
Therefore, it is concluded that flue gas recirculation would have no effect
on controlling NOX emissions. S0£ and CO emissions were also increased
significantly, probably because of lowered oxygen partial pressure and gas
residence time.
A series of tests was conducted to determine if a combination of two
stage combustion and NH3 injection would result in further reductions in NOX
emissions. The runs were made by adding the control techniques one at a time
and then in combination to a run made under conditions typical of those used
in the NOX control program. It was found that either two stage combustion or
ammonia injection decreased NOX emissions 25 to 30%. However, the combination
of the two techniques resulted in a reduction of only 26%. Therefore, the
combination did not offer any advantages over the use of either method alone.
It should be mentioned that only one run was made and that additional testing
might be needed to determine if a positive interaction exists.
Regeneration studies were carried out in the modified bench regenerator.
The unit now is equipped for continuous feeding of sulfated sorbent and with-
drawal of regenerated sorbent. The regenerator vessel is refractory lined to
an inside diameter of 9.5 cm and has an interior height of 4.6 m. The air
and fuel addition systems are similar to those used in the miniplant regen-
erator. A below-grid grid burner provides most of the heat required by the
system. Additional fuel is added above the grid to create a reducing zone
where CaS04 reduction occurs. Supplementary air is added higher in the bed
to create the oxidizing zone needed to oxidize the undesirable CaS byproduct
and complete combustion of the fuel. The test program was divided into two
segments, natural gas fueled regeneration and coal fueled regeneration. Most
of the work concentrated on developing satisfactory operating techniques. The
work done with natural gas-fueled regeneration included a study of the most
effective way to introduce fuel to the regenerator and a limited study of
operating variables and their effect on sorbent regeneration. Studies made
11
-------
with coal-fueled regeneration were preliminary 1n nature and were Intended to
determine the feasibility of operating a regeneration unit with coal. The
Incentive to use coal Instead of natural gas of course, 1s to reduce operating
costs and minimize dependence on natural gas supplies.
Attempts were made to add all the natural gas fuel directly to the bed,
using the below-grld burner only to preheat the bed to temperatures high
enough to Initiate methane combustion. This was done after 1t was found that
about 30% of the natural gas fed to the burner was used to overcome heat
losses from the burner plenum. The attempts to add all the fuel directly to
the bed were not completely successful due to the occurrence of frequent high
temperature excursions. The temperature excursions were believed to be caused
by unstable combustion conditions in the regenerator. Operations with natural
gas addition to the burner as well as to the bed were more successful. A
series of six runs was made. However, because of the limited amount of data
obtained it was not possible to draw conclusions concerning the relationship
between operating variables and the extent of regeneration. The S02 concen-
tration 1n the off gas varied from 0.2 to 0.6%, equivalent to 15 to 30% of
the calculated equilibrium concentrations. These low SO? concentrations
occurred since the concentrations were not limited by kinetics or thermo-
dynamics, but rather by heat and material balance requirements.
A series of runs was made with coal added to the bed instead of natural
gas. The results of the few runs were promising. The only serious operating
problem was maintaining good temperature control. Some of the runs were
subject to high and erratic bed temperature. Modifications, such as the use
of nitrogen instead of air as the coal transport medium, improved temperature
control significantly. However, control problems persisted until the end of
the program, possibly due to poor solids and gas mixing caused by the small
regenerator diameter. Solids regeneration levels (the percentage of the sul-
fated sorbent converted to CaO in the regenerator) were around 40%, somewhat
lower than the levels of 40 to 80% measured with natural gas fueled runs.
S02 concentrations in the off gas were about 17% of the calculated equilibrium
concentrations, roughly comparable to the levels measured with natural gas
fuel. More work is needed before coal-fueled regeneration can be completely
evaluated.
12
-------
SECTION II
INTRODUCTION
The pressurized fluidlzed bed combustion of coal is a new combustion
technique which can reduce the emission of $03 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
fluidlzed bed material. In addition to emissions control, this technique
has other potential advantages over conventional coal combustion systems
which could result in a more efficient and less costly method of electric
power generation. By immersing steam generating surfaces in the fluidized
bed, the bed temperature can be maintained at low and uniform temperatures in
the vicinity of 800 to 950°C. The lower temperatures allow the use of lower
grade coals (since these temperatures are lower than ash slagging tempera-
tures), and also decrease NOX emissions. Operation at elevated pressures, in
the range of 600 to 1000 kPa, offers further advantages. The hot flue gas
from a pressurized system can be expanded through a gas turbine, thereby
increasing the power generating efficiency. Operation at the higher pressure
also results in a further decrease in NOX emissions.
In the fluidized bed boiler, limestone or dolomite is calcined and
reacts with SOg and oxygen in the flue gas to form CaS04 as shown in reaction
v' / •
CaO + S02 + 1/202 + CaS04 (1)
Fresh limestone or dolomite sorbent feed rates to the boiler can be
reduced by regeneration of the sulfated sorbent to CaO and recycle of the
regenerated sorbent back to the combustor. One regeneration system, studied
by Exxon Research and Engineering Company in the past, is the so-called one
step regeneration process in which sulfated sorbent is reduced to CaO in a
separate vessel at a temperature of about 1100°C according to equation (2).
The goal is to produce SOg in the regenerator off gas at a sufficiently high
concentration to be recovered in a by-product sulfur plant.
CO C02
CaS04 + H2 + CaO + S02 + H20 (2)
A diagram of the pressurized fluidized bed combustion and regeneration
process is shown in Figure II-l.
Exxon Research and Engineering Company, under contract to the EPA, has
built two pressurized fluidized bed combustion units to study the combustion
and regeneration processes. The smaller of the two units, the bench scale
unit, was built under contract CPA 70-19 and was described in previous
reports (1,2,3,4). Those reports also described regeneration and combustion
studies carried on in the bench unit. This report describes work done in the
13
-------
FIGURE ll-l
PRESSURIZED FLUIDIZED BED COAL COMBUSTION SYSTEM
AIR COMPRESSOR
GAS TURBINE
CONDENSER
COAL AND
SORBENT MAKEUP
HIGH EFFICIENCY
SEPARATOR
SOLIDS
TRANSFER
SYSTEM
r\TO SULFUR
RECOVERY
SEPARATOR
I
DISCARD
COMBUSTOR
Tl
FUEL
REGENERATOR
-------
bench combustor evaluating various NOx emission control methods. Regenera-
tion studies were conducted in the bench regenerator evaluating fuel injection
methods and the use of coal as the regenerator fuel. These results are also
included in this report.
The larger unit, called the mini pi ant, was designed under EPA Contract
CPA 70-19 and built under Contract 68-02-0617. Figure II-2 shows a photo-
graph of the miniplant. The shakedown and operation of the unit was funded
under Contract 68-02-1312. Previous reports (1,2,3,5) described design,
shakedown and operation of the unit. This report includes additional results
from the operation of the combustion and regeneration sections of the mini-
plant. The combustion program included tests to verify and document further,
the ability of PFBC to reduce S02 emissions by 90% or more. The dynamic
response of S02 emissions to a rapid change in coal sulfur content and dolo-
mite to coal feed ratio was also measured. This was done to provide the basis
for the evaluation of an S02 emission control concept by General Electric
Company under a separate EPA contract. Additional data on the emission of
NOX, other gases and particulates were also obtained. Particulate control
studies evaluating high efficiency cyclones, a high temperature/pressure
ceramic fiber filter, a granular bed filter and a low temperature/pressure
electrostatic precipitator and bag house were also completed. This work was
done in cooperation with a number of other EPA contractors including Acurex
Corporation, Southern Research Institute and Air Pollution Technology, Inc.
A series of regeneration tests was also conducted. Three extended runs
were completed in which activity maintenance of the regenerated sorbent, SOg
concentrations in the regenerator off gas and S02 retention in the combustor
were measured.
A series of sampling campaigns was conducted in cooperation with another
EPA contractor, GCA/Technology Division, to provide samples for Level I and
Level II comprehensive analysis test programs. Samples were taken during
operation of the combustor alone and during regeneration tests in which the
combustor and regenerator were both in operation.
The period of performance discussed in this report is August 12, 1977
to August 7, 1979.
15
-------
FIGURE 11-2
EXXON FLUIDIZED BED COMBUSTION MINIPLANT
16
-------
SECTION III
MINIPLANT COMBUSTION STUDIES
Combustion studies have been carried out in the EPA/Exxon pressurized
fluidized bed combustor referred to as the miniplant. The miniplant is shown
schematically in Figure III-l. This unit has provisions for continuous
addition of coal and sorbent and continuous withdrawal of sulfated sorbent.
It normally operates at pressures of about 9 atm, temperatures of 840 to
950°C, with superficial velocities of about 1 to 2.5 m/s, feeding 100 to
160 kg/hr of coal and 15 to 20 Nm3/min of combustion air. As of August 1979,
the combustor has been operated for a total of approximately 3700 hours in a
series of individual runs up to 250 hours duration. This section of the
report describes the combustor equipment, operating procedures, combustor
performance and combustion results. A discussion of the regeneration work is
given in Section V.
EQUIPMENT, MATERIALS, PROCEDURES
This section will focus on the major system components which include:
(1) solids feeding system, (2) combustor with internal subcomponents, (3)
flue gas system, (4) temperature and pressure control, (5) flue gas sampling
and analytical system, (6) process monitoring and data generation system, (7)
combustor safety and alarm system, (8) coal and sorbent properties, and (9)
operating procedures. A detailed description of each of these systems can be
found in earlier reports (1.2) and only a brief discussion will be included
here. Changes in the equipment made since the last report are also included.
Soljds Feeding System
Figure III-2 displays a schematic of the miniplant coal and sorbent
feeding system. Crushed and sized coal and limestone or dolomite are held
in separate storage bins (20 tonnes for coal and 3 tonnes for sorbent) under
atmospheric conditions. On demand, the solids from the bins are proportioned
to a specific coal/sorbent ratio. Inverters control the motor speeds of
separate coal and sorbent screw feeders and volumetrically control the coal/
sorbent ratio. A blending screw transports the mixture into a solids feed
vessel. The coal/sorbent mixture is held in this vessel until refill of the
injector vessel is required.
The solids feeding system provides for continuous solids delivery (coal
and sorbent) from the injector vessel to the pressurized combustor, while
allowing intermittent refilling of the injector vessel (193 kg operating
capacity). Load cells located under the injector vessel monitor the solids
feed rate and actuate control signals for the refill cycle. Prior to initia-
tion of a refilling operation, the injector vessel, feed vessel, and the pair
of solids storage bins remain isolated from each other. When the load cell
under the injector vessel detects a solids loading of less than 102 kg, 91
kg of solids are automatically transferred pneumatically from the feed vessel
17
-------
FIGURE 111-1
EXXON FLUIDIZED BED COMBUSTION MINIPLANT
COOLING
WATER
TO SCRUBBER
CO
ORIFICE
r
THIRD STAGE
SEPARATOR
AIR (PRESSURE CONTROL)
CYCLONE
SEPARATOR
TOOLING
COOLING WATER
OUT IN
TO
SCRUBBER
CYCLONE
SEPARATOR
SOLIDS
DISCHARGE
SOLIDS
REJECT
VESSELS
FEED
WATER
RESERVOIR
COAL AND
LIMESTONE
FEED SUPPLY
AUXILIARY
AIR
COMPRES
r
LIQUID FUEL STORAGE
NATURAL
GAS
COMPRESSOR
MAIN AIR
COMPRESSOR
-------
LIMESTONE
BIN
COAL
BIN
FIGURE 111-2
COAL & LIMESTONE FEED SYSTEM
CONTROLLERS
AP SIGNAL -eee*>
[FEEDER I | FEEDER |
BLENDER
J
VENT
04 _n|
3j~
FEED
VESSEL
h*18"-»-
S- &3
HIGH PRESSURE AIR
D-D
TC
I I I I
AP TEMP
t
i >^-
INJECTOR
VESSEL
MAI
I II II U U I
COMBUSTOR
A
1/2" S.S. PIPE
(LOOPED)
-------
to the pressurized injector vessel without interrupting feed to the combustor.
Refilling is usually completed in about 5 minutes. After refilling, the feed
vessel is again isolated from the injector vessel, vented, and filled with
solids from the storage bins. The feed vessel is then isolated and repres-
surized to await another cycle.
Solids in the injector vessel are continuously aerated by the pressurized
air stream, and the vessel is automatically controlled at a pressure level
slightly above that in the combustor. Originally the port used to sense com-
bustor pressure was near the top of the combustor where plugging with bed
material would not occur. Maintenance of the proper pressure differential
between the combustor and the injector vessel during minor bed upsets was dif-
ficult. These upsets sometimes resulted in hot solids back flow into the
injector vessel. For this reason, the combustor pressure tap was relocated
to a port very near the coal injection nozzle. A large purge flow rate in the
pressure sensing line and a high coefficient of damping 1n the measurement
system were used to control plugging and oscillations. This system has worked
without incident for over one year.
The coal and sorbent mixture 1s discharged from the injector vessel
through a 1.3 cm diameter orifice and pneumatically conveyed by a stream of
dried transport air through a 1 .6 cm ID stainless steel pipe. Originally
the pipe was S-shaped. To increase the pressure differential between the
combustor and the injector a longer conduit with one 180° and four 90° bends
was installed. The conduit is 17 m long with 30 cm radius bends. The pres-
sure differential between the injector vessel and the combustor was increased
from about 20 kPa to 70 kPa. The erosion of the coal transport conduit at
the location of the four bends consistently occurred after about 200 hours of
operation when the conduit was made of 0.37 cm thick, type 304 or 316 stain-
less steel heavy wall pipe. Heat treated, type 410 stainless steel pipe
bends although much harder than 304 or 316 stainless steel, did not produce
any better results. Carbonized steel pipe failed after only 60 hours of
operation. Limited data were obtained with type 304 stainless steel pipe
that had been alonized. After approximately 150 hours of operation, the
alonized conduits were still functioning.
A comparison of the locations where the holes were eroded in the conduit
bends of various materials showed that in all cases the failure occurred when
the angle of solids impingement upon the wall was between 15 and 20 degrees.
A segment from a type 304 stainless steel bend showing the hole location
after 200 hours of operation is shown in Figure III-3.
An auxiliary solids feed system was constructed to allow changing from
one coal/sorbent feed to another instantaneously. The capacity of this new
vessel is equivalent to about 8 hours of operating inventory. It 1s situated
on load cells to monitor feed rate. There is no provision for on-line refil-
ling. The coal/sorbent blend exits the auxiliary vessel through a 1.3 cm
orifice and is pneumatically conveyed by a stream of dried transport air
through a 1.5 cm ID polyurethane tube. This material was chosen to minimize
line erosion, because of the excellent resilience of polyurethane. This tube
is shielded in the event of rupture. The tube is 12.5 m long and winds once
around the 2.4 m diameter vessel before it reaches the combustor.
20
-------
FIGURE 111-3
CROSS SECTION OF COAL/SORBENT FEED LINE
I
-------
Final entry of solids Into the combustor 1s through a 1.3 cm ID nozzle
located 28 cm above the fluldlzlng grid and horizontally extending about
2.5 cm beyond the reactor wall. Both the primary and the auxiliary feed
systems have their own Identical entry ports located 180" apart on the com-
bustor. The tip of the probes Include ten 0.79 mm diameter holes which sur-
round the solids feed opening. They are used to continuously Inject an
annular stream of sonic-velocity air to assist penetration of the solids
feed Into the fluldlzed bed and to protect the feed nozzles from blockage
with bed solids. Only one probe Is used at any one time. The probe not 1n
use 1s continuously purged to prevent blockage of the openings with sol Ids.
The flow of solids Into the combustor 1s controlled to maintain constant
temperature 1n the combustor.
Combustor
The combustor consists of a 61 cm ID steel shell, refractory lined to an
Inside diameter of 33 cm. The 9.75 m high unit 1s fabricated 1n flanged sec-
tions to allow Insertion and removal of the cooling colls. Various ports are
strategically located to allow for material entry and discharge. Numerous
taps are also provided for monitoring both pressure and temperature. A
schematic of the combustor 1s shown 1n Figure III-4.
Heat removal from the combustor 1s provided by cooling colls located in
discrete vertical zones above the grid. Each coll has a total surface area
of 0.55 m* and consists of vertically-oriented loops constructed of 1/2-Inch
Schedule 40 316 stainless steel pipe. The number of colls normally varies
from one to four depending on the combustor operating conditions and the
amount of cooling required. A high pressure pump is used to pump the cooling
water through a closed-loop arrangement consisting of a dem1neral1zed feed
water reservoir, cooling coils, and a heat exchanger. The flow rate and exit
temperature from each coil can be separately controlled and monitored. Baf-
fles were installed on the cooling colls to prevent excessive erosion by the
bed solids (2).
Only one failure of the baffled cooling colls has been recorded after
1200 hours of operation. A hole (Figure II1-5) developed on the lower por-
tion of the coil closest to the bed support grid, facing the center of the
combustor, approximately 45° of arc along a bend. The lowest portion of the
coll was only 23 cm above the fluldlzlng grid. The section of coil was
analyzed and the failure was attributed to erosion, Very little oxidation
or sulfidatlon was present. The lack of sulfldation 1s not surprising, since
the coil water outlet temperature 1s controlled at 140°C. At this Internal
temperature, little sulfidatlon 1s expected. Most of the 180° bends 1n the
lower section of the coll also exhibited a decided sharpening of the leading
edge. This is also attributed solely to erosion. The entire coil 1s shown
in Figure III-6. A similar coll located further from the fluidlzing grid was
exposed over the same time period with no significant erosion. The erosion
problem was apparently solved by welding small rectangular steel "guards"
horizontally at the bottom of the colls. These should prevent erosion but
have been used for only 500 hours of operation to date.
22
-------
FIGURE 111-4
COMBUSTOR VESSEL
23
-------
FIGURE 111-5
MINIPLANT COMBUSTOR COOLING COIL 1A
AFTER 1200 HOURS OPERATION (CLOSE UP)
24
-------
FIGURE 111-6
MINIPLANT COMBUSTOR COOLING COIL 1A AFTER
1200 HOURS OPERATION
25
-------
The combustion air to the unit 1s provided by a main air compressor
having a capacity of 40 Nm3/min at 1030 kPa (1400 SCFM at 150 pslg). Preheat
of the combustor during startup 1s made possible by a natural gas burner which
1s housed 1n the bottom plenum section of the combustor. Once the fluldlzed
bed temperature reaches approximately 430°C, a liquid fuel system 1s used to
heat the bed further to the coal Ignition temperature. Both systems are more
completely described In a previous report (1).
With continuous addition of sol Ids to the combustor, there must be pro-
vision for continuous removal to control combustor expanded bed height. The
expanded bed height can be controlled at any level above 2.3 m by the con-
tinuous withdrawal of bed solids through a port located 2.3 m above the
fluldlzlng grid. Originally, solids flowed by gravity through a refractory
lined pipe Into a "pulse pot" from where they were pneumatically transported
by controlled nitrogen pulses to a pressurized lock hopper. This system did
not give precise solids flow control and was replaced with a dual valve sys-
tem. Solids now flow by gravity to a refractory lined sliding gate valve.
This valve and a ball valve behind 1t alternately open and close an adjustable
number of times per hour. Each cycle rejects 4 kg of bed material, the amount
of material that fills the volume between the two valves. From the ball valve.
the sol Ids empty Into a lock hopper which 1s manually emptied at regular
Intervals.
The combustor fluldlzlng grid consists of 332 3/32 Inch holes and four
Independent cooling water loops consisting of five channels each. It has
performed since run 50, without Incident. A more complete description Is con-
tained 1n a previous report (1).
Flue Gas System
Combustion gases exit the combustor and Immediately pass through two
cyclones 1n series. From the second cyclone, the gas passes through a long
duct to a large pressure vessel. This vessel was designed to contain a Ducon
granular bed filter or other high efficiency hot gas cleanup device. After
completion of the granular bed filter tests (run 66), the filter was replaced
with a third cyclone Installed within the vessel. When the third cyclone Is
used, no flow enters the vessel; 1t 1s merely a pressure containment device.
After passing through a third cyclone (or the granular bed filter 1n the
earlier tests), the flue gas 1s sampled, expanded and sent to a wet scrubber
for final cleanup. The dimensions of all three cyclones are shown 1n Figure
III-7. The dimensions of the second stage cyclone have been modified since
It was originally built. Originally, the barrel diameter was 31.8 cm and the
inlet section sized to give an Inlet velocity of about 10 m/s. After run 18.3,
the inlet dimensions were changed to Increase the inlet velocity to about
30-35 m/s. The outlet diameter was decreased. The final change was made
after run 47 to the dimensions in Figure III-7. These changes decreased the
barrel diameter but maintained the inlet section dimensions to keep the Inlet
velocity range fixed.
The granular bed filter, used 1n tests prior to run 67, 1s described in
Section IV.
26
-------
FIGURE 111-7
CYCLONE DESIGN DIMENSIONS
w h—
INLET
SECTION
T
^o
I
\ i
i,
DIMENSIONS (CM)
SYMBOL
Db
W
h:
DESCRIPTION
Inlet Type
Barrel Diameter
Inlet Width
Inlet Height
Outlet Pipe Insertion
Barrel Height
Cone Height
Outlet Pipe Diameter
Outlet/Inlet Ratio
Diameter of Cone at Bottom
Function
CYCLONE
NO. 1
Tangential
32.4
7.62
25.4
20.3
43.2
45.7
14.6
0.87
12.7
Recycle
CYCLONE
NO. 2
Tangential
17.8
4.45
10.2
13.3
35.6
35.6
8.89
1.37
9.0
Cleanup
CYCLONE
NO. 3
Tangential
15.2
3.81
7.62
11.4
40.6
20.3
6.2
1.04
9.0
Cleanup
27
-------
The primary Intent of the first cyclone 1s to redrculate larger unburned
carbon particles back to the combustor to Improve combustion efficiency. The
particles fall down the dlpleg and are relnjected with a nitrogen pulse about
66 cm above the fluldlzlng grid.
Particulates collected 1n the second stage cyclone are removed through
a lock hopper. The cyclone, which was originally cast with Grefco LUecast
75-28 refractory, was recast after run 47 with Resco RS-17-E, and Is still 1n
service after 2-1/2 years. None of the severe pitting that was present with
the earlier refractory was found.
From the second cyclone, the flue gas enters the large pressure vessel
which contains the third cyclone. This cyclone 1s of conventional design and
constructed of 316 stainless steel. Particulates captured by the third
cyclone are removed through a lock hopper.
Following the third cyclone, the flue gas may either be expanded through
a pressure control nozzle or a turbine test section supplied by General
Electric under contract with the Department of Energy (DOE Fireside Corrosion
Contract No. EX-76-C-01-2452). The test section contains 24 stationary
blades of various materials that simulate both the Impulse and reaction vanes
of a gas turbine. The turbine section was tested for 1100 hours of operation
with three stage cyclone partlculate cleanup as part of the DOE Fireside
Corrosion program. A more complete description 1s given 1n a report to DOE
(6).
In order to simulate gas turbine operation, care must be taken to prevent
the gas temperature from dropping below 843°C (1550°F). Below this tempera-
ture, gas phase alkali will condense on duct walls or on partlculate matter.
This would unrealIstlcally reduce the vapor phase alkali concentration to
which the turbine blades are exposed. For this purpose, four natural gas
Injection ports were located between the second and third cyclones. The
locations of the four ports were staggered to maintain gas temperatures
between 857°C and 900°C. The total amount of natural gas (typically 0.6 to
0.9 Nm3/m1n) 1s controlled to maintain a set third cyclone Inlet temperature.
The flow distribution between the four probes 1s manually controlled. The
Injection probe consists of a closed end 3/8 Inch alonlzed tube projecting
to the duct centerllne with a 0.10 cm diameter hole facing downstream. The
probes are purged with nitrogen when not 1n use.
The natural gas Injection system 1s frequently used to hasten thermal
equilibrium In the mlnlplant off gas system. Heating the refractory to
steady state would otherwise require 8 to 12 hours. With natural gas Injec-
tion, this 1s reduced to two hours. The effects of natural gas Injection
will be reported In a following section.
Temperature Control
The rate of solids feed to the combustor 1s automatically controlled 1n
order to maintain a specific operating temperature within the combustor. This
1s accomplished, through a series of controls which adjust the pressure
28
-------
differential between the combustor and the Injector vessel, thereby varying
the coal feed rate 1n a way to maintain the proper temperature, This system
1s described in detailed In a previous report (1).
Pressure Control
The FBC mlnlplant combustor has the capability of operating at pressure
levels of up to 10.5 atmospheres (140 pslg). Normally, pressure control Is
achieved by restricting the discharge flow of all the gas from the combustor,
so as to achieve an Increase 1n back pressure. This 1s done by use of a
silicon carbide converging nozzle Inserted 1n the discharge line. Adjustment
of the combustor pressure 1s accomplished by metering high pressure air Into
the discharge line just upstream of the flow nozzle. A 2 Inch-ball valve
equipped with a pneumatic positioner and actuator regulates the amount of
air added 1n response to a signal from the pressure controller. A more com-
plete description 1s contained 1n a previous report (1).
During the time that the combustor gas passed through the General Electric
turbine test section, this turbine section acted as the back pressure regulator
for most of the gas. The converging nozzle was disconnected during these
tests. However, to maintain proper pressure control, a small slip stream of
gas (-10%) was diverted prior to the test section and expanded through a smaller
variable nozzle. Makeup air was added just upstream of this nozzle through
the same valves and controllers that controlled pressure 1n the normal opera-
ting mode. This variable nozzle (Figure III-8) was adjusted to maintain 5 to
9 Nm3/min makeup air flow. The nozzle was constructed of 316 stainless steel
and performed satisfactorily for over 700 hours of testing.
Sampling and Analytical Systems
Gas Sampling System--
Flue gas is sampled at a point about 20 m downstream of the third stage
cyclone. The system Is designed to produce a solids-free, dry stream of flue
gas at approximately ambient temperature and atmospheric pressure whose com-
position, except for moisture, is essentially unaltered from that of the
original flue gas. The conditioned gas 1s analyzed in a series of continuous,
on-Hne analyzers for S02, CO, C0£, Og and NOX concentration. The flue gas
sampling and continuous analytical system was described in a previous report
(2). Another sample of flue gas can be extracted which has been filtered,
cooled and depressured, but not dried. This system 1s used to obtain batch
samples of flue gas for analysis of SOa and SOa by wet chemistry methods.
The method to determine $03 concentration was modified since the publication
of the previous report.
Previous measurement of $63 from the FBC mlnlplant consisted of flowing
the flue gas through an 80% Isopropyl alcohol (IPA) Impinger. Because of
potential problems with this method, such as possible contamination of IPA
lots with small quantities of peroxides which oxidizes SOg to SOa, collection
of S03 1s now accomplished by using a controlled condensation coll. The coil
1s maintained at 60eC, which is well above the water dew point and low enough
to remove most of the 503 from the gas phase.
29
-------
FIGURE 111-8
VARIABLE PRESSURE REDUCING NOZZLE
30
-------
Figure III-9 shows the sampling train as it is now used. Participate
material is removed at 9 atm by a Balston filter heated to 288°C, well above
the $03 dew point under FBC conditions. The cleaned gas is delivered to the
gas sampling train through a heated (288°C) stainless steel line followed by
heated teflon tubing. Collection of S03 as H2S04 is accomplished by using a
Goksoyr-Ross controlled condensation coil. The 80% IPA impinger is used as
a backup for 863 collection with IPA known to have a negligible peroxide
contamination.
H2S04 collected in the condensation coil and the IPA impinger is titrated
with barium perchlorate reagent and thorin indicator. Comparable results
were obtained with an acid/base titration using bromphenol indicator. Two
H202 impingers are used to trap S02 which is also determined using barium
perchlorate/thorin titration.
Particulate Sampling Systems
The miniplant has three particulate sampling systems. These systems were
installed or modified as the need for more and different particulate sampling
information arose. The oldest system referred to as the "Balston Filter Par-
ticulate Sampling System" is described in a previous report (1). The Balston
filter elements consist of borosilicate glass fibers bonded with an epoxy
resin. A newer system is capable of sampling at high temperature and high
pressure with an in-situ size distribution measuring device. This system is
referred to as the high temperature high pressure (HTHP) particulate sampling
system. It was constructed during this reporting period to sample the flue
gas with a 5-stage cyclone train developed by the Southern Research Institute
or a University of Washington 7-stage impactor. Both the Balston filter and
the HTHP sampling systems sample particulates in the gas leaving the tertiary
cyclone. The third system samples the particulates in the gas entering the
tertiary cyclone. This system, referred to as the Upstream Balston Filter
Particulate Sampling System, was only installed after run 94. It was used to
determine cyclone capture efficiency. The locations of all of these systems
is shown in Figure 111-10.
The old Balston Filter Particulate Sampling System has been modified
since its description in a previous report (1). Originally the system con-
tained three 18 cm Balston total filters in series followed by coolers, a
water knockout, flow control valve and rotameter. The secondary and tertiary
Balston filters were removed from the system. During the sampling it was
found that their weight gain contributed only a negligible amount to the
total loading and what was captured had the appearance of wet corrosion pro-
ducts. A wet test meter was also added after the rotameter. This allowed
for a more accurate measurement of sample flow than was previously possible
with a rotameter and a timer. A schematic of the system as modified at the
end of this reporting period is shown in Figure III-ll. This system has
provided a large data base of particulate measurements. It has been used with
probe sizes between 1.09 and 0.46 cm and isokinetic flow rates between 20 and
200 Ndm3/min, all with good results,
31
-------
SLIP
FROM
PFBC EXHAUST
FIGURE 111-9
CONTROLLED CONDENSATION COIL
ATTACHED TO A SO3-SO2
IMPINGER TRAIN
to
ro
S
TEFLON
TUBING
BALSTON
FILTER
STAINLESS STEEL MANIFOLD
HEATED SECTIONS
TEMPERATURE 288°C
CONTROLLED
CONDENSATION
COIL
.
1
!
00
FRITTED
WET TEST METER
0
0
0
0
"\
H^ H2O2 BLANK IPA
SO3-SO2 IMPINGER TRAIN
ICE
BATH
-------
FIGURE 111-10
MINIPLANT FLOW SCHEMATIC
OJ
CO
.^-~— >.
/^N
c
M
B
U
S
T
0
R
^i
<
N/G,
Kl //-»
k ) 2ND IN/ii
\X rvrinMF
1ST
Cyclone
sS
x^
1
1
N/G '
UPSTRI
BALST
EAM
ON -
1
G.E.
TURBINE TEST
HTHP
PARTICULATE
SAMPLING
J
FILTER
PAKIICULAIh T"
SAMPLING ..L
SYSTEM N/G
SYSTEM
\ /
V
(
SECTION
_ - -i
r >
1 ' 1 '
ALKALI '
\ — SAMPLING
SYSTEM
BALSTON
FILTER
PARTICULATE "~
SAMPLING
SYSTEM
. _ GAS SAMPLING
»-*-~ SYSTEM
x-v
"?_
t~*—>3
MAKE UP AIR
OFF GAS
COOLER
N/G i
ACUREX ESP
OR BAGHOUSE
N/G = NATURAL GAS INJECTORS
3RD
CYCLONE
t
-------
FLUE
GAS
1/2"
PROBE
CO
FIGURE 111-11
BALSTON FILTER PARTICULATE SAMPLING SYSTEM
TC
TEMP.
VALVE
BALSTON
Fll TFR
NLTtK
STEAM HEAT
EXCHANGER
WATER
CONDENSER
FLOW
MEASUREMENT FLOW CONTROL
ROTAMETER / VALVE
WATER '
KNOCKOUT
WET TEST METER (1 SCF/Rev)
-------
The HTHP sampling system is designed to obtain isokinetic, isothermal
samples of participate laden flue gas. The system consists of a long 0.95 cm
(3/8 inch) diameter tube projecting into the flue gas stream along the duct
centerline. Once outside the flue gas duct, the tube is heated with electric
heaters to maintain a set gas temperature. A high temperature (Kamyr) ball
valve provides system isolation during combustor operations. Following the
valve, the gas enters a pressure furnace which contains the sampling device.
This furnace is large enough for most in-stack sampling devices. It is
electrically heated to a controlled temperature (usually 1600°F) and pres-
surized with helium to prevent sample leakage or high temperature corrosion
of the steel. After the furnace, the gas passes through a backup Balston
filter and some gas coolers. Then the gas passes through a water knockout;
another filter to remove the condensed acid mist; and the flow control sys-
tem. Initially, the flow control system was automated to maintain a set flow
rate over a span of sampling device pressure drops. This automatic system
was prone to plugging and corrosion failure. The automatic flow controllers
were replaced with a manual valve. This valve has performed satisfactorily
for many samples. Following pressure let down, the gas passes through a gas
saturator and a wet test meter that measures total flow. Figure 111-12 is a
schematic of the original system. Figure 111-13 is a schematic of the system
as modified at the end of the reporting period.
One additional modification that was made to the HTHP sampling system,
that could not easily be shown in the figure, is the provision to bypass the
furnace and the small (7 cm) backup Balston filter and install an 18 cm
Balston total filter in their place. In this way, the probe and flow control
system can be used for sampling at 500°C with one 18 cm Balston filter as well
as at 850°C with the other devices. The advantage of the Balston filter is
rapid and reliable particulate concentration measurements.
The in-situ size distribution measuring devices used in the HTHP parti-
culate sampling system were a Southern Research Institute 5-cyclone train and
a University of Washington 7-stage impactor. The 5-cyclone train (Figure
111-14) was housed in the pressure furnace in the vertical position. Two
cyclones were upside down during sampling. A complete description of the
cyclones is given in a separate report by SoRI which also details cold flow
calibrations (7). The 7-stage impactor was also installed in a vertical
position, with a precutter cyclone before the impactor. The cyclone was there
to prevent overloading of the first stage due to large concentrations of
large particles. The impactor is described in a technical bulletin (8), a
picture is shown in Figure 111-15.
The third particulate sampling system was installed after run 94 to
measure particulate loadings entering the tertiary cyclone. This system was
built mainly to verify the high collection efficiency of the tertiary cyclone.
This efficiency was previously obtained by completing a mass balance with the
outlet filter sample and the captured lock hopper material. This sampling
system consists of a 6 mm probe projecting to the duct centerline. Filter
isolation is provided by two bellows seal valves (650°C temperature limit).
Immediately after the valves is the 18 cm Balston total filter. Following the
filter a large cooling coil provides natural convective cooling. This is
followed by a water knockout, manual flow control valve, gas saturator and
35
-------
FIGURE 111-12
ORIGINAL HTHP PARTICULATE SAMPLING SYSTEM
PROBE I
NITROGEN X
PURGE T_
HEATER
KAA)
FLOW
CONTROLLER
SAMPLE
NITROGEN
PURGE
*
CO
o>
BACK PRESSURE
REGULATOR
0!
-tk-L
t
BALSTON
FILTER
WATER
KNOCKOUT
BACK PRESSURE
'REGULATOR
TEMP.
VALVE
WATER
GAS
SATURATOR
SAMPLING
DEVICE
HEATER
VENT
PRESSURE
VESSEL
WET TEST
METER
HEATER
VESSEL
PRESSURIZING
AIR
FLUE
GAS
-------
FIGURE 111-13
HTHP PARTICIPATE SAMPLING SYSTEM AS MODIFIED
BALSTON
CO
--4
HIGH TEMP
FILTER
FLOW CONTROL
VALVE
WATER
KNOCKOUT
GAS
SATURATOR
BALSTON
V
OFF GAS
COOLER
1
VATER
1-ILItK
(BACK
UP)
1
VENT
SAMPLING
DEVICE
HEATER
WET TEST METER
PROBE
NITROGEN I
PURGE X
HEATER
SAMPLE
NITROGEN7HIGH rAAAn
. PMRnp 4TEMP r v v v I
VALVE
HEATER
PRESSURE
VESSEL
—H!
VESSEL
PRESSURIZING
Helium
3/8"
PROBE
FLUE
GAS
-------
FIGURE 111-14
SOUTHERN RESEARCH INSTITUTE 5-CYCLONE TRAIN
38
-------
FIGURE 111-15
UNIVERSITY OF WASHINGTON 7-STAGE IMPACTOR
39
-------
a wet test meter. The flow 1s set to 1sok1net1c conditions by timing the gas
flow through the wet test meter. This system has verified efficiencies
obtained earlier by the mass balance technique. Low flow rates are required
due to the limited natural convectlve cooling.
Alkali Probe Train
An alkali probe train was designed and constructed by Exxon Research to
enable acquisition of a hot pressurized flue gas sample before the turbine
test section (Figure 111-10). The probe was designed to measure vapor phase
Na and K concentrations of the flue gas at 840°C and 9 atm pressure. Figure
II1-16 1s a schematic of the alkali probe train. The temperatures shown are
those for test 4 of run 78; these are representative of normal operating
conditions.
The hot pressurized flue gas enters the probe at system temperature and
pressure and 1s preflltered through three layers of astroquartz. Na and K
vapors 1n the flue gas are then condensed on the walls of an air cooled quartz
tube, which lowers the flue gas temperature from approximately 840°C to
200°C. Alkali metals which condense on partlculates In the flue gas are col-
lected In a Balston filter after exiting the quartz tube. Flow 1s manually
controlled by a needle valve and measured by a wet test meter.
Process Monitoring and Data Generation System
Data characterizing the mlnlplant operating conditions are recorded on
5 multipoint recorders. In addition, at one minute Intervals, the same out-
put 1s recorded by a data logger system consisting of a D1g1trend 210 data
logger with printer and a Kennedy 1701 magnetic tape recorder. Approximately
100 pieces of data are logged with one-half Involving temperature measurement
while the rest deal with pressure, material flow rate or flue gas composition,
The system 1s described 1n a previous report (1).
Signals from the data logger are scanned every minute and are stored on
magnetic tape. The magnetic tape, containing about 6000 Items of data per
hour of run time, 1s fed to a computer which converts the logger output to
flow rates, pressures, etc. with the proper dimensions. The data are then
averaged and standard deviations are calculated over preselected time Inter-
vals (usually 10 to 30 m1n.). other quantities are also calculated. This
Includes average bed temperature, based on four thermocouple readings covering
the 15-114 cm Interval above the flu1d1z1ng grid, superficial gas velocity,
excess air, as well as the Important gas concentrations.
Combustor Safety and Alarm System
A process alarm system was designed to warn of Impending operational
problems. Two general alarm categories exist. The first, dealing with less
critical situations, alerts the operator of the problem so that appropriate
corrective action can be taken. The second class of more critical alarms
results In the Immediate or time delayed shutdown of the complete system or
specific subsystems. An alarm condition is brought to the attention of the
40
-------
FIGURE 111-16
SCHEMATIC OF ALKALI PROBE TRAIN*
25°C
VENT
100 kPa
NEEDLE VALVE
WET TEST
METER
FLOW RATE
0.06
(2 SCFM)
SECOND
BUBBLER
710kPa T>70°C T >210°C
(160°F) (410°F)
FIRST
BUBBLER
AIRCOOLED
S.S. PROBE
=-£
T > 877°C
(1610°F)
^—FLUE
GAS
/
/ QUARTZ
*- TUBE
BALSTON
FILTER
KNOCK OUT
FRONT
FILTER
* TYPICAL TEMPERATURES AND CONDITIONS DURING TEST 4 OF RUN 78
-------
operators by a flashing light above the control panel accompanied by a high
pitch sound. The sensitivity of the Individual alarms 1s controlled by pot-
entiometers located beneath the control panel. The system 1s described 1n a
previous report (1).
Coal and Sorbent Properties
Coal —
Coals used 1n the mlnlplant variables study were a high volatile bitum-
inous coal from the Consolidation Coal Company's "Champion" preparation plant
1n Pennsylvania, an Illinois No. 6 seam coal obtained from Carter Oil Company's
Monterey No. 1 mine, and an Ohio coal obtained from the Valley Camp mine.
The Champion coal was partly classified to remove fines smaller than 50
U.S. Mesh. The Illinois coal was screened to 6 x 40 U.S. Mesh to prevent
plugging of the primary injector feed vessel. Partlculate size distribution
and composition data for the coals are shown in Figure 111-17 and Table III-l
respectively. The Valley Camp coal was only used on a very limited basis
during a coal strike. This coal was purchased from Curtiss-Wright and no size
analysis was obtained. The coal was prescreened 6 x 50 U.S. Mesh.
Sorbent
Sorbent—
Grove limestone (BCR No. 1359) and Pfizer dolomite (BCR No. 1337) were
the primary sorbents used in the mlniplant variables study. The composition
of these stones 1s given in Table III-2. Most of the runs were made with the
stone screened to 8 x 25 U.S. Mesh to give the distribution of dolomite shown
1n Figure 111-18. The limestone was only used on a limited basis, no size
distribution was obtained.
Operating Procedures
Prior to initiating a run, a detailed checkout procedure Is followed to
Insure that the system is ready for operation. This includes various equip-
ment checks, alarm system checks, calibration of flue gas analyzers, activa-
tion of process monitoring and control systems, and the turning on of all
cooling water systems. All runs are begun with an initial bed of sorbent In
the combustor. This consists of either a charge of sulfated limestone or
dolomite, or the bed material from the previous run. Fresh, uncaldned lime-
stone has been used, but is not preferred.
The first operation of start-up involves heating the gently fluidlzed
sorbent bed by burning natural gas in the burner plenum followed by injection
of kerosene into the bed. Once the bed temperature is ~650°C, coal injection
is begun. As soon as coal ignition is confirmed by a sharp temperature
increase, the other fuels are shut off. Combustor temperature, gas flow rate
and pressure are than rapidly Increased to their operating values. A more
complete procedure 1s contained In a previous annual report (1).
42
-------
FIGURE 111-17
COAL PARTICLE SIZE DISTRIBUTION
CO
PARTICLE SIZE, \im
-------
TABLE III-1. COAL COMPOSITION
Run Number
60-65
66-70
71
72
73-75
76 and 77
78.1
78.2-78.10
79
80
81-88
89-96
97,98,99.1,
99.2,99.5-99.7
99.3 and 99.4
100
101 and 102
103 and 104
105
106
107-115
Coal Type
Champion
Illinois
Illinois
Ohio
Ohio
Ohio
Illinois
Illinois
Illinois
Illinois
Illinois
Champion
Champion
Illinois
Champion
Champion
Champion
Champion
Illinois
Champion
Moisture
2.19
1
1
1
1
1
3
12
12
12
12
2
3
13
3
2
2
2
11
2
.96
.60
.76
.82
.8
.16
.82
.5
.86
.9
.3
.12
.0
.12
.28
.94
.52
.36
.40
Ash
12.50
9.68
9.47
8.76
7.27
6.77
11.57
8.88
7.98
8.54
8.12
9.82
12.20
9.18
8.91
6.64
8.61
6.97
7.89
8.34
Ultimate Anlysls
Total
Carbon H S N Cl
71.35
67.35
68.73
73.41
74.33
75.13
65.58
57.76
60.63
67.67
58.81
71.91
68.04
60.34
70.44
74.35
72.43
79.41
67.39
--
4.74
5.40
5.41
5.23
5.35
5.52
5.63
5.02
5.48
5.81
4.78
5.24
4.76
4.70
4.88
5.23
5.20
5.87
5.10
--
1.40
4.00
4.00
2.84
2.63
2.47
4.17
3.52
3.33
3.53
3.37
1.65
1.92
3.4
1.87
1.38
1.61
1.64
3.52
1.84
1
1
1
1
1
1
2
2
1
1
1
1
1
1
1
1
1
1
1
.54 0.07
.24 —
.24 -
.35 -
.29 --
.16 -
.41 -
.26 -
.17 --
.29 --
.09 -
.43 -
.42 -
.02 -
.48 -
.58 —
.49 -
.47 -
.33 --
—
6.31
10.37
9.58
6.70
7.93
7.18
7.55
11.0
10.11
11.99
10.94
7.70
8.63
9.65
9.41
8.58
7.81
2.04
4.43
—
Heating
Value
Ib/BTU
12,514
12,575
12,645
13,171
13,351
13,711
12,042
10,988
11,043
11,528
10,892
12,809
12,581
10,912
12,984
13,728
13,151
13,612
11,969
12,874
-------
TABLE III-2. SORBENT COMPOSITION
Quarry
Grove
Pfizer
Sorbent Type
Limestone
Dolomite
Weight Percent
CaO
94.64
55.78
MgO
1.01
37.04
S102
0.04
0.59
A1203
0.46
0.45
Fe203
0.117
0.23
-------
FIGURE 111-18
CO
o
8
DOLOMITE SIZE DISTRIBUTION
Run 100
T—I—I T
—• »o oo NJ a-
88888
o
g
o
>
Z
PARTICLE SIZE,
-------
EXPERIMENTAL RESULTS AND DISCUSSION
Combustor Operations
During the month of August 1979, experimental miniplant studies under the
present contract were completed with run 115. The combustor was shut down
after completing over 3700 hours of coal combustion operation. Twelve runs
were made of over 100 hours duration; five of these were over 200 hours dura-
tion. The longest continuous run of the combustor was 250 hours long.
The miniplant also made many short runs in rapid succession. During the
ceramic bag filtration studies, fourteen 8-10 hour runs were completed in
fifteen working days. This was possible because of the rapid startup proce-
dure developed for the miniplant. Coal was typically burned 1 to 1-1/2 hours
after the start of the first shift.
The ability to turn the unit around rapidly was also demonstrated. For
example, during the month of June 1979, the combustor was operated for nearly
400 hours in 3 run segments for two different test series. These runs were
coordinated with two other EPA contracts (Acurex and GCA Corp.). Several runs
were also made in a "piggyback" fashion with up to three other EPA contractors
participating.
Although the operation of the miniplant became rather routine, some
operating problems still occurred. The impact of these problems was usually
minimized by employing a number of operating procedures which were developed
over a period of time.
Initially coal feeding was one of the more serious operating problems.
After a number of equipment modifications described in an earlier report (3),
the system performed satisfactorily. Some problems still persisted and were
dealt with by modified operating procedures. The erosion of the coal feed line
described in an earlier section of this report was such a problem. The 316
stainless steel line eroded through after about 200 hours operation at a bend
in the line. The impact of this problem was minimized by replacing all bends
in the coal feed system as soon as one bend eroded. This required only 10-15
minutes, during which time the combustor was fed liquid fuel. Before long
runs, entirely new bends were installed, further reducing run interruptions.
Coal feed plugging was only a minor problem during this period. Plugging
of the coal feeder would occur occasionally if coal was fed with a combination
of high moisture and high fines content. Feeding of coal with a moisture con-
tent as high as 112 was possible if fines smaller than 420 ym (40 U.S. Mesh)
were removed. Drier coals with moisture contents of 1 to 3% were fed without
incident when fines less than 300 ym (50 U.S. Mesh) were removed. The drier
coals were easier to process and generated less waste as fines.
The availability of properly processed coal was much more a problem than
coal feed plugs. No supplier of dry Illinois No. 6 coal was found for the
quantities needed. The high moisture coal was used because of this. Unpre-
pared Illinois coal was dried by spreading it in a parking lot and raking it.
47
-------
Coal prepared "In this manner contained 11% moisture. The supplier who did this
work elected to discontinue after processing only our most urgent supply needs.
Regeneration runs were made with low sulfur Champion coal because no supplier
of prepared Illinois could be found.
Problems associated with the back flow of hot bed solids through the coal
transfer line into the coal injector vessel became severe during the testing
of the granular bed filter. During the filter blow back cycle, a high pressure
pulse would upset the pressure differential between the combustor and the
injector vessel. Hot solids would back flow to the injector vessel and ignite
the coal in the vessel. Several runs were terminated for this reason. The
problem, however, disappeared when testing of the granular bed filter stopped.
Corrosion of the wet scrubber used to ensure compliance with EPA emission
standards was a continual problem. It required frequent repairs during turn-
around times. The severity of the problem was lessened with the injection of
NH3 into the scrubber to neutralize the acids formed.
Operations during winter months were difficult. Because of the start/stop
nature of many of the mini pi ant runs, frequent draining and blowing of water
systems was necessary. Even with these precautions, freeze ups still occurred.
Productivity during the winter was much lower for these reasons.
Small problems with the main fluidizing air compressor frequently required
attention during a run. One failure during a run gave a chance to observe
the effect of bed slumping and hot start. When the compressor failed, it
stopped all fluidizing air to the bed. The temperature profiles during this
period are shown in Figure 111-19. The unit was restarted within 1/2 hour
after repairs were made to the compressor. The restart was completely suc-
cessful with no bed agglomeration.
Hot Corrosion/Erosion Testing of Materials--
During this reporting period, a total of 1117 hours of combustor operation
were devoted to "Hot Corrosion/Erosion Testing of Materials for Application to
Advanced Power Conversion Systems Using Coal-Derived Fuels." This work was
sponsored by DOE, under a cooperative agreement with EPA, to test the effect
of exposing gas turbine blade and boiler tube materials to real PFB conditions.
The gas turbine specimen test section was designed and built by General
Electric. The air cooled boiler tube probes were designed and built by
Westinghouse. The evaluation of results from these two systems is the res-
ponsibility of General Electric and Westinghouse. The combustor flue gas
system was modified as shown 1n Figure 111-20 to permit good pressure control
in the presence of the turbine test section. Most of the flue gas was expanded
through the turbine test section; 10% was diverted for pressure control.
A picture of the blade specimens after 1000 hours of exposure is shown in
Figure 111-21 . As can be seen 1n the photograph, little visible erosion is
present. Further metallographic examination is being done by General Electric.
Results will be reported by them 1n the future.
48
-------
FIGURE 111-19
SELECTED COMBUSTOR TEMPERATURES DURING CRASHDOWN AND RESTART
RESTART
• Primary Cyclone Discharge
A 3rd Cyclone Inlet
• Port #9 (in-bed)
Numbers refer to height above
fluidizing grid, (cm)
1415
1400
1345
1330
o
o
450
- 350
CLOCK TIME
-------
FIGURE 111-20
MINIPLANT FLOW SCHEMATIC
HOT CORROSION/EROSION TEST CONFIGURATION
en
o
p.
c
0
M
B
S
T
O
R
1st
\ /
\
Cyclone
x
S
/
/
N/G-<
i
2ND
' CYCLONE
h
VG-^
1
ALKALI
HTHP PR°BE
PARTICULATE TURBINE TEST
PROBE SECTION
- OFF GAS
T COOLER
1
N/G
I 1 I » t_ -J
-------
FIGURE 111-21
GE TURBINE BLADE SPECIMENS
L
1000 HOURS
S T CASCADE
I SECOND CASCADE
FOURTH CASCADE
1 C"
51
-------
Twenty-eight of the air cooled boiler tube probes designed by Westinghouse
were exposed for up to 1117 hours. These were controlled at specified tem-
peratures, and located both in and above the fluidized bed. Each probe con-
tained two test materials. A picture of two probes is shown in Figure 111-22.
Results of the metallographic examination of the specimens will be reported
separately by Westinghouse in the future.
Exxon will report run conditions including particulate loadings and
other exposure conditions in a report to DOE later in 1979.
Effect of Natural Gas Injection--
Natural gas injection is used in the miniplant to maintain temperatures
above 840°C (1550°F) for the DOE corrosion tests as well as to shorten the
time necessary for thermal equilibrium in downstream components. Without
natural gas injection, the temperatures at the tertiary cyclone would only
reach their steady operating values after about 12 hours of operation. With
natural gas injection this is reduced to 1-1/2 to 2 hours. For these reasons,
natural gas injection is used routinely during miniplant operations when
tests of components outside the combustor are run.
The effects of natural gas injection on flue gas composition and parti-
culate characteristics have been investigated. One concern is that instan-
taneous combustion of the gas would cause hot spots, which could increase
formation of NOX, or alter particulate characteristics. Temperature measure-
ments have shown that the gas does not begin to burn until it is several cen-
timeters downstream of the injection point. Combustion then occurs uniformly
over the next 2 meters before it is complete as indicated by a temperature
decline. The temperature profile for the injection of natural gas can be
seen in Figure 111-23. During this study the amount of natural gas normally
distributed among 4 probes was injected in only the first probe. During
normal operations, the temperature rise is no more than 20°C, however, the
shape of the profile for each probe is similar.
The only effect of natural gas injection on flue gas composition that has
been measured besides a 0.5 to 1.0% reduction in flue gas oxygen is a drastic
reduction in CO. During run 81, CO emissions fell from 525 ppm 1-1/2 hours
after coal injection was begun to 30 ppm after natural gas injection into the
flue gas lines was started. Reduction in CO emissions was also observed in
runs 79 and 80 with natural gas injection. CO emissions observed in the past
for miniplant runs without natural gas injection were in the range of 100 to
200 ppm. For the runs with natural gas injection, the residence time of the
flue gas in the piping at the temperature range of 815 to 940°C was about
twice as long as for runs without natural gas injection. Thus, natural gas
injection could provide increased opportunity for burnout of CO.
The effect of natural gas injection on flue gas particulates is also of
concern, especially if the gas is responsible for the high collection
efficiencies measured with the tertiary cyclone. Studies in the miniplant
have shown that cyclone collection efficiencies are not significantly affected
by the injection of natural gas. The slight effect that is measured is usually
explained by the effect of temperature on the cyclone inlet velocity.
52
-------
FIGURE 111-22
WESTINGHOUSE BOILER TUBE PROBES
After 1117 Hours Exposure
(850°C In-Bed)
Top
ALLOY --
Haynes 188
Hasten oy X
Bottom
Top
ALLOY —
Hasten oy X
Haynes 188
Bottom
53
-------
FIGURE 111-23
NATURAL GAS INJECTION - TEMPERATURE PROFILES
T
T
900
850-
U
o
10
DISTANCE DOWNSTREAM OF 2ND CYCLONE (M)
76.7
METHANE
FLOW
(Ndm3/MIN)
15
-------
$02 Emission Control
As a result In a change 1n the New Source Performance Standards for coal
fired utility boilers, the allowable S02 emission level was decreased from
1.2 Ib/MBTU to a sliding scale which, for higher sulfur coals, requires 90%
SOz retention. A series of runs were made to demonstrate that PFBC could
attain the 90% SOg retention level using dolomite or limestone sorbents. Ohio
and Illinois coals, containing 2.5 and 4% sulfur respectively were tested.
Ca/S molar ratio was the primary variable in the study. With dolomite sorbent,
the Ca/S ratio was varied from 1.5 to 2.1. With limestone, it was varied from
3.7 to 7.6, Superficial gas phase residence time in these tests was varied
slightly from 1.6 to 2.2 s. Other conditions are summarized in Table III-3,
which also tabulates the S02 retention results. As seen, 90% SO? retention
can easily be reached with either sorbent. Retention levels as high as 99+%
were measured. As expected, dolomite is more active. A Ca/S ratio of 1.5
will assure 90% S02 retention at a gas residence time of about 2s while lime-
stone use will require a Ca/S ratio of between 3.5 and 4.0.
In subsequent runs, additional data were obtained in the high S02 reten-
tion area. All SO? retention data were then plotted against Ca/S ratio for
both dolomite and limestone sorbents and analyzed. In the plots, the Ca/S
ratio set on the coal/sorbent blender was used as the correlating parameter.
In earlier runs, a Ca/S ratio calculated by mass balance was used because of
blender speed control problems. Since the problems were solved by installing
an electronic speed control system on the blender, it was decided to use the
"set" Ca/S ratio since it appeared to be more reproducible, possibly because
it did not require reliance on a series of chemical analyses of the solid and
gaseous products.
Results with Dolomite Sorbent—
Figure 111-24 gives S02 retention results using Pfizer dolomite sorbent,
Champion, Illinois and Ohio coals, at gas residence times from 1.5 to 2.5s.
Retention results shown in the figure were corrected to a residence time of 2s
using the simple first order rate expression described in the previous report
(2). Figure 111-25 gives results of runs made with a residence time of 2,5
to 3.5s, corrected to a residence time of 3s. Since a significant amount of
data scatter occurs on each plot, values from the smooth curves were used to
construct a curve relating the first order rate constant for the S02 sorbent
reaction, k, to the sorbent utilization. A set of values for k was calculated
from the first order rate equation by using a range of retentions at constant
residence times. The sorbent utilization was obtained from the smooth curves
by using the expression for a sulfur balance:
% Ca Utilization * S02 retention (%)
ratio
If the S0£ reaction rate is indeed first order, then the values from the
curves in Figures 111-24 and 111-25 would overlay on a k vs. utilization curve,
all other variables being equal. However, when this was done, two k vs. uti-
lization curves were obtained, one for each residence time. The reason for
this was that the 3s residence time data did not show a significant improve-
ment in S02 retention over the 2s data, as would be predicted by the first
55
-------
TABLE I I 1-3. RUN SUMMARY 90$ S02 RETENTION STUDY
Nominal Operating Conditions Run 68 Run 69 Run 70 Run 71 Run 72 Run 73 Run 74.1 Run 74.2 Run 75
Run Length (hrs) 10 11 11 13 13 11.5 4.5 5.75 9
Pressure (kPa) 930 930 930 930 930 930 930 930 930
Temperature (°C) 946 947 936 933 943 939 933 937 936
Superficial Velocity (m/s) 1.7 1.8 1.9 1.7 1.8 1.8 1.7 1.7 1.7
Expanded Bed Height (m) 2.8 3.0 3.0 3.8 3.7 3.8 3.7 3.7 3.8
Gas Residence Time (s) 1.6 1.7 1.6 2.2 2.1 2.1 2.2 2.2 2.2
Excess A1r («) 17 14 23 13 11 10 16.5 18.9 17
Coal Feed Rate (kg/hr) 127 132 127 122 116 117 113 112 115
Ca/S Molar Feed Ratio 2.0 1.5 1.5 1.5 1.8 2.1 7.6 4.8 3.6
Sorbent PD PD PD PD PD PD 6L GL GL
Coal IL IL IL IL OH OH OH OH OH
<* Emissions
S02 (ppm) 50 197 94 60 60 4 1 30 170
NO' (ppm) 79 56 81 40 30 50 80 60 50
02 (%) 3.2 2.6 4.0 2.4 2.1 2.0 3.3 3.7 3.4
C0? (%) 16 16 17 17 22 18 19.5 19 15.2
CO2 (ppm)* ............ ......
S02 Retention 98.5 94.2 96.9 98,1 99.0 99.7 99.9 98.2 90.9
Notes: PD * Pfizer 1337 Dolomite
GL = Grove 1359 Limestone
IL * Illinois Coal - 4% S
OH * Ohio Coal - 2.5% S
* CO Analyzer Malfunctioned
-------
FIGURE 111-24
SO2 RETENTION ADJUSTED TO 2 S. GAS RESIDENCE TIME (tg).
DOLOMITE RUNS HAVING ACTUAL tg BETWEEN 1.5 AND 2.5 S.
100
90-
80-
70-
z
o
*• 60
CM
50-
40-
30
0
Ca/S MOLAR RATIO
57
-------
10
90
80
Z
o
70
CN
60
FIGURE 111-25
SO2 RETENTION ADJUSTED TO 3 S. GAS RESIDENCE TIME (tg).
DOLOMITE RUNS HAVING ACTUAL tg BETWEEN 2.5 AND 3.5 S.
T
T
50
40
30
t
0
Ca/S MOLAR RATIO
58
-------
order expression. A plot of the mean k vs. utilization curve derived from the
two retention curves is presented in Figure 111-26. This curve is a compromise
between the upper and lower limits formed by the 2s and 3s residence time data.
At present, there is no explanation for the lack of uniformity in the two
groups of data, but one obvious observation would be that the reaction does not
strictly comply with first order kinetics. This may be a problem inherent in
large scale FBC operations where additional operating parameters, such as
solids recycle and rejection, compound the difficulty of obtaining representa-
tive samples necessary to characterize the system.
To study the effect of residence time on S02 retention, the k vs. utiliza-
tion curve in Figure II1-26 was used to construct retention curves for four
residence times. 2s and 3s retention curves were calculated and plotted with
the miniplant data curve in Figure 111-27. As seen, the two calculated surves
straddle the data curve reasonably well, indicating that effects of gas resi-
dence time on S02 retention become more difficult to discern as residence time
Increases. This can be seen more clearly in Figure 111-28, where retention
curves for all four residence times were drawn. Here it becomes obvious that
below 2s, residence time has a significant effect on SOg retention, while
above 3s this effect would necessarily become less substantial, particularly
at Ca/S ratios greater than 1.
Additional work is obviously needed before the effect of gas phase resi-
dence time as well as other kinetic parameters can be accurately predicted.
Figure 111-28 can be used as an approximate guide until a better understanding
is developed.
Results with Limestone Sorbent--
Figure 111-29 gives S02 retention results obtained with limestone sorbent
at temperatures high enough to calcine the limestone substantially. As seen
in the previous figures with dolomite, the data scatter is significant. With
limestone, the degree of calcination and porosity of the calcined stone are
other parameters in addition to the kinetic effects discussed above. Again,
more work is needed to improve understanding of the desulfurization system.
It should also be mentioned that limestone not only is less active than
dolomite under normal PFB combustor operating conditions, but becomes com-
pletely inactive at temperatures at or below 760°C (2). This must be con-
sidered if limestone is to be used and 1f low temperatures are expected, for
example, to aid in decreasing the steam output from the combustor.
S02 Dynamic Response
The new S02 emission standards call for 90% retention of the S02 from new
coal fired utility boilers. This retention must be averaged over 30 days. It
is expected that utilities would operate at slightly higher retentions to avoid
exceeding the new limit and avoid costly violations. The incentive therefore
exists to develop a system to control S02 retention close to the desired level
without large safety margins. With this goal in mind, General Electric has set
out, under contract to the EPA, to develop a control system that would adjust
sorbent feed rates to assure a specified S02 retention despite variations in
59
-------
FIGURE 111-26
Z
o
u
as
Q
Of
O
FIRST ORDER RATE CONSTANT VS. CALCIUM UTILIZATION
MINIPLANT DOLOMITE RUNS
(APPROXIMATE ONLY)
3.
3.
2.5
2.0
1.5-
1.0
.5
T 1 1 Tl 1 1 1 1 r
_L
_L
J.
J_
0 10 20 30 40 50 60 70
CALCIUM UTILIZATION (%)
_L
J.
80 90 100
60
-------
FIGURE 111-27
MEASURED AND CALCULATED SC>2 RETENTION VS. Ca/S FOR
MINIPLANT DOLOMITE RUNS
100
90
80
g
t—
LLJ
70
cs
60
50
40
30
CALC. FOR tg = 3 S.
MINIPLANT DATA
FOR tg = 2 to 3 S.
CALC. FOR tg = 2 S.
0.5
1.5 2
Ca/S MOLAR RATIO
2.5
61
-------
FIGURE 111-28
CALCULATED SO2 RETENTION VS. Ca/S
MINIPLANT DOLOMITE RUNS
1.0 1.5 2.0
Ca/S MOLAR RATIO
2.5
3.0
62
-------
100
90-
80-
, 70
Z
o
FIGURE 111-29
SO2 RETENTION VS. Ca/S RATIO FOR
LIMESTONE NO. 1359 AT CALCINING CONDITIONS
tu
Of
CM
8
60-
50
40
0
345
Ca/S MOLAR RATIO
8
63
-------
the sulfur content of the coal, and In the reactivity of the sorbent. Tests
were carried out on the miniplant in cooperation with GE, to determine the
emissions response to a step change in either coal sulfur or dolomite feed
rate, with other variables held constant.
Experimental Equipment and Procedures--
At GE's request, the nvfniplant was run with no primary cyclone recycle
during the S0£ response tests. The combustor was allowed to reach a steady
state for 4 to 5 hours before a change of condition was made. During this
period, the old, continuous, primary feed system was used. Once steady state
emissions were achieved, the change was made to the auxiliary feed system
containing a different coal/sorbent blend. The auxiliary feeder was used for
8 hours, the approximate capacity of the vessel. After 8 hours on the
auxiliary feeder, the change was made back to the primary feeder and the
original coal/sorbent blend. The combustor was then run for 18 to 20 hours
with the primary feeder. The 18 hours were required to assure two changes of
the bed material Inventory.
The procedure for changing from one feeder to another consisted of:
1. Calibrating the analyzers 30 minutes before the change.
2. Closing the valve at the bottom of the active feeder.
3. Allowing the coal transport Hne to clear of sol Ids (~20 sec.)
4. Changing over AP taps, transport air and electronic controls
to the other vessel.
5. Opening the valve at the bottom of the other vessel.
6. Confirming coal feed by non-zero SOg emissions and combustor
temperature increase.
The time required for the change of feed vessels was 40 to 80 sec, depending
on crew experience and operating conditions.
Results and Discussions —
Shakedown—Shakedown of the auxiliary feed system went smoothly during
runs 97 and 98. Several modifications made after run 97 allowed the time
required for the change to be reduced from 4 minutes to 40 sec during run 98.
The resultant combustor temperature drop was reduced from 260 to 70°C. Table
III-4 lists the results of step changes in the coal source during shakedown
as well as 1n the actual runs. The time listed 1s the time required from the
confirmation of coal feed by non-zero SOg emissions to the first time the S02
emission reached the equilibrium value. A large, short S02 "spike" was
Ignored in this summary since it was due to the rapid combustor temperature
drop and subsequent coal feed rate increase caused by the action of the coal
feed controllers. These temperature fluctuations may be unique to the mini-
plant and may not appear in a commercial installation.
Change in Coal Sulfur—The first actual response test conducted on the
minlplant (run 99) was a step change in coal sulfur content at a constant
dolomite mass feed rate. The minlplant was started with Champion coal and
dolomite (1.9% sulfur, Ca/S = 1.40). After 4 hours, the combustor was switched
to the auxiliary feeder which contained Illinois No. 6 coal and dolomite (3.4%
sulfur, Ca/S » 0.76). Then, after 8 hours, the combustor was switched back to
64
-------
TABLE III-4.
S02 DYNAMIC RESPONSE SUMMARY
in
Run #
97(D
970)
98(D
98(D
98^
98^
99
99
100(3>
100(3>
Type
of
Change
% S
%-S
% S
% S
% S
% S
% S
% S
ca/s
Ca/S
/0\
From(2)
Champion
Illinois
Champion
Illinois
Champion
Illinois
Champion
Illinois
Ca/S = 1 .43
Ca/S =0.38
To
Illinois
Champion
Illinois
Champion
Illinois
Champion
Illinois
Champion
Ca/S =0.38
Ca/S = 1 .43
S02
Before
(ppm)
90
675
225
675
225
675
225
675
375
500
S02 at Steady
State (ppm)
675
225
675
225
675
225
675
225
500
300
Time from
Feed Change to
Steady State (min)
6-3/4
7
6
4
5
7-1/2
7-1/2
8-1/4
110
200
(1) Shakedown Runs
(2) Champion Coal - 1.92% Sulfur (Ca/S = 1.40 during tests where % S changed)
Illinois No. 6 coal - 3.4% Sulfur (Ca/S = 0.76 during tests where % S changed)
(3) Champion Coal Used
Pfizer dolomite used in all tests
-------
the Champion coal, dolomite blend for 20 hours. The S02 emission, averaged
over 10 minute intervals, over the entire run is shown in Figure 111-30.
Figures 111-31 and 111-32 show the continuous S0£ analyzer output immediately
before and after the two changes in coal sulfur. The response times of 7.5
and 8.2 minutes were determined from these figures. During the change to the
higher sulfur coal, the dolomite mass feed rate actually increased 19% although
it was intended to hold it constant. The effect of a 19% increase in dolomite
mass feed rate was superimposed on the 104% increase in coal sulfur mass feed
rate occurring because of the change. The change back to the lower sulfur coal
also had a 19% decrease in dolomite mass feed rate. The unintended change in
the dolomite mass feed rate was caused by slight changes in the sulfur content
and heating value of the coal actually used, versus prior analyses of other
batches used to set conditions.
During the combustion of Champion coal, S02 emissions were close to the
predicted level of 170 ppm. This prediction comes from Figure 111-25 of the
previous section. The predicted SQ2 emission for the Illinois No. 6 coal is
1080 ppm, much higher than the 725 ppm measured.
Change in Sorbent Feed Rate—During run 100, the S02 emission response to
a setp change in Ca/S from 1.43 to 0.38 using Champion coal and dolomite was
investigated. The combustor was started with the higher Ca/S which was then
rapidly changed to the lower Ca/S by switching to the auxiliary feed system.
The first attempt to switch to the auxiliary feed vessel failed due to a plug
in the transport line. The plug was cleared while the combustor was maintained
on liquid fuel. The low Ca/S coal was fed from the auxiliary vessel for 20
minutes to determine the likelihood of another plug. The combustor was switched
back to the high Ca/S blend until the SOg emission stabilized at the preswitch
level (about 2 hours). The combustor was again switched to the low Ca/S blend
in the auxiliary feeder. After 8 hours of operation on the low Ca/S blend,
the change was made back to the higher Ca/S blend. During this final run seg-
ment, coal was fed for 19 hours to assure that steady state had been attained.
From Figure 111-33 the response time for the step from Ca/S = 1.43 to 0.38 is
shown to be 110 minutes. The reverse step required 300 minutes to achieve
steady state. It is unexplained why the original steady emission, at startup
(about 375 ppm), was not the same as the final segment emission (300 ppm). At
no time during the run was the average SOg emission as close to the predicted
emission as expected. The predicted emission for Ca/S = 1.40 Is 170 ppm. The
actual average emission during the 19 hour final portion of the run was 300
ppm. During the low Ca/S portion (mid portion) of the run, the actual average
emission was only 500 ppm compared to 750 ppm predicted. This may be due to
the fact that the low Ca/S operating region has not been well defined experi-
mentally. The time required during the second change (to the higher Ca/S) to
reach the "startup emission" of 375 ppm was approximately 120 minutes. This
suggests that there is a reversibility in the response times for these
changes. The SOg emission just prior to and just after the coal sorbent blend
changes can be seen in Figures 111-34 and 111-35. The immediate effect of
both changes ts not noticeable since the response times are fairly long.
Conclusions—
For the typical 90% S02 retention levels expected to be used in most pres-
surized fluidlzed bed combustors, the response to 100% increase in coal sulfur
66
-------
FIGURE 111-30
CT»
Q_
O.
z
O
CN
1100
1000
900
800
700
600
500
400
300
200
100
CHANGE IN COAL SULFUR/SO2 EMISSIONS (RUN 99)
PREDICTED
EMISSION
170 PPM
COAL FEED
UPSETS
CHAMPION
COAL
Ca/S=1.40
I
PREDICTED EMISSION
1080 PPM
i r
8.2 MINUTE
RESPONSE
7.5 MINUTE
RESPONSE
ILLINOIS NO. 6 COAL
Ca/S = 0.76 I
i i i I i i i i
I
CHAMPION COAL Ca/S= 1.40
i i i I i i i i I i i
0
10 15 20
TIME INTO RUN (HRS)
25
30
-------
FIGURE 111-31
en
CD
+22.5
-22.5
INSTANTANEOUS SO2 RESPONSE (RUN 99)
(CHAMPION TO ILLINOIS NO. 6 COAL)
AFTER CHANGE
ILLINOIS NO. 6 COAL
Ca/S = 0.76
COAL SULFUR CHANGE
BEFORE CHANGE
CHAMPION COAL
Ca/S = 1.40
I
300
600
900
1200 1500
1800 2100 2400
SO2 EMISSIONS (PPM)
-------
FIGURE 111-32
IO
+22.
+15.0-
LU
O
z
I
u
o
u.
-15.0
-22.5
0
INSTANTANEOUS SO2 RESPONSE (RUN 99)
(ILLINOIS NO. 6 TO CHAMPION COAL)
AFTER CHANGE
CHAMPION COAL
8.2 MINUTES
Ca/S = 1.40
PRIOR TO CHANGE
ILLINOIS NO. 6 COAL
Ca/S = 0.76
600
900 1200 1500
S02 EMISSIONS (PPM)
1800
2100
2400
-------
FIGURE 111-33
RUN TOO SO2 EMISSIONS
(10 MINUTE AVERAGES)
600
i i i i | I i i i | i
PREDICTED EMISSION 750 PPM
i i
500-
I 400
Q_
to
z
§ 300
—
LLJ
200
100
ATTEMPTED
CHANGE
Ca/S=1.43
— PREDICTED
EMISSION
170 PPM
GAS SAMPLING
SYSTEM INOPERATIVE
Ca/S =1.43
PREDICTED EMISSION 170 PPM~
I
10 15
TIME INTO RUN (MRS)
20
25
30
-------
FIGURE 111-34
+22.5
+15.0-
o
z
<
u
o
Qi
-15.0
-22.5
RUN 100 INSTANTANEOUS SO2 RESPONSE
(CHAMPION COAL)
600
AFTER CHANGE
Ca/S = 0.38
CHANGE IN Ca/S
PRIOR TO CHANGE
Ca/S= 1.43
900 1200 1500
SO2 EMISSIONS (PPM)
1800 2100 2400
-------
FIGURE 111-35
+22.5
+15.0-
+7.5-
LLJ
0
z
u
o
exi
Ll_
UJ
5
-7.5-
-15.0-
RUN 100 INSTANTANEOUS SO2 RESPONSE
(CHAMPION COAL)
CHANGE IN Ca/S
PRIOR TO CHANGE
600
900 1200 1500
SO2 EMISSIONS (ppm)
1800 2100 2400
-------
was fairly rapid (~8 minutes). The response to a 73% change in dolomite feed
rate required 2 to 5 hours to reach a stable emission. This would indicate
that 1t may not be possible to prevent sudden changes in SO? emissions which
respond rapidly to variations in coal sulfur content with changes in sorbent
feed rate which require much longer response times.
The fact that most emissions were far from the predicted emission levels,
may suggest that there is an initial short response time coupled with a very
much longer response time. Chemical analyses of the sorbent, bed extract
material, and cyclone dumps have been sent to GE. They are incorporating these
data into their S02 emission response model and will issue a separate report
with their conclusions.
NQX Emissions
NOX emissions measured in all minlplant runs are plotted in Figure 111-36
as a function of percent excess air. Data follow the same trend Tine shown
in previous reports (1). NOX emissions are well below the recently modified
New Source Performance Standard of 0.6 Ib/MBTU for coal fired utility boilers.
Most PFBC design studies have indicated the excess air level will most likely
be around 20%. At this excess air level, NOX emissions are expected to be
0.2 4^0.1 Ib/MBTU. Emissions are expected to be below 0.4 Ib/MBTU, even at
excess air levels in the range of 60 to 100%.
An evaluation of methods which could decrease NOX emissions below those
in Figure 111-36 was carried out in the bench PFB combustion unit. The
results of this study are given in Section VII.
Other Gaseous Emissions
503 emissions in the combustor flue gas were measured for most runs. For
runs up to run 75, the EPA Method 8 method was used. For subsequent runs, the
controlled condensation method described in Section III was used. Results
for all measurements is given in Appendix Table G. Two measurements out of a
total of 40 appeared to be in error and were discarded. These were readings
of 213 pprn from run 71 (Method 8), and 73 from run 100 (controlled condensa-
tion). Excluding these two measurements, 16 Method 8 measurements averaged
12 + 12 ppm, 22 controlled condensation measurements averaged 6 +_ 9 ppm. All
meaFurements averaged 9 £ 11 ppm. In both cases, the range of measurements
was 0 to 30 ppm. This is the same range reported previously when only the
Method 8 technique was used (1). Aside from the fact that the controlled
condensation method gave an average reading exactly one half of the Method 8
average, no positive conclusions could be drawn from the results. Even the
difference between the average measurements is not statistically significant
due to the large degree of uncertainty. In both cases, the degree of uncer-
tainty was approximately +, 100% of the average values. Large variations were
also measured within a single run. Therefore, the variation appears to be
random and no conclusions could be drawn regarding the cause of SOs formation
or the factors affecting the degree of $03 formation.
CO emissions were again in the range of 50 to 200 ppm as reported pre-
viously (1). A few measurements were 1n the 300 to 500 ppm range, but could
have been caused by analyzer problems.
73
-------
FIGURE 111-36
CORRELATION OF NO EMISSIONS
-------
Reduced sulfur compounds, H2S, COS and C$2 1n the flue gas were measured
by gas chromatography (GC) a number of times and were consistently less than
the detectability limit of 1 ppm. Hydrocarbons were also measured by GC.
Methane averaged 7 +_ 5 ppmt ethane 4 + 4 ppm, 03 through Cc hydrocarbons were
generally below the detectabillty limit of 1 ppm. GC results are given in
Appendix Table F.
Emission of sodium, potassium, chlorine and vanadium in the flue gas was
also analyzed using the system described previously. In this system, a sample
of high temperature, pressure flue gas is extracted from the ducting following
the third stage cyclone, partly filtered, cooled, filtered again, cooled fur-
ther and bubbled through an impinger train (see Figure 111-37). Figure 111-37
also shows the different locations where the four elements are condensed,
deposited or absorbed and subsequently analyzed. The filters are extracted
with boiling water and the extract analyzed. Unused filters are also extracted
to give a blank for the filter material itself. The quartz tube is washed and
the wash solution analyzed. Table III-5 gives the distribution of the four
elements and particulates found in the six locations in the train. Sodium
totalled 2 to 3 wppm in the inlet flue gas, potassium 0.3 to 0.5 wppm.
Generally, over 90% of these elements were found on the front and final fil-
ters (sampling locations 1 and 3 in Figure 111-37). The front filter was not
retaining all the particulates impinging on it, as seen by the large fraction
(60 to 70%) of particulates captured by the final filter in samples 78 and
79-1. The front filter media was removed for sample 79-2. However, the
fraction of total alkali captured on the final filter was greater than the
fraction of particulates captured on the final filter. Therefore, sodium and
potassium compounds were condensing in the quartz tube on the surface of
particulates which were removed by the final filter. Very little alkali
compounds were condensing and collecting on the surface of the quartz tube
itself.
The concentration of chlorides measured in the flue gas was about 50
wppm. This represents about 60% of the total Cl entering with the coal and
dolomite feeds. As seen in Table III-5 chlorides were not detected on the
particulates but only in the knockout condensate. Therefore, chlorides were
probably all present as HC1 at the low temperature conditions in the sampling
probe. Sodium and potassium were then probably present as the sulfates in the
sampling system, although they probably occurred as chlorides in the high
temperature flue gas prior to sampling.
No vanadium was detected in any of the samples.
Particulate Emissions
Particulate emissions from the miniplant depend heavily upon the parti-
culate cleanup system used. The emission when the flue gas is cleaned by a
granular bed filter is very different from that when it is cleaned with cyc-
lones. During this reporting period almost all runs were made with 3 stages
of cyclones in series. For this reason, this section will deal mostly with
emissions measured in the flue gas cleaned in this manner. Emissions using
other control technology are contained in Section IV.
75
-------
FIGURE 111-37
SCHEMATIC OF THE ALKALI PROBE TRAIN
CT»
VENT
1 ATM
T
WET TEST
METER
2ND
IMPINGER
1
1
I
ER
NEEDLE
VALVE
1
1--,
1 1
3\ -r
T\ 1
1 1ST '
IMPINGER
(T)
>240°C
//
¥ n
* 1
, \
\ BALST
\ FILT
KNOCK O
CON DEN SA
©<~:
o
ANALYZED SAMPLE LOCATIONS
AIR COOLED
S.S. PROBE
843°C
ATM
FILTER QUARTZ
TUBE
-------
TABLE II1-5. EMISSION OF SODIUM, POTASSIUM, CHLORINE, VANADIUM IN FLUE GAS
Run/Sample No.
Element
Concentration (wppm)
Distribution (wt. %)
Location
1 - Front Filter
2 - Quartz Tube
3 - Final Filter
4 - Condensate
5 - 1st Impinger
6 - 2nd Impinger
78/4
Na K V
2.06 0.54 0
7
< 1
89
3
< 1
< 1
37
3
59
1
0
0
79/1
irt.
• ••
39
0
61
0
0
0
Na
1.84
1
2
96
1
0
0
K
0.28
5
14
80
1
0
0
V Cl
0 47.4
- 0
- 0
- 0
- 100
- 0
- 0
Part.
™ ™
31
0
69
0
0
0
Na
3.23
< 1
2
95
2
0
0
79/2
K
0.38
6
8
83
3
0
0
y___ci
0 53.8
- 0
- 0
- 0
- 100
- 0
- 0
Part
~ *
3
0
97
0
0
0
-------
Particulate emissions during this reporting period generally ranged from
0.03 to 0.15 g/Nm3 (0.013 to 0.065 gr/SCF) with 3 stages of cyclone cleanup.
These gas particulate concentrations were measured with Balston total filters
in either the Balston filter sampling system or the HTHP sampling system.
The size distribution of particulates in the filter cake was obtained with the
Coulter Counter utilizing techniques described 1n Appendix D. The size dis-
tributions of most samples are recorded 1n Appendix M. Table III-6 gives par-
ticle size distributions in the flue gas leaving the third cyclone for two
runs which span the expected range.
TABLE 1II-6. PARTICULATE EMISSION PARTICLE SIZE DISTRIBUTION
Particle Size (urn)
Run 5% 10% 25% 50% 75% 90$ 95%
No. Less Than Less Than Less Than Less Than Less Than Less Than Less Than
78
80
0.58
0.56
0.66
0.66
0.98
0.85
1.9
1.2
4.6
1.8
8.5
3.4
11.0
5.4
The particulate loading of the PFBC effluent at various locations within
the off gas system during a typical run (run 80) is shown in Figure 111-38.
Run 80 is typical of good long term operations. The concentrations are broken
down into five principal size ranges for material leaving each cyclone in the
gas stream. These loadings were obtained from mass balance calculations around
the cyclones. Size analysis was performed on a Balston filter catch, as well
as second and third cyclone dump material taken at a similar time as the Bal-
ston filter flue gas particulate samples were being collected. These size
analyses together with the mass collected by each cyclone and on the Balston
filter were used to calculate the loadings and size distributions leaving the
primary and secondary cyclones. The only assumption made was that the par-
ticles did not change size significantly during their pass through the cyclone.
Total particulate concentrations and median particulate sizes in the flue gas
at the same three points, and median particulate sizes in material captured in
the secondary and tertiary cyclones, are given in Table II1-7. These data
cover the range of values measured in a large number of individual samples from
many runs. A comparison of the data in Figure 111-38 and the range of values
in Table III-7 shows that run 80 was a typical run.
TABLE III-7. PARTICULATE CONCENTRATION AND SIZE
RANGES REPRESENTING A NUMBER OF RUNS
Cyclone
Recycle
(Primary)
Second
Third
Part. Cone,
(g/m3)
Passing
8-12
0.4-1.2
0.03-0.15
Mass Median Size (pm)
CapturedPassing
20-25
20-25
3-5
3-5
1-3
Cyclone Efficiency
(X)
90-95
85-94
78
-------
PARTICULATE LOADING (gr/SCF)
o — •
o o — ' o
o o • • •
— — — 0 0
n -o
-— ^^ — ~
r^ O >
vy ^
m -y ^
rn
to
ECONDAR
CYCLON
OUTLET
*
o Q ~*
5 P 3
p"* /"^ H^
™ z 3
rn "^
1 1 1 1 1 1 1 1 } 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 i r\ ^r\t im /*"*"* i*1 o' ^
1 IU £U|irn \£.i »Q /Q)
20-40 jim (0%)
>40|im (0%)
(0-5 um (92 0%}
— | 5-1 Opm (7.3%)
10-20|im (0%)
20-40 nm (0%)
>40pm (0%)
i t i I 1 1 1 1 i i t i i i i 1 1 i i i i i 1 1 1 1 i i t i i i 1 1 1 1
-o
70
n
m
O
O
y^
O
rn
z
"^
IS
zz
•
n
O
z
m
s
oo
0
c
— <
r™
m
—I
0 0 — -
* • • \J
o — o x
-n
O
c
70
m
T
&
oNiavoi
-------
The primary cyclone was Intentionally designed with a low efficiency.
Its primary function Is to return the larger particles back to the bed. Dur-
ing normal operations, the material in the cyclone dlpleg has the size dis-
tribution shown 1n Figure II1-39. At times this dlpleg has become plugged
and operations have continued as In run 81. During this type of operation,
there is a large Increase 1n the material captured 1n the second cyclone, a
slight Increase 1n the third cyclone capture and a large reduction in the bed
overflow solids rate. This change in the mass balance envelope reflects the
greater amount of material entrained from the combustor and recycle cyclone
and captured in the last two cyclones. At the exit of the third cyclone, the
only significant difference from normal operations when the primary cyclone
dlpleg plus is a 3-5 percentage point increase in particles over 10 ym. The
particulate concentration and average particle size are not changed signi-
ficantly. A further discussion of run 81 is contained in another report (6).
The secondary cyclone is in reality the first cleanup cyclone. As shown
in Table III-7 this cyclone captures 90 to 95% of the particulates in the flue
gas exiting the recycle cyclone. The average particle size of the particulates
captured by and passing through the cyclone is 20-25 urn and 3-5 um respectively.
The third cyclone in the train ~ the second cleanup cyclone -- captures
approximately 90% of the particulates which exit the second cyclone. The par-
ticulates exiting this cyclone with the flue gas have a mean diameter of 1 to
3 ym. Of the particulates leaving the cyclone with the flue gas, normally
80 to 90% are smaller than 5 ym.
During the course of particulate sampling, many samples of particulates
were found to have magnetic particles in them. These samples required special
techniques for Coulter Counter size analysis, these techniques are described
in Appendix D. The possibility of enhanced particulate collection by magnetic
interactions was investigated. Figure 111-40 is a magnetization curve for
particles captured by the third cyclone during runs 99 and 100. Figure 111-41
shows the effect of temperature on magnetic induction. Clearly this effect
1s not useful in the temperature range of interest (800-900°C). Furthermore
based on the type of response curves seen in the two figures, it is evident
that the magnetization is primarily due to the presence of ferrites in the
ash.
Combustion Efficiency
Carbon combustion efficiencies for runs made since the last report were
consistently above 99%. Only three runs out of a total of 39 were below 99%.
The correlation published in the previous report (1) which relates combus-
tion efficiency to the average bed temperature, excess air level and gas
residence time, was used to predict combustion efficiencies for these runs.
The calculated efficiencies were consistently lower than the measured effi-
ciencies. The average calculated efficiency was 98.5% compared to 99.3%
measured, an average miss of 0.8 combustion efficiency units. The standard
error for the correlation was previously reported to be 0.6 units but without
the bias shown 1n the recent runs. If a consistent measurement error occurred
80
-------
FIGURE 111-39
PRIMARY CYCLONE DIPLEG SIZE DISTRIBUTION (RUN 85)
oo
z
<
ffi
Z
u_
2000 1000 600 400
200
100 60 40 20
PARTICLE SIZE (jim)
1086
-------
CO
ro
FIGURE 111-40
MAGNETIZATION CURVE FOR THE MINIPLANT FLYASH SAMPLE AT 25°C
100
I
I
oo
LLJ
to
I/O
o
*.
5
•*
z
B 50
u
Q
z
u
I—
UJ
Z
o
0*
50 100
MAGNETIC FIELD, H, OERSTEDS
T50
-------
FIGURE 111-41
EFFECT OF TEMPERATURE ON THE MAGNETIC INDUCTION
FORCE OF THE MINIPLANT FLYA5H SAMPLES
0.03
200
300 400
TEMP. °C —
500
600
700
83
-------
in the recent runs, the error was apparently less than 1%. Therefore, a
conservative estimate of the average combustion efficiency from recent runs,
obtained from the correlation, would be 98.5%. This is within the targeted
range for a commercial PFBC facility.
85
-------
SECTION IV
PARTICULATE MEASUREMENT AND CONTROL
An important technical issue to be resolved before pressurized fluidized
bed combustion can be applied commercially is the degree of particulate
removal needed to protect the gas turbine. A related issue is which tech-
nology to use to achieve the needed degree of the particulate control. In
addition to meeting the particulate removal requirements set by the gas tur-
bine, the environmental New Source Performance Standards published by the
Environmental Protection Agency must also be met.
The degree of particulate removal needed to protect the gas turbine
from erosion, hot corrosion and deposition has not been defined as yet.
Studies, including the 1000 hour turbine cascade exposure test recently com-
pleted in the mi nip!ant for the Department of Energy, are now underway to
define the needed degree of particulate control. Although the EPA is not
directly concerned with studying particulate control to protect gas turbines,
it is responsible for the evaluation of such particulate control devices to
the extent that these devices could determine the level, size and composition
of particulate matter emitted from a PFBC system. Therefore a particulate
control program was begun in which pre-turbine devices such as granular bed
filters, a ceramic fiber filter and high efficiency cyclones were evaluated.
In addition, post-turbine devices were also tested, based on the realization
that the degree of pre-turbine cleanup may not be sufficient to meet the more
stringent New Source Performance Standards for particulates. In this case, a
trailer mounted electrostatic precipitator (ESP) and a trailer mounted bag
house were connected to the miniplant flue gas system and tested under typical
low pressure, low temperature conditions. A number of these tests were con-
ducted in cooperation with other EPA contractors. A ceramic fiber filter,
built and tested by Acurex Company, was provided by Acurex and tested jointly
with them. The mobile ESP and bag house were also tested in cooperation with
Acurex. These devices are operated by Acurex, for the EPA, on a number of
industrial gas sources.
The cyclone tests were carried out primarily by Exxon. However, Southern
Research Institute and Air Pollution Technology, both EPA contractors, con-
ducted a series of particulate sampling tests to confirm cyclone performance
results obtained by Exxon using a different particulate sampling system.
In addition to evaluating the particulate control devices, it was also
necessary to develop and Improve particulate measurement systems. Such sys-
tems were used to determine particulate concentrations in the flue gas before
and after the control devices and to measure particulate size distributions.
The results of the particulate measurement and control programs are
reported in this section.
86
-------
PARTICULATE MEASUREMENT
Particulate measurements In the high pressure, high temperature flue gas
are difficult to make. A lot of data have been obtained with 18 cm Balston
high temperature total filters. These filters were used on all three sampling
systems described earlier. The particles captured on the filters were sized
with a Coulter Counter Model TA11 as described in Appendix D. However, there
is some question as to whether the particulate size distribution measured by
the Coulter Counter is the same as that actually occurring in the flue gas
before sampling. For this reason sampling was attempted with the Southern
Research Institute (SoRI) 5-cyclone train and the University of Washington
(UoW) 7-stage impactor using the HTHP sampling system for an 1n-s1tu size
distribution measurement. Verification of results was also made by other EPA
contractors (Southern Research Institute and Air Pollution Technology) with
Impactors designed for the high temperature, high pressure conditions. In
order to understand better the instantaneous particulate emission, an on-line
particle concentration monitor was used. This device was developed by IKOR,
Inc. of Gloucester, Massachusetts. It is based on a charge transfer principle.
Balston Filter Measurements
Balston filter samples have been used to build a large data base of cyc-
lone, granular bed filter and ceramic barrier filter performance. This total
filter gives a good measurement of the particulate concentration provided good
sampling procedure is followed. Obtaining the size distribution of this
material can be a problem. The Coulter Counter method used to size the parti-
culates requires that a small amount of solids be dispersed in an agitated
electrolyte (Isoton II). The particles are then drawn through a precision
orifice. A small current continuously passing through the orifice is Inter-
rupted each time a particle passes through. The degree of Interruption Is
proportional to the particle volume. An accumulator records the number of
these interruption pulses whose magntiude falls 1n each of the 16 accumulator
channels. The particle diameter reported Is for a sphere of the same volume
as the particle.
One question that arises is whether the particulates are agglomerated in
the flue gas or on the filter and to what degree they are redlspersed in the
agitated electrolyte. Depending on the relative extent of agglomeration and
redispersion, the measured particle size distribution could be either coarser
or finer than that occurring 1n the flue gas. Also sufficient particulate
must be captured to form a filter cake. If there 1s no cake formed, filter
material may contaminate the sample. Filter material consists of fine, long
cylindrical particles with a volume mean diameter of about 9 ym.
A cyclone or other cleanup device sees the aerodynamic equivalent dia-
meter. This 1s different from the volumetric diameter measured by the Coulter
Counter. In order to determine the validity of Coulter Counter measurements
for approximating aerodynamic size, some material was sent to General Electric
for size analysis with a Bahco particle size analyzer. A comparison between
the Coulter Counter and the Bahco results for both second and third stage cyc-
lone capture material is shown 1n Figures IV-1 and IV-2. The distributions
87
-------
FIGURE IV-1
oo
00
LLI
N
Q
LU
<
O
I—
z
_i
O
COMPARISON OF PARTICLE SIZE DISTRIBUTION VIA BAHCO VS. COULTER COUNTER
OF SECONDARY CYCLONE CAPTURED MATERIAL (RUN #78)
EXXON (COULTER COUNTER AND SONIC
SIFTER)
-------
00
10
FIGURE IV-2
COMPARISON OF PARTICLE SIZE DISTRIBUTION VIA BAHCO VS. COULTER COUNTER
OF TERTIARY CYCLONE CAPTURED MATERIAL (RUN #78)
N
co
O
LLJ
co
Z
<
I
I—
oe.
UJ
s
Qi
O
O
100
90
80
70
60
50
40
30
20
10
0
n
GE-BAHCO
EXXON (COULTER COUNTER)
Q
\
i i i
.81.0
2.0 3.0 4| 5 | 6 7 8 910
4.2 5.4
PARTICLE SIZE (jim)
20
30 40
-------
are remarkably similar for two such different techniques. The second cyclone
material was preclassified with a sonic sifter for the size distribution above
45 ym and combined with the Coulter Counter size distribution to give the
Exxon result. Balston filter captured particulates cannot be analyzed with
the Bahco, because there is an insufficient quantity of material available for
this instrument.
High Temperature High Pressure
Particulate Sampling
The question as to the difference between the size distribution measured
by the Coulter Counter and the true size of particulates in the flue gas before
sampling was also addressed. It was answered as part of a joint program with
Southern Research Institute (SoRI) and Air Pollution Technology, Inc. (APT),
both EPA contractors. The joint program was intended to resolve the question
of the validity of the filter/Coulter Counter method and thereby confirm cyc-
lone efficiencies based on this method.
The minimum particle size detectable by the Coulter Counter in this
application was estimated to be approximately 0.6 \m. Thus, the possibility
existed that the fraction efficiency results might have been biased by using
the mass balance technique with Incomplete size distributions. Further, the
ultrasonic deagglomeration of particles prior to performing the Coulter anal-
yses may have resulted in aerodynamically large agglomerates being measured
by the Coulter Counter as individual primary particles. The magnitude of the
possible errors in the fractional efficiency curves resulting from these
biases are difficult to assess.
Cascade impactors have been used for a number of years for determinations
of control device fractional efficiencies over the size range from approx-
imately 0.3 ym to 10 ym. Recent work at APT (9), has shown that cascade
impactor performance at temperatures and pressures like those in the miniplant
flue gas lines can be predicted to good accuracy by current theories. A
pair of cascade impactors which were designed by APT, to operate at high pres-
sures and temperatures was made available by the EPA for a series of indepen-
dent tests of the miniplant tertiary cyclone outlet size distribution and
fractional collection efficiency.
Inconel shim substrates were used for particle collection surfaces for
each stage. Ceramic fiber backup filters were used to collect those particles
which were not removed by the impaction stages. Qualitative verification of
the performance of the Impactors was obtained by Coulter Counter analyses
(where applicable) and by electron microscopy of the various stage catches
from typical runs. Previous experience by Exxon and the samples obtained dur-
ing this joint program by Southern Research and APT showed that the particles,
at the sampling conditions, were highly adhesive. This permitted valid
impactor results to be obtained even though bare metal substrates were used.
(Ceramic fiber substrates were on hand for use had this not been the case.)
90
-------
Figure IV-3 shows comparative results of size distributions of the par-
ticles In the tertiary cyclone exhaust stream obtained by Coulter analyses of
filter catches from mlnlplant run 105 by Exxon and those obtained by means of
cascade impactors during the same run by SoRI and APT.
As seen, the cascade impactor results Indicate a larger concentration of
fine particles and a mass median particle size of about 0.8 micron, where the
Coulter Counter mass median particle size averaged 1.6 microns.
Electron micrographs of material captured on the various impactor stages
Indicated that, with the exception of the first stages, the participate matter
collected in the Impactor stages was fine, non-agglomerated participates.
Electron microscopy also revealed that much of the aerodynamically large*par-
ticulate matter on the first impactor stage at the cyclone Inlet was agglo-
merates of smaller particles.
These findings indicated that the results of the Balston filter/Coulter
Counter method used by Exxon differed somewhat from cascade impactor results
but to the degree expected, based on measurements made in other particulate
systems. They also indicated that the Coulter Counter results were not being
biased toward the finer particles by breakdown of agglomerates in the aqueous
dispersing medium used in the Coulter Counter, since the Coulter Counter gave
a coarser distribution than the impactor.
Prior to the joint impactor sampling program with SoRI and APT, samples
were taken by Exxon with a University of Washington (UoW) Mark III cascade
Impactor. The 7-stage impactor was used to obtain background experience with
impactors at high temperature and high pressure. The impactor was used with
the HTHP particulate sampling system. It was heated to 750°C and pressurized
to 900 kPa. Several runs were made using the Impactor which was designed for
the dilute sampling conditions. These runs showed that the particulate was
very adhesive and would stick to bare metal substrates. Also 1t was shown
that the pre-cut cyclone used ahead of the Impactor to prevent overloading
of the first stages was not needed. It captured very little material and
after most runs the first two stages of the impactor were empty.
Particle size distributions obtained with the UoW Impactor were Indeter-
minate. The fmpactor was erroneously run at very high flow rates. This tended
to increase jet velocities to the point where much of the material that should
have remained on a given stage may have been "washed" off onto a following
stage. This may be the reason that the Balston backup filter captured a sub-
stantial portion of the particulate. The stage cut diameters were calculated
with classical formulas, and the size distribution obtained. The size dis-
tribution was much finer than that of a Balston filter (with Coulter Counter
size analysis) taken at the same time. This 1s further evidence that the flow
rates were much too high.
Mini Cyclone Train Sampling--
The SoRI 5-cyclone train was used to try to obtain gram size quantities
of particulates which pass through the tertiary cyclone. This sampling device
was also intended to give an in-situ distribution of the particulates once it
91
-------
FIGURE IV -3
LU
N
oo
LLJ
U
BALSTON FILTER/COULTER COUNTER VS. APT IMPACTOR
PARTICLE SIZE DISTRIBUTION
FLUE GAS FOLLOWING THIRD CYCLONE (RUN 105)
10.0
I*
6.0-
5.0-
4.0
3.0
2.0
l.(
0.3
0.2
0.1
2%
CASCADE IMPACTOR
BALSTON FILTER/
COULTER COUNTER
1
1
1
1
1
i
i
i
I
I
5 10 20 30 40 50 60 70 80 90 95
PERCENTAGE SMALLER THAN PARTICLE SIZE
98%
92
-------
was calibrated. The cyclones were used with the HTHP sampling system. They
were heated to between 720 and 870°C and pressurized to 900 kPa. The problems
associated with the 5-cyclone train were mostly due to the very dilute nature
and small size of the particulates in the flue gas. Long sample times were
required and small amounts of particles had to be removed from the relatively
large catch cups. The system was very prone to leaks and contamination by
oxidized metal and the anti-seizing compound used on the nuts and bolts that
hold it together. The total loading of particulates measured was typically
one-half the loadings obtained with concurrent Balston filter sampling. Also,
the size differentiation of the cyclone train was much too broad at the samp-'
ling conditions to describe the particle size distribution. The cyclone cap-
ture efficiencies probably due to re-entrainment of the dust from the catch-
pots on each cyclone. Therefore, with the cyclone stage cut diameters a func-
tion of sampling duration, calibration was not possible. For these reasons
the SoRI 5-cyclone train was not deemed suitable for use in the downstream
location to measure particulate loadings or size distributions. Modifications
may have improved its performance under these conditions, but this was out of
the scope of this program.
Continuous Particulate Monitor
A continuous, on-line particulate monitoring system was evaluated during
run 80. The monitoring system (manufactured by IKOR, Inc., Gloucester,
Massachusetts) makes use of the surface charge that particulates accumulate
during their flow through the combustion unit.
The IKOR probe measures the total electrical charge of impacting solids.
The probe consisted of a 0.63 cm diameter solid rod of type 316 stainless
steel that was 35.5 cm long. The last 8.9 cm of the probe was immersed in
the 10 cm diameter exhaust duct of the miniplant at a location 2.1 meters
downstream of the GE turbine test cascade. The probe was mounted on the
duct's center line and normal to the gas flow direction. A constant purge of
dry N£ was maintained around the probe to prevent condensation on the probe
insulator.
The output of the probe was correlated with gas temperature measured near
the probe location and with average particulate loadings measured by Balston
filter samples. The data plotted in Figure IV-4 compares the IKOR signal
strength with gas temperature during run 80 at relatively constant dust load-
ing. The particulate loading data obtained using the Balston filter during
run 80 are summarized in Appendix M-6. The particulate data show that the
average mass mean diameter of the particles captured by the Balston technique
following the third cyclone was 1.2 ym and the average loading was 0.051 g/Nm^
The loading varied between 0.034 and 0.067 g/Nm3 during the IKOR testing
period. The samples contained 6% of particles larger than 5 ym.
The data in Figure IV-4 show that when the particulate loading downstream
of the third cyclone averaged 0.051 g/Nm3, the IKOR probe signal strength
varied almost linearly with gas temperature. Natural gas was injected into
the combustor exhaust stream upstream of the third cyclone. The gas tempera-
ture at the probe location was varied by adjusting the amount of natural gas
flow.
93
-------
FIGURE IV-4
EFFECT OF TEMPERATURE ON IKOR MONITOR READING (RUN 80)
0
^—
II
>
SENSITI'
©
LU
w
— J
D
u_
tfS
IUU
90
80
70
60
50
*J\J
40
30
20
10
0
65
O -, o-
Z Z °co
0^1-7
< ^ co \
O t °. d
— 1 ^^ ^^
i j
• v\
" XCONSTANT
"•/J LOADING
* IBUT TEMP.
•"/• JCHANGEVI;
f ICH4 INJEC-
./ [TION
/ !
• J '
• / !
COMBUSTOR ON LIQUID FUEL X1 \
1 \
• i i * i i !
0 700 750 800 850 S
-
—
-
<\
-
-
>0(
GAS TEMPERATURE @ PROBE (°C)
-------
The unexpected sensitivity of the IKOR signal to probe temperature made
comparison of signal amplitude with gravimetric sampling very difficult.
Because of this and the unknown effect of thermionic emission from the probe
or pressure vessel walls, further testing was stopped. IKOR agreed to investi-
gate the problems and possibly develop a temperature compensator for the unit.
CYCLONE STUDIES
The miniplant was initially designed with only two stages of cyclone
cleanup. The gas was sampled and expanded after the secondary cyclone. The
particulate loading of the gas was much too high to pass through any gas tur-
bine. A granular bed filter was then added to the flue gas stream. The pro-
blems of this device are contained in this and a previous report (1). A con-
ventional cyclone was installed upstream of the granular filter to lower the
dust loading and possibly improve efficiency. This cyclone performed very
well and lead to a halt in granular bed filter testing. All subsequent hot
gas cleanup was accomplished with three stages of cyclones. However, further
tests of hot gas cleanup systems were conducted on a slipstream of flue gas
withdrawn after the second stage cyclone. Low temperature systems were also
tested. These tests are discussed in subsequent sections.
Cyclone Efficiency Testing,
Particulates in the flue gas leaving the miniplant recycle cyclone pass
through the two additional cyclones which remove ~99% of the particulate
matter. Mass balance calculations based on particulate material captured in
the second and third cyclones and in the flue gas leaving the third cyclone
are used to determine the particulate size and concentration at the exit of
all three miniplant cyclones.
These measurements and calculations show that material captured by the
second cyclone has a mass median particle size of 20 to 25 urn. The overall
cyclone efficiency is about 95%. Material captured by the third cyclone has
a mass median size of 3 to 5 ym. The overall efficiency of this cyclone is
about 90%. Particulate concentration in the flue gas leaving the third cyc-
lone is generally 0.03 to 0.15 g/Nm3. The mass median size of the particulate
ranges from 1 to 3 microns as determined by Coulter Counter. These results
are summarized in Table IV-1 . Detailed particulate loading and size data
for each run are shown in Appendices M-3 through M-6.
TABLE IV-1. PARTICULATE CONCENTRATION AND
SIZE MINIPLANT CYCLONE SYSTEM
Part. Cone. Mass Median Size
(g/Nm3) (Microns) Cyclone
Cyclone Passing Captured Passing Efficiency (%)
Recycle 8-12 -- 20-25
Second 0.4-1.2 20-25 3-5 90-95
Third 0.03-0.15 3-5 1-3 85-94
95
-------
The particulate emissions after the tertiary cyclone have been as low as
0.03 g/Nm3. This is slightly below the New Source Performance Standard (0.03
Ibs/MBTU) which corresponds to -0.035 g/Nra^. The performance was not con-
sistent however, and usually exceeded the emission standard. This performance
is not good enough to use cyclones alone to meet the emission standards.
The third cyclone fractional efficiency is of greatest interest, since
if a three stage cyclone system is used in a commercial PFBC system, the third
stage must be very efficient to prevent damage to the gas turbine behind it.
The collection efficiency of the tertiary cyclone has been measured over some
20 runs with run durations between 8 and 250 hours. Throughout this period
the total collection efficiency averaged 90% with a standard deviation of
only +_ 3%. This summary is shown in Table IV-2. As can be seen from the
table, these measurements were made with 3 different coals and 2 different
sorbents with a variety of Ca/S ratios.
Both the total efficiency and the fractional (grade) collection efficiency
was much higher than expected. This cyclone, which was designed with classical
handbook formulas, averaged a cut diameter (50% efficiency), over this period,
of 0.88 vim ^0.02. Table IV-3 shows the grade efficiencies measured during
runs 68 through 100. These efficiencies were obtained from outlet samples
and captured samples sized with a sonic sifter and a Coulter Counter. The
inlet concentrations were calculated by size differentiated mass balance under
the assumption that no size change of particles occurred through the cyclone.
During runs 99 and 100 this assumption was checked with filter samples taken
upstream of the cyclone. Although cyclone performance may not be sufficient
to meet emission standards, cyclones performing as well as those used on the
miniplant may be sufficient to protect the gas turbine from serious damage by
erosion.
The main purpose of the particle size distribution study (described ear-
lier) by the Southern Research Institute and A1r Pollution Technology was to
measure cyclone fractional collection efficiency. Figure IV-5 shows the frac-
tional efficiency of the third cyclone during run 105 as calculated from; (1)
the SoRI/APT cascade impactor data, (2) the Exxon total filter/mass balance
technique with Coulter Counter size analysis, (3) the Lelth and l_1cht (10)
cyclone fractional efficiency model. The cyclone operating conditions are
shown 1n Table IV-4. The Impactor and the total filter/Coulter Counter effi-
ciencies agree fairly closely except in the small particulate size range.
The cyclone cut diameter (50% efficiency) 1n both cases was approximately 0.7
microns. Therefore, the cyclone efficiencies calculated from cascade Impactor
data substantially confirms the efficiency obtained from total filter/Coulter
Counter data. However, the predicted fractional efficiency curve is signifi-
cantly lower than the measured results. The Lelth and Licht model, shown in
Figure IV-5, comes closer to predicting miniplant third cyclone performance
than other available models tested. The reason for the lack of agreement
between the measured and predicted results is not understood at the present
time. It may be due to the fact that many of the models are semi-empirical
and based on data obtained with particulate generated by other systems at lower
temperatures and pressures.
96
-------
TABLE IV-2. TERTIARY CYCLONE TOTAL EFFICIENCY SUMMARY
Run #
Total Eff.
67.1
67.2
67.3
68
71
72
73
74.1
74.2
75
78.2
78.4
78.10
79
80
81
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
Average Of
All Runs
+ Is
88.4
83.4
88.7
97.0
82.9
91.3
94.3
94.6
94.5
89.1
89.2
90.0
90.1
93.9
84.6
91.2
92.4
95.9
94.7
92.0
94.2
92.6
88.6
90.9
91.1
88.6
89.8
89.9
90.4
91.1
90.7
86.9
89.6
88.4
90.6 +3.3
Coal Type
Illinois
Illinois
Illinois
Illinois
Illinois
Ohio
Ohio
Ohio
Ohio
Ohio
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois/Ohio
Ohio/Champion
Champion
Champion
Champion
Champion
Champion
Champion
Champion
Champion
Champion
Champion
Champion
Sorbent Type Ca/S Ratio
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Limestone
Limestone
Limestone
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomi te
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
1.25
1.25
1.25
2.00
1.52
1.79
14
60
80
65
45
1.45
1.45
1.45
1.41
1.56
1.45
1.45
1.45
1.45
1.45
1.45
1.40/1.87
1.87
1.50
1.50
1.50
1.50
1.50
1.50
1.40
1.40
1.40
0.38
97
-------
TABLE IV-3. TERTIARY CYCLONE FRACTIONAL EFFICIENCY SUMMARY
Particle Size Captured with Stated Efficiency (urn)
Run No. 50% 75% 85% 90% 95%
68 N/M N/M 1.15 1.60 2.15
71 1.10 1.80 2.30 2.60 3.20
72.1 N/M 0.95 1.55 2.10 3.15
73.1 0.70 1.00 1.15 1.20 1.50
74.1 N/M 0.85 1.15 1.55 2.10
74.2 N/M N/M 0.80 1.55 2.50
75 0.80 1.45 1.85 2.10 2.60
78.2 0.85 1.20 1.50 1.70 2.50
78.4 1.05 1.40 1.70 1.85 2.45
78.10 0.75 1.20 1.55 2.00 N/M
79 1.15 1.45 1.65 1.80 2.00
80 1.05 1.65 1.95 2.30 4.50
81 N/M N/M 1.35 1.60 2.20
86 N/M N/M 1.55 2.20 3.20
87 N/M 1.85 2.20 2.35 2.70
96 N/M N/M 1.45 2.00 3.00
99.7 N/M N/M 0.90 2.00 4.50
99.7* N/M 1.00 2.50 4.00 6.35
100.3 0.70 1.25 1.60 2.00 2.30
100.3* 0.65 1.25 1.60 1.90 2.35
Is 0-88 ^ 0.02 1.31 + 0.3 1.57 +. 0.4 2.02 + 0.6 2.91 +. 1.13
* Efficiency calculated based on upstream Balston total filter sample.
N/M = Beyond range of Coulter Counter.
98
-------
100
90
80
70
60
50
40
FIGURE IV-5
TERTIARY CYCLONE COLLECTION EFFICIENCY
(RUN 105)
u
LLJ
z
o
O 30
U
20
10-
0
15
BALSTON FILTER/COULTER COUNTER
CASCADE IMPACTOR
LEITH AND LIGHT MODEL (THEORETICAL)
J I I I I
_l_J I I I I
10 9 8 7 6
.9 .8 .7 .6 .5 .4
PARTICLE SIZE, jim
-------
TABLE IV-4. THIRD CYCLONE OPERATING CONDITIONS (RUN 105)
Pressure 700 kPa
Temperature 635°C
Flow rate 14.6 Nnr/min
Inlet Velocity 36 m/sec
Pressure drop 4 kPa
Knowlton and Bachovln (11) found that pressures up to 5.6 M Pa had little
effect on cyclone performance. Their work was based on much higher dust load-
ings than the current study. Perhaps other effects, peculiar to the PFBC sys-
tem such as the nature of the particulates, or the operating temperature and
pressure are responsible for the high efficiency. Additional work will be
needed to explain the observed results.
The conclusions reached by this study were basically to confirm the high
third cyclone efficiencies measured with the Balston filter/Coulter Counter
technique. The Balston filter/Coulter Counter technique was not being biased
toward finer size distributions by particle deagglomeration in the aqueous
dispersing medium used in Coulter Counter measurements. In fact, the Coulter
Counter gave a coarser size distribution than the impactor, which is as expected
based on other measurements made in other particulate systems.
Further work is required to optimize cyclone performance under PFB con-
ditions. The emissions of particulate with 3 stage cyclone cleanup have, at
times, met the new source performance standard of 12.9 ng/J (0.03 Ibs/MBTU),
set by the EPA. However, the performance was not consistent enough to meet
the standard over the long term. The improvement in performance required is
small and may be met by cyclone optimization.
Cyclone Optimization and Variables Study
A cyclone variables and optimization study was undertaken during the last
7 mini plant runs. During these runs (nos. 109 through 115) three different
cyclones were exposed to various PFB operating conditions in the tertiary
position. Each cyclone was located inside the vessel that once contained the
granular bed filter elements. In this way, the plunger type valves that were
used to provide blowback shutoff of the granular filter elements could be
used to shut off the cyclones. All three cyclones could be tested, one at a
time, during every run. Each cyclone was equipped with an orifice in the inlet
line. In this way, valve leakage could be detected and flow to the cyclones
verified. A AP cell was used to measure the pressure drop between the pressure
vessel inlet and outlet. This was the same location used in previous measure-
ments with only one cyclone.
The dimensions of the cyclones are shown in Table IV-5 and Figure IV-6.
The "A" cyclone was the original miniplant tertiary cyclone which had almost
2000 hours of combustor operating time at the start of the test program. The
"B" and "C" cyclones were newly constructed, scaled down versions of cyclones
currently being tested by American Electric Power and Stal-Laval in England
at the National Coal Board facility in Leatherhead.
100
-------
TABLE IV-5. TERTIARY CYCLONE DESIGN DIMENSIONS
(cm)
ymbol
--
%
W
hi
ho
"B
he
Do
0/1
Dd
Dp
hp
•essur
Description
Inlet Type
Barrel Diameter
Inlet Width
Inlet Height
Outlet Pipe Insertion
Barrel Height
Cone Height
Outlet Pipe Diameter
Outlet/Inlet Ratio
Diameter of Cone at
Dust Pot.
Dust Pot Diameter
Dust Pot Height
e Drop @ 30.5 m/sec
Cyclone
A
Tangential
15.2
3.81
7.62
11.4
40.6
20.3
6.2
1.04
None
None
None
6.6
Cyclone
B
Scroll
17.8
3.66
8.18
7.62
22.9
44.5
5.25
0.76
1.94
12.4
25.4
11.0
Cyclone
C
Scroll
17.8
3.66
8.18
7.62
22.9
44.5
3.51
0.31
1.94
12.4
25.4
28.6
Inlet Velocity 8409C
930 kPa (calculated)
101
-------
FIGURE IV-6
BASIC CYCLONE DESIGN
102
-------
The parameters that were changed in this study included; cyclone geometry,
inlet velocity, inlet temperature, inlet particulate loading and combustor
coal feed. Cyclone geometry was studied by using more than one cyclone at
each condition. Inlet velocities were lowered during one test by lowering
the combustor fluidization air by 26%. The lower combustor superficial
velocity also lowered the cyclone inlet temperature. Cyclone inlet tempera-
ture was also varied by not injecting, or injecting very little natural gas,
This in turn varied the inlet velocity. Inlet particulate loading was varied
by deactivating the secondary cyclone for one series of tests. The effect
of combustor coal feed on cyclone performance was determined by varying the
combustor feed coal from the Champion coal used for most of the test to
Illinois No. 6 coal. During this test the dolomite sorbent feed rate was
approximately constant.
The cyclone performance tests were carried out in one day (6-10 hour)
runs. The combustor was allowed to equilibrate for 1-1/2 to 2 hours before
any testing was done. After steady state was attained, Balston total filter
samples were taken simultaneously at the cyclone inlet and outlet. At the
conclusion of the sampling a different cyclone was put on line. The next
cyclone reached steady state during the turn around of the filter (one-half
hour) and the sampling was started all over again. The cyclone particulate
catch was also sampled before and after every gas particulate sample.
The total collection efficiencies are shown in Table IV-6. For the
baseline case the total efficiency of the "A" cyclone was 84 +_ 2%. The B and
C cyclones averaged 83 +_ 4% and 72 + 6% total collection efficiency respec-
tively under baseline conditions. The standard deviations are such that
most of the non-baseline data are within the data scatter of the baseline
data. This obscured many of the effects intended to be studied.
The cyclone grade efficiencies were also calculated for each of the
tests. These were obtained from the inlet and outlet filter sample concen-
tration and size distributions. These grade efficiencies contained as much
data scatter as the total efficiencies. For the baseline case, the cyclone
cut diameters averaged 1.5, 1.1, and 1.7 ym for the A, B and C cyclones
respectively.
Further data concerning particle size distribution in the flue gas
prior to, and after, the cyclones are presented in Appendices M-5 and M-6.
The pressure drop measured across all cyclone systems was fairly high
(60-90 kPa). However, if the pressure drop due to the orifices and the rest
of the flow distribution system are subtracted, the pressure drops of the
cyclones were close to the calculated values. The pressure drop obtained for
the "A", "B" and "C" cyclones was 6, 12 and 40 kPa respectively. These com-
pare favorably with calculated pressure drops of 6, 12 and 29 kPa.
The conclusion reached from this study was that no significant effects of
pressure, inlet velocity, temperature, and combustor coal feed were found. An
increase in Inlet particulate loading did increase overall efficiency signi-
ficantly. The slight decrease in the cut diameter it caused was within the
data scatter of the measurements. None of the cyclones worked as well as
expected.
103
-------
TABLE IV-6. SUMMARY OF CYCLONE TEST PROGRAM RESULTS
Particle Size
Collected with
Run
No.
109
m
113
no
112
113
114
105
109
111
113
no
no
112
114
109
m
115
112
Cyclone
AM \
A ri
i 1 i
A
A
A
A
A/o ^
A(3)
B(D
B/f {
gU )
B
B
B
B
C(D
C\ f
/I \
C(D
C
Inlet
Velocity
(m/sec)
42
42
42
39
25
42
42
36
41
41
41
34
38
25
41
41
41
41
25
Inlet
Temp.
CC)
840
840
840
760
700
840
840
635
840
840
840
650
760
700
840
840
840
840
700
Coar2^
Type
CH
CH
CH
CH
CH
IL
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
Inlet
Loading
(q/m3)
1.87
1.61
0.85
1.35
1.54
1.73
4.83
1.62
1.23
1.13
1.79
1.36
1.24
1.25
5.06
1.19
0.91
0.54
1.14
Inlet
Loading Total Collecti
(g/m3) Efficiency (J
0.268
0.143
0.163
0.160
0.152
0.269
0.304
0.176
0.255
0.179
0.249
0.240
0.138
0.302
0.393
0.387
0.191
0.171
0.236
85
85
81
88
90
85
94
89
79
84
86
82
89
76
92
68
79
68
79
Stated Efficiency
on (u)
;) 50%
NA
NA
1.5
1.0
1.2
1.4
0.7
1.2
1.0
1.2
1.0
1.4
1.2
1.3
1.1
2.1
1.3
1.7
1.2
75%
2.0
NA
1.8
1.5
1.4
1.7
1.3
1.6
1.6
1.4
1.4
1.8
1.4
1.6
1.4
2.8
1.8
2.1
1.3
902
4.0
NA
2.1
2.8
1.7
2.2
1.9
2.1
2.9
2.0
1.8
2.4
1.8
2.2
2.0
3.8
2.7
3.4
1.8
(1) Baseline expected commercial/design operating conditions
(2) CH = Champion Coal (Ca/S = 1.25)
IL = Illinois No. 6 Coal (Ca/S = 0.76)
(3) Pressure Reduced to 700 kPa, All others at 900 kPa
NA - Data Not Available
-------
The "A" cyclone did not perform at the 907» collection efficiency measured
over the previous 2000 hours. This was probably due to changes in geometry
due to erosion, repair and thermal stress. The "B" and "C" cyclones also
did not collect particulates with the expected efficiency. Ironically, this
may be due to a lack of exposure. Any rough welds or edges had not had suf-
ficient exposure to the erosive gas to smooth them to their steady operating
cond-itions. As can be seen in Table IV-2 the total collection efficiency of
the "A" cyclone was also lower and slightly more eratic during Its first 100
hours of exposure (run 66-68} than later on.
Comparisons of "B" and "C" cyclone performance with the Van Tongeren
Model 850 cyclones used by AEP and Stal-Laval in England are underway. The
two cyclones used 1n England are geometrically similar to the "B" and "C"
cyclones. The results from these tests will be followed to develop more
reliable scale up procedures for cyclones in PFBC service.
CERAMIC FIBER FILTER EVALUATION
As mentioned previously, cyclones do not appear to be capable of consis-
tently meeting the New Source Performance Standard for particulates. However.
cyclones may be adequate to protect gas turbines, based on early results from
extended materials tests. Two devices, with the potential for meeting the
environmental emission standard as well as the turbine requirement, were
tested on a small flue gas slipstream from the miniplant. The first of these
was a ceramic fiber filter developed and supplied by Acurex Corporation under
an EPA contract.
Previous development work on the ceramic fiber filter concept is
reported elsewhere (28). The testing of the fiber filter on the miniplant
will also be described in a separate report by Acurex.
Test Description
The test device used in this study was provided by Acurex Corporation.
It consisted of a single filter element contained in a heated pressure vessel
with all necessary cycle controls. The filter consisted of a loosely packed
mat of Saffil alumina generally about 1 cm thick sandwiched between an open
weave fine gage 304 stainless steel support screen. The alumina mat, support
screen sandwich was formed into a cylindrical filter element which was slipped
over a heavy gage support cage and clamped at the top and bottom.
The objective of this study was to demonstrate the feasibility of ceramic
filtration under actual PFBC conditions. Data were also obtained on the effects
of cleaning cycle, face velocity, temperature and bag age on filtration
efficiency.
A cross section of the filter housing is shown in Figure IV-7. Hot,
dusty inlet gas enters the unit from the side, below the test filter. This
gas impacts against a plate on the dust hopper. Heavy particles may remain
in the hopper while others travel upwards to the filter element. The filter
element was 10 cm in diameter by 45.7 cm long. A heater element surrounded
105
-------
FIGURE IV-7
FILTER HOUSING PRESSURE VESSEL CROSS SECTION
r\ GAS
V OUTLET
REVERSE FLOW
INLET
GAS INLET [)
TEST FILTER
HEATER ELEMENT
DUST HOPPER
106
-------
the test filter and was used to maintain gas temperature in the test filter
zone. After removal of particles by the test filter, hot gases exit the
chamber through a pipe in the top of the vessel.
Figure IV-8 illustrates the installation of the test filter on a flue
gas slipstream downstream of the miniplant second stage cyclone. The flue gas
flow rate through the filter was 1.0 to 2.3 Nm3/min.
The electronic time sequencer controlled the operation of the cleaning
cycle. Cleaning cycle parameters were adjustable but the basic cleaning
sequence was as follows:
t Start on a timed interval by closing a solenoid valve downstream
taking the filter off-line
t Start a gentle reverse flow of unheated gas
• Release one or more cleaning pulses, (amplitude, duration and
pulse interval are all adjustable)
• Wait several seconds for dust removed during pulsing to fall into
the dust hopper
t Stop reverse flow
• Open the downstream solenoid valve, returning the filter to service
Figure IV-9 is a photograph of the filter unit installed on the miniplant.
The slipstream for the bag filter leaves the main flow duct through a 1 inch
pipe. Two high-temperature valves, a 1-inch Mohawk ceramic gate valve and a
1-inch Kamyr ball valve, were used to isolate the filter from the PFBC. Just
before the filter vessel, a gas bypass line allowed extra gas to be withdrawn
from the PFBC to maintain temperature in the inlet line. This bypass line was
also used to preheat the inlet line prior to the start of filtration.
The filtered gas leaving the top of the filter pressure vessel cooled
down to 440°C before it entered the Balston total filter shown in Figure IV-8.
The weight gain of this filter was used to determine the outlet particulate
concentration. The gas was further cooled and the water removed in a knockout
vessel before it was measured through a flow orifice and expanded through a
ball valve. Pressure drop across the ceramic filter was continuously measured
and recorded. Inlet particulate concentration was measured by extracting a
sample and passing it through a Balston total filter.
The filter was evaluated during runs 82 through 96. Runs 82 to 85 were
devoted to system shakedown. Typically, a run lasted for one working day. At
the end of that time the Acurex technician changed the test filter and a new
test was begun the following day.
Several problems with valves occurred during the shakedown runs 82 through
85. The Kamyr valve failed during run 83. It was removed and not replaced.
The Mohawk valve bonnet leaked during run 84. That gasket was replaced with a
copper gasket and the valve performed satisfactorily until the alumina gate
cracked during run 93. The solenoid valve that shuts off the filtered gas
107
-------
FIGURE IV-8
ACUREX TEST FILTER INSTALLATION SCHEMATIC
GAS BYPASS
o
00
KAMYR
WATER
KNOCK
OUT
SAMPLE PROBE
Reverse
1
1
^
1
Flow
Valve
1
PULSE '
WAIVE
BALSTON
TOTAL
FILTER
COOLER
FLOW ORIFICE
910 kPa
CLEAN AIR
1300 kPa
CLEAN AIR
ELECTRONIC
TIME
SEQUENCER
%* 1/2" KAMYR
i
.j
1" BALL
(FLOW RATE
CONTROL)
1" SOLENOID
OPERATED BALL
•>TO SCRUBBER
-> TO BALSTON FILTER SAMPLING SYSTEM
1/2" KAMYR
1" MOHAWK
OFF GAS FROM SECOND CYCLONE
-------
FIGURE IV-9
ACUREX HTHP CERAMIC BAG FILTER SITE
FILTER SUPPORT CAGE
-------
flow during cleaning failed in run 85. The teflon seat of this valve had
become damaged by hot gas. The valve was replaced with a solenoid operated
ball valve which functioned well for the remainder of the test. Otherwise,
shakedown ran smoothly and was mainly used to optimize the cleaning phase of
the filtration cycle.
Test Results
Pressure drop across the filter bags varied as a function of time in a
manner typical of fabric filters. Figure IV-10 illustrates typical pressure
drop and flow recordings. Baseline pressure drop was defined as the pressure
drop of a bag at the start of the filtration cycle. Graphically, this point
corresponds to the first point in each cycle in Figure IV-10. Many filters
were operated for several hours at baseline pressure drops close to that of
a new bag. The baseline pressure drop was always between 0.1 and 5.0 kPa,
usually between 0.2 and 2.0 kPa. As shown in Table IV-7, the baseline
pressure drop did increase with time in a number of runs, but this is typical
of fabric filters during initial operation. Pressure drops before cleaning
were never allowed to exceed 14 kPa to reduce the chance of bag failure.
High pressure drops generally caused the inner filter support screen to bulge
into the cage on which the filter was fastened. Baseline pressure drops were
slightly higher when filtering Champion coal than when filtering Illinois
No. 6 coal under similar conditions. Outlet particulate loadings were
slightly lower with Champion coal than with Illinois No. 6 coal. The overall
influence of coal type was small, and could not be quantified from the
relatively few tests completed at the miniplant. Filter bags were used for
from 4.5 to 19 hours. A summary of test conditions, pressure drops, and out-
let particulate loadings is shown in Table IV-7.
Filtration efficiencies for the Acurex ceramic bag filter were generally
over 90%, ranging from 96 to 99.5%. An exact filtration efficiency was dif-
ficult to determine because of problems in measuring the filter inlet parti-
culate concentration. Filter inlet particulate concentration was measured or
calculated by three methods: (1) Balston total filter catch on an extracted
sample, (2) mass balance around the third miniplant cyclone, (3) mass balance
around the ceramic bag filter. The results obtained by these three techniques
were not consistent as shown in Table IV-8.
The test filter inlet line (Figure IV-11) was 1-inch schedule 80 pipe
taking a sample from a 4-inch schedule 5 pipe. Isokinetic flow would have
been 1.15 Nm3/min. This flow alone, would have cooled to 450°C before reach-
ing the filter vessel. A bypass flow of 1.4-2.3 Nm3/min in addition to the
filtered gas, was drawn through the line to help maintain temperature near
800°C. The flow into the filter inlet line was therefore 200 to 300 percent
isokinetic. The Balston total filter inlet sample (1/4-inch probe) was incor-
rectly operated isokinetically with respect to the filtered gas only, neglect-
ing the bypass gas which was present at that point. Therefore, the Balston
total filter samples were taken at only 30 to 50% isokinetic rates. For this
particle size range and loading, isokinetic flow appears to be important, and
this inlet loading, measured with the Balston filter, may be treated as a
lower limit of particle concentration.
110
-------
FIGURE IV-10
ACUREX HTHP CERAMIC BAG FILTER PRESSURE DROP AND FLOW
O 70
CN
I
E 60
O 50
m 40
0£
30
at 20
LLJ
E 10
0
CLEANING
^PULSES
OBASELINE
5 MIN
TIME
?1.5
0
"-0.5
cd
LLJ
1—
_l
0
s ,
-
CLEANINGr1
CYC LE ^
TIME
-5 MIN
111
-------
TABLE IV-7. ACUREX HTHP BAG FILTRATION SUMMARY
(1)
(2)
(3)
(4)
(5)
Run No
83
D4
85
86
87.1
87.2
87.3
88.1
80.2
88.3
89.1
89.2
89.3
89.4
90
91.1
91.2
91.3
92.1
92.2
93.1
93.2
94
95
96.1
96.2
96.3
96.4
96.5
96.6
96.7
Coa
Outlet
1 Load
Flow Face Average
Rate Velocity Tempera tui
Ba
-------
TABLE IV-8. ACUREX HTHP BAG FILTER INLET PARTICULATE LOADINGS
Run No.
86
87.1
87.2
87.3
88.1
88.2
88.3
89.1
89.2
89.3
89.4
90
91.1
91.2
91.3
92.1
92.2
93.1
93.2
94
95
96.1
96.2
96.3
96.4
96.5
96.6
96.7
Inlet
Sample by
Balston Total
Filter at Inlet
0.40
0.31
0.30
0.32
0.30
0.60
0.60
0.48
0.48
0.37
0.37
0.47
0.40
0.40
0.57
N/A
N/A
1.22
1.22
0.53
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
Particulate Loading (g/m )
Calculated from Calculated from
Mass Balance Mass Balance
Around 3rd Cyclone Around Filter
0.79 N/A
1.09 N/A
1.09 N/A
1.09 N/A
0.77 N/A
0.77 N/A
0.77 N/A
1.08 N/A
1.08 N/A
1.08 N/A
1.08 N/A
1.35 0.48
1.06 0.48
1.06 0.48
1.06 0.48
0.92 0.84
0.92 0.84
1.16 0.84
1.16 0.84
1.16 0.73
0.96 0.73
1.06 0.73
0.94 0.73
0.94 0.73
1.65 0.73
1.65 0.73
1.65 0.73
1.65 0.73
N/A = Not Available
113
-------
FIGURE IV-11
ACUREX HIGH-TEMPERATURE CERAMIC BAG FILTER GAS INLET SCHEMATIC
ACUREX
CERAMIC
BAG
FILTER
BYPASS GAS TO SCRUBBER
1.4-1 .9 Nm3/mm
ACUREX BAG FILTER
INLET PROBE (1" SCH. 80)
1 .9-3.4 Nm3/min
(200-300% ISOKINETIC)
KAMYR
VALVE
MOHAWK
VALVE
1 .0 to 2.0 Nm /min
BALSTON FILTER
SAMPLE FLOW
0.034-0.056 Nm3/min
FLUE GAS FROM
2ND CYCLONE
VENT
BALSTON
FILTER
X
4 INCH SCHEDULE
5 PIPE
WET TEST
METER
•LUE GAS
16.6 Nm3/min
TO 3RD CYCLONE
-------
A mass balance around the filter test vessel was attempted to resolve the
Inlet loading Issue. Weight of the filter bags was not determined before
exposure, so a tare of 0,2 kg was assumed by weighing other unexposed bags.
These inlet loadings, intermediate to the other two results, can still be con-
sidered low because of the multitude of places where particulates could have
been lost during cleaning and dismantling operations. However, from these
three calculation methods a reasonable estimate can be made of the actual
inlet particulate concentration.
Size distribution of the inlet particulate matter was not obtained since
the samples were not taken under Isokinetic conditions.
The bag filter outlet particulate concentration was determined by passing
the entire filtered gas flow through a Balston total filter. The total par-
ticulate concentration was obtained by weighing the total filter before and
after exposure. Overall ceramic bag filter efficiencies are shown 1n Table
IV-9. These were calculated using the three methods of determining inlet
concentration discussed previously. Despite some uncertainty in the inlet
particulate concentrations, the collection efficiencies calculated by the
three methods were generally in good agreement.
A size distribution of the outlet particulates could not be obtained.
The amount of particulates on the Balston filter was so low that Insufficient
material was available for Coulter Counter analysis. The filters were washed
off with Isoton II in an attempt to remove particulates without mechanical
brushing. This method caused enough Balston filter material to be washed Into
solution to obscure completely the flyash particulates. A clean Balston
filter, not exposed to any flyash but also washed with Isoton gave a sample
which had a size distribution similar to that obtained from a used filter.
During the tests at the miniplant, eight single and one double thickness
bags were exposed to PFBC conditions, as shown in Table IV-7. Most bags were
exposed for 6 hours or more. Averaging the face velocity and exit particulate
concentration over the first 6 hours of new bag exposure and plotting outlet
loading as a function of face velocity provided the data shown 1n Figure IV-12.
Bag number 4 results were not recorded on Figure IV-12 because of problems with
the outlet filter. Bag number 7 was a double thickness bag ("2 cm). Normally
filter media thickness was about 1 centimeter. Bag 7 was physically less dis-
torted and its pressure drop was less than bag 8 (a 1-cm thick bag which was
run at similar conditions), and its filtration efficiency was much lower. The
reason for the lower AP and lower efficiency for the double-thickness bag is
not certain (it probably had a leak). As seen in Figure IV-12, the parti-
culate penetration of a bag seems to increase with face velocity. The trend
lines shown on the curve were selected by drawing a line through the Illinois
No. 6 results, which fell on a straight line, and then drawing a suitable
parallel line for the Champion coal. However, the data may not be as precise
as implied by this curve. For example, one objective of the off-line cleaning
used in this study is to offset the increased penetration with velocity that
is normally seen in filter tests and suggested by Figure IV-12. More testing
1s still required to determine better the effect of face velocity and off-line
cleaning on penetration.
115
-------
CT>
Run No.
TABLE IY-9. ACUREX HTHP BAG FILTER COLLECTION EFFICIENCY
Outlet
Particulate
Loading
(g/m3)
86
87.1
87.2
87.3
88.1
88.2
88.3
89.1
89.2
89.3
89.4
90
91.1
91.2
91.3
92.1
92.2
93.1
93.2
94
95
96.1
96.2
96.3
96.4
96.5
96.6(D
96.7(1)
0.0093
0.001
0.009
0.003
0.021
0.014
0.011
0.023
0.023
0.023
0.023
0.009
0.007
0.003
0.004
0.006
0.005
0.007
0.005
0.066
0.012
0.016
0.007
0.007
0.016
0.030
0.187
0.244
Collection Efficiency (%) Based
Upon Alternative Inlpt nptprmi nations
Sampled by
Balston Total
Filter at Inlet
97.6
95.2
97.1
99.1
93.0
97.7
98.1
95.3
95.3
93.9
93.9
98.1
98.3
99.1
99.4
__
..
99.4
99.6
87.6
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
Calculated from
Mass Balance
Around 3rd Cyclone
98.8
98.7
99.1
99.8
97.3
98.2
98.5
97.9
97.9
97.9
97.9
99.3
99.4
99.7
99.7
99.4
99.5
99.4
99.6
94.3
98.8
98.6
99.4
99.4
98.6
97.4
83.5
78.4
Calculated
from
Mass Balance
Around Fil
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
98.1
98.6
99.3
99.2
99.3
99.4
99.2
99.5
90.9
98.4
97.8
99.1
99.1
97.9
95.9
74.4
66.6
ter
Average
98.2
97.0
98.
99.
95.
98,
98.
96.
96.
95.
95.
98.
98.8
99.4
99.4
99.
99.
99.
99.
90.
98.
98.
99.
99.
98.
96.
78.
72.5
(1) Bag Failed
N/A = Not Available
-------
FIGURE IV-12
ACUREX HTHP BAG FILTER OUTLET LOADING VS. FACE
VELOCITY (AVERAGED OVER FIRST 6 HOURS OF EXPOSURE)
FACE VELOCITY, (FPM)
0.06
CO
E
Z 0.05
CD
Z
< 0.04
O
LU
=1 0.03
u
Qi
°-
5 0.02
i—
O
0.01
5 10 lo zo
i i 1*1
7*
• CHAMPION COAL
O ILLINOIS NO. 6 COAL
8 BAG NUMBER
* DOUBLE FILTER THICKNESS BAG
•
•
3A^
9 3^^^
**~'&T^
^6
i i I i i i
123 456
0.030
0.025
u.
0)
o
•
o
ro
o
LOADING,
0.015 uj
D
y
fe
0.010 £
LU
5
O
0.005
FACE VELOCITY, m/min
117
-------
FIGURE IV-13
ACUREX CERAMIC BAG FILTER—BAG NO. 5
PARTICIPATE PENETRATION HISTORY
0.015
0.010
z
o
u
o
u
LJJ
13
y 0.005
c*.
o_
O
\
\RUN 90
\RUN 91.1
x
l
iii
l
5 10
BAG AGE (HOURS)
15
118
-------
Outlet loading plotted in Figure IV-12 was averaged over the first 6
hours of operation. Outlet loading actually decreased as a function of time
in the same fashion that a conventional filter media test would show in similar
tests under ambient conditions. This decrease as a function of time is shown
as Figure IV-13. This bag was exposed to Champion coal at 775°C for a total
of 13 hours. Along with the decrease in filter particle outlet loading there
was an increase in baseline pressure drop from 0.1 to 3.0 kPa as expected as
the filter cake formed and the efficiency improved.
The effect of bag cleaning conditions was not studied. After adjusting
cleaning conditions during the first two runs, conditions were fixed for the
rest of the runs at 6 pulses of 1300 kPa air for about 0.75 s each with a 3 s
interval between pulses.
Examination of the filters after a test showed that the dust cake was
deposited mostly on the surface of the filter media and the dust cake could be
removed easily. Figure IV-14 shows bag number 3 immediately after removal
from the filtration vessel. Figure IV-15 is a close up of the same bag after
a strip was vacummed clean. This strip had the appearance of a virtually new
bag indicating very little dust penetration into the media.
At the conclusion of the series of short runs, run 96, a long continuous
test of the bag filter was attempted at conditions deemed optimum for extensive
testing. Filtration commenced smoothly, however, the baseline pressure drop
across the bag continued to increase during the first 6 hours of filtration.
A possible reason for this increase became clear at the conclusion of the run
when it was discovered that the pressure regulator used to set the pressure of
the reverse flush air was set only slightly higher than filter vessel system
pressure. It is possible that a slightly higher combustor pressure could have
reduced reverse flush air flow to a level too small to clean the filter
effectively. Since the reverse flush air flow rate was not measured, this
hypothesis cannot be confirmed.
After 9 hours of filtration during run 96, a pressure drop of over 12.5
kPa across the filter was thought to be excessive for continued bag life. The
vessel pressure was reduced to 300 kPa and a full high pressure (1300 kPa)
pulse blow back was initiated into the lowered system pressure. This blow
back reduced the pressure drop almost back to the clean baseline condition
(0.5 kPa). However, outlet particulate concentration increased over the next
10 hours until it was almost identical to the inlet concentration. The run
was terminated at that point. As expected, the bag had failed (see Figure
IV-16). The bag failure probably began with the high pulse pressure blow back
against the low system pressure, and was made worse by subsequent blow backs.
The final cause of the failure appeared to be high temperature corrosion of
the thin 304 stainless steel filter support screen. The blown out appearance
was probably caused by the high pressure pulse into the lowered system pressure
environment after the corrosion weakened support screen failed.
It is recognized that corrosion of a metal support screen is a potential
problem in long term applications. The metal support screens used in these
tests were used only for ease of construction. In addition, the early failure
119
-------
FIGURE IV-14
CERAMIC FILTER BAG NO. 3
120
-------
Vacuumed Strip
FIGURE IV-15
CERAMIC FILTER BAG NO. 3 CLOSEUP
OF VACUUMED STRIP
,1
I • ' I ••
*'
v,
r,,Jh,v'
Ere
•H
\ I
121
-------
FIGURE IV-16
CERAMIC FILTER BAG NO. 9 AFTER RUN 96
122
-------
of the support screen experienced in run 96 was probably caused by a malfunc-
tion in the temperature control system which caused the filter element to be
overheated. Flexible woven ceramic fabric screens are available to perform
this function for applications requiring greater corrosion resistance.
Summary and Conclusions
During the miniplant tests, baseline pressure drops of under 2 kPa were
maintained for over 6 hours average duration at face velocities of up to 6.0
m/min with removal efficiencies of 95 to 99%.
The average particulate concentration at the bag filter outlet was measured
to be 0.013 +_ 0.005 g/Nm3. This is less than one-half the EPA New Source Emis-
sion Standard for particulates (0.03 Ibs/MBTU). The particulate concentration
was as low as one order of magnitude below the emission standard. In compar-
ison to 3 stage cyclone cleanup, the average improvement was approximately a
factor of 7. The particulate concentration with the ceramic filter easily
meets both the EPA and most gas turbine standards,
These tests proved the ceramic filter was cleanable while subjected to
flyash generated under PFBC conditions. High collection efficiency at high
face velocity was also shown. In general, the test filter exhibited perfor-
mance similar to that which would have been expected from a filter unit opera-
ting under more common conditions.
The failure of the filter element in the final test was a result of an
unusual combination of events and is not an indication of an inherent filter
problem. Further investigation of this potentially promising technique 1s
felt to be warranted.
GRANULAR BED FILTRATION STUDIES
Granular bed filters were also evaluated as high efficiency particulate
control devices. Originally, the objectives of the test program were to measure
particulate removal efficiency, operational stability and long term life of the
filter hardware. However, operating difficulties, primarily caused by plugging
of the filter inlet sections by the particulates, poor bed cleaning and loss
of filter media during cleaning caused a change in the program direction.
The program was then directed toward modifying the granular bed filter system
to overcome the operating problems. The previous report describes the filter
systems studied, the problems encountered and modifications made to solve the
problems (1). These modifications allowed a 24 hour run to be completed.
However, filtration efficiency decreased sharply during the test. It was
found that the filter beds had not been adequately cleaned and gas was blowing
through the filter beds in "rat holes." Additional modifications were sub-
sequently made and the modified systems tested in an attempt to improve perfor-
mance. These latter tests are described in this report.
Equipment
The filter configuration used at the start of the test program discussed
in this report was described in detail in the previous report (1). This system
123
-------
was installed on the miniplant following the second cyclone, and was capable
of accepting the total flue gas flow. It consisted of two filter elements,
each composed of five filter beds, as shown in Figure IV-17. A schematic of
one of the filter beds is shown in Figure IV-18. The filter media contained
in each bed consisted of granular alumina or quartz. This design had no
screens across the dirty gas inlet opening at the top. This was done to pre-
vent inlet screen plugging. The freeboard height was also increased to 18 cm
to prevent loss of the filter media during the reverse (upflow) cleaning step
with clean compressed air. In the cleaning step, the beds are fluidized by
reverse-flow air, the fine particulates retained by the bed from the previous
filtration step are entrained from the filter beds and blown out of the inlet
opening at the top of the beds in a reverse direction. Figure IV-19 is a
schematic which illustrates the operation of a filter element during the
filtration and blow back (cleaning) steps. The fluidizing grids at the bottom
of each bed were also redesigned to give better gas distribution during blow
back. Each grid contained 56, 0.5 cm diameter holes.
Each filter element was installed in a shroud as shown in Figure IV-19.
Flyash blown from an element during blow back is retained within the shroud,
falls to the bottom and is collected in a lock hopper. Two filter elements
were installed in parallel after the second stage cyclone in a single large
refractory lined pressure vessel. The inside dimensions of the vessel are
approximately 2.4 m diameter by 3.4 m inside height. The vessel is connected
to the second stage cyclone by 12 inch pipe refractory lined and sleeved with
stainless steel to an inside diameter of 10 cm.
Dirty gas entering the filter vessel is piped to each shroud, passing
through orifices which measure flow rate to each filter element. Clean gas
leaving each shroud fills the interior of the pressure vessel before leaving
through a single outlet line. Blow back air enters each filter element
through flanges at the top of the pressure vessel and flows in a reverse
direction through each filter element. A blow back nozzle and seal plate are
dropped down to engage the top of the filter element during blow back and
direct the blow back gas in the proper direction. This is shown in Figure
IV-19. The blow back gas leaving a filter element flows in a reverse direc-
tion through the clean gas inlet system into the other filter element which
is in a filtration step. Each element is blown back separately.
A natural gas burner was installed to preheat the interior of the pres-
sure vessel to a temperature above the dew point of the combustor flue gas
before starting a filtration test. The burner fires into the vessel through a
side port for an 8 to 12 hour period prior to the start of a run. The filter
vessel is at atmospheric pressure during this period.
Natural gas injection into the flue gas piping between the second stage
cyclone and the filter vessel was used to keep the flue gas temperature above
840°C. Gas was injected at four points, 2.3 m, 6.7 m, 9.8 m and 14.6 m down-
stream of the second stage cyclone. Approximately 0.03 Nm3/m1n of natural gas
was injected through each of the four inlets.
All of the above systems are described in more detail in the previous
report (1).
124
-------
FIGURE IV-17
MODIFIED GRANULAR BED FILTER ELEMENT
(WITHOUT SHROUD)
125
-------
FIGURE IV-18
MODIFIED FILTER BED
tMspmw
DIRTY
C, A ci ••—•
INLET
1
—
______ _
1
._ ____.
CLEAN GAS
OUTLET
l 1 .3 cm
t
21.9cm
FLUIDIZING GRID
"(56-^.36 cm
i did. holes)
-7.3 cm
21 .9cm
126
-------
FIGURE IV-19
GRANULAR BED FILTER SCHEMATIC
FILTRATION CYCLE
BLOWBACK CYCLE
FILTER MEDIA
SHROUD
Ly/y>
CLEAN
GAS EXIT
DIRTY GAS
FIVE BEDS
!f
1
7
—
8*
M
[be
u
•-
C
-1
11
C
»
i
:
)
i
j
i
j
OB
M
aa
J
ml
FLY ASH,
F-
'*^ DIRTY GAS
FLUIDIZED
FILTER MEDIA
LOCK
HOPPER
LOCK
HOPPER
127
-------
System Modifications--
In order to improve the operation of the filter a few modifications were
made to the filter system. The modifications were used at various times dur-
ing the test program. None of them were permanent or needed for all tests.
The section dealing with test results will specify which modifications were
used for the various runs. This section will describe the modifications.
Blow back gas ejector--A blow back gas ejector was installed to replace
the blow back gas inlet nozzle and seal plate assembly. It was described in
the previous report (1). Briefly, the ejector used a small amount of motor
air during blow back, but pulled in additional clean gas by the action of the
ejector to supplement the blow back gas. There were no moving parts and no
sealing surfaces. The ejector was tested during two runs and removed.
Inlet screens--In the above sections describing the design of the filter
it was mentioned that the inlet screens were removed to prevent plugging and
the bed freeboards increased to prevent loss of filter media during blow back.
However, loss of filter media still occurred despite the increased bed free-
board and attempts to minimize blow back gas velocity. Therefore, for most
runs, screens were strapped over the inlet slot. Three screen mesh sizes were
used, 10 x 10, 20 x 20 and 50 x 50.
Baffles—Another modification was made to prevent loss of filter media,
in this case, by the installation of baffles in the filter beds near the gas
inlet. Figure IV-20 is a sketch of the baffle arrangement. The baffles were
tested in one run.
Filter media —Initial tests were carried out using quartz granules with
a particle size range of 250 to 600 ym. In subsequent tests, 850 to 1400 urn
alumina particles were used. A very dense iron oxide (speculite) was also
tested in two runs in an attempt to minimize filter media losses. The specu-
lite particle size range was 400 to 2000 ym.
Outlet piping--Initially. the clean gas from the filter elements filled
the interior of the pressure vessel and left through a single vessel outlet
line. However, heat losses from the vessel were high and the temperature
drop between the filter inlet and outlet lines was too great. The filter
element outlets were then piped directly to the vessel outlet and the new
piping and the filter shrouds were insulated to reduce heat loss.
Ambient Temperature Models —
Because of various operating problems, especially loss of filter media
and poor bed cleaning, transparent plastic models of the filter beds were
built and tested. All the models operated at ambient temperature and close
to atmospheric pressure. The first model was simply a plastic pipe, 2.4 cm
in diameter. The pipe was filled with granular filter media, dust was added
manually and the bed blown back with compressed air.
A two-dimensional plexiglas model of a granular bed filter was also
built and tested. A schematic of the model is shown in Figure IV-21. The
system was designed to operate in both the filtration mode and blow back
mode although it was only used to study blow back operations. The model
128
-------
FIGURE IV-20
FILTER BED WITH INTERNAL BAFFLE
I / / / / /// / 77
BAFFLES
vr//// ///n
///////
1.3cm-*| [««-
j
i
WA
V //////.
/ /
i
r
u
i™
}
/ /
/ \
1
"^
^
r
i-
=
j
s
*
•>
:,
n
•
=
D
S
•
6.9 cm
129
-------
FIGURE IV-21
PLEXIGLAS MODEL GBF
CA>
O
FILTRATION
VENT
BLOW BACK
VENT
FILTER
3-WAY
VALVE
IROTAMETER
SOLENOID
VALVE
/^PRESSURE GAUGE
T
.ATE
TOR
i^^^mmmmm
\
w,
1
jr
*
M 1
FILTRATION
FLOW SHUT-OFF
VALVE
1
—
W
—
m
PL!
—
M
—
VA
:X
GLA.
-^~~
BACK PRESSURE
FLOW CONTROL
VALVE
IROTAMETER
BLOWBACK
AIR
PRESSURE
REGULATOR
-------
consists of a two dimensional slice of two filter beds each with a filtra-
tion area of 0.01 m2. Each filter bed is 23 cm high with two 1.3 cm inlet
slots. The filter media is supported by a 50 x 50 mesh screen supported by
a fluidizing grid. Dirty gas enters the filter through two inlets, each feed-
ing one side of the filter beds. Filtered gas exits the filter through a
5 cm central core and is passed through a backup filter to allow an estimation
of collection efficiency to be made. Blow back air enters through the central
core and is distributed to the filter beds. After passing through the filter
beds, the blow back air is then vented to the atmosphere.
Another simple plastic model was built. This was 7.6 by 7.6 cm in cross
section and 7.6 cm high. The chief feature of this unit is the design of the
blow back system. Blow back air can be added through the bed support grid as
in the other units, but it can also be added through a ring sparger, located
at the top of the bed along the wall. The blow back air from the sparger is
directed horizontally and was intended to shear the dust layer from the top
of the bed without disturbing the filter bed itself to a large degree.
A second high temperature/high pressure filter was built and tested.
This was a much smaller unit operated on a slipstream of flue gas withdrawn
after the second cyclone. This unit and results of the tests are described
in a later section.
Experimental Results
Six runs were made on the larger filter system during the reporting
period. A summary of run conditions is given 1n Table IV-10. The changes 1n
filter configuration described above were made during the test program. In
addition, blow back conditions were varied to examine the effect of blow back
gas velocity, duration and frequency on cleaning of the filter beds. At the
conclusion of the six runs, an attempt was made to use the granular bed filter
in a 100 hour test as part of the DOE sponsored program on hot corrosion and
erosion of gas turbine materials.
The experimental program was not successful. Three major problem areas
developed which were not satisfactorily resolved:
• low particulate outlet concentrations could not be maintained
for more than a few hours,
• loss of filter media could not be prevented,
• serious operating problems persisted to the end of the program.
Outlet Particulate Concentration and Size Distribution--
Outlet particulate concentrations as low as 0.07 to 0.11 g/Nm3 were
measured in several runs. However, in all cases, the concentrations Increased
during the runs as much as a factor of three In 8 to 10 hours. Results from
two typical runs are given 1n Figure IV-22. It is believed that this poor
removal efficiency and the increase in particulate concentration with time
were due to poor dust removal from the beds during blow back. This was suppor-
ted by the observation that the filter beds, after completion of a run, were
generally found to contain a very high concentration of dust intimately mixed
131
-------
TABLE IV-10. GRANULAR BED FILTER RUN SUMMARY
GO
ro
Run No.
Filter Description
Number of Parallel Elements
Number of Beds/Element
Retaining Screen
Miter Media
Bed Depth (cm)
Filter Media Part. Size (ym)
Other Information
Operating Conditions
Preheat Temperature (°C)
Filter Inlet Temperature (°C)
Filter Outlet Temperature (°C)
Filter Inlet Pressure (kPaa)
Filter Face Velocity (m/s)
Filter AP After Blow Back (kPa)
Filter AP Before Blow Back (kPa)
Run Length (hrs)
Blow Back Conditions
Superficial Velocity (m/s)
Duration (s)
Interval Between Blow Back (min)
Partlculate 'Emissions
Outlet Partlculate Concentration (g/Nm )
60
2
5
50 Mesh
Quartz
6.4
250-600
61
2
5
50 Mesh
Alumina
6.4
850-1400
Ejector on One Element Ejector on One Element
Run Unsuccessful
Because of Poor Flow
Distribution and High
AP Measurement
760
840
690
900
0.26 and 0.47
9.6 and 30({
24 and 450)
5.5
0.61
8
10-15
1.1
(1)
0)
(1) Values for the two ftlter elements are different due to unequal flow split.
-------
TABLE IV-10 (CONT'D). GRANULAR BED FILTER RUN SUMMARY
CO
CO
Run No.
Filter Description
Number of Parallel Elements
Number of Beds/Element
Retaining Screen
Filter Media
Bed Depth (cm)
Filter Media Part. Size (urn)
Other Information
Operating Conditions
Preheat Temperature (°C)
Filter Inlet Temperature (°C)
Filter Outlet Temperature (°C)
Filter Inlet Pressure (kPaa)
Filter Face Velocity (m/s)
Filter AP After Blow Back (kPa)
Filter AP Before Blow Back (kPa)
Run Length (hrs)
Blow Back Conditions
Superficial Velocity (m/s)
Duration (s)
Interval Between Blow Back (min)
Particulate Emissions
Outlet Particulate Concentration (g/Nm )
62.1
2
5
10 Mesh
Alumina
6.4
850-1400
0.30
8
15
0.069
62.2
2
5
10 Mesh
Alumina
6.4
850-1400
760
830
700
900
0.41
14 to 20
35
2
__
830
700
900
0.41
20 to 24
35
2
0.30
4
10
62.3
2
5
10 Mesh
Alumina
6.4
850-1400
830
700
900
0.41
24
35
2
0.30
2
10
0.48
63
2
5
None
Speculite
1.9
425-2000
760
900
800
900
0.47
22 to 33
41
12
0.46
8
6
0.11,0.16,0.27
-------
TABLE IV-10 (CONT'D). GRANULAR BED FILTER RUN SUMMARY
CO
Run No.
Filter Description
Number of Parallel Elements
Number of Beds/Element
Retaining Screen
Filter Media
Bed Depth (m)
Filter Media Part. Size (ym)
Other Information
Operating Conditions
Preheat Temperature (°C)
Filter Inlet Temperature (°C)
Filter Outlet Temperature (°C)
Filter Inlet Pressure (kPaa)
Filter Face Velocity (m/s)
Filter AP After Blow Back (kPa)
Filter AP Before Blow Back (kPa)
Run Length (hrs)
Blow Back Conditions
Superficial Velocity (m/s)
Duration (s)
Interval Between Blow Back (min)
Partlculate Emissions
Outlet Particulate Concentration (g/Nm )
64
2
5
None
Speculite
6.4
425-2000
760
950
840
900
0.36
10 and 14
22 and 26
7
65
(2)
,
U
0.82
4 to 5
3 to 5
0.64,0.66,0.62,0.27
2
3
20 Mesh (One Element)
Alumina
6.4
850-1400
Baffles On One Element (No Screen)
760
850
800
860
0.51
7 to 34
34 to 62
8.5
0.61
4
4
0.11,0.14
(1) Values for the two filter elements are different due to unequal flow split.
(2) A third cyclone was installed between the second cyclone and the filter vessel during this
run, thus reducing the inlet particle concentration to the filter
-------
FIGURE IV-22
INCREASE IN OUTLET PARTICULATE CONCENTRATION WITH TIME
w
i_
&
z
O 0.12
U
O 0.08
u
3
ID
y
t 0.04
ID
O
0
0
a MINIPLANT RUN 63
O MINIPLANT RUN 65
D
O
EPA EMISSION STANDARD
j 1 i ' '
4 6
TIME INTO RUN (MRS)
8
0.3
0.2
O
c
70
^H
n
c
n
O
Z
n
m
Z
0.1 5
I
CO
10
135
-------
with the filter media. Dust concentrations of 10 to 3Q% were typically mea-
sured. Tests made in the transparent models showed that the dust was only
partly removed during blow back, some of it adhered to the filter media and
was mixed into the bed during blow back by the motion of the fluidized filter
media granules. The fine particles present in the PFBC flue gas are apparently
very adhesive and are not removed easily during blow back. This mixing pro-
cess probably explains the observed uniform distribution of the fine parti-
culates throughout the filter beds after extended runs. Once the fine parti -
culates were mixed into the filter media, they could be carried out through the
bed support grid during the filtration step because of locally high gas veloc-
ities at the grid holes. As seen in Figure IV-22, at no time were outlet con-
centrations measured which meet the current emission standard of 0.03 Ib/MBTU.
Attempts were made to improve bed cleaning by blowing back more frequently
for longer periods and at higher velocities. The blow back velocity was
limited by loss of filter media (with alumina or quartz media) to 0.3 to 0.6
m/s. The duration of the blow back was usually 4 to 6 s, although durations as
short as 2 s and as long as 8 s were tested. The frequency of blow back was
reduced from once every 15 min. to once every 4 min. As a result of these
changes, the degree of filter cleaning appeared to improve, but the best run
made, run 65, still showed an increase in outlet particulate concentration with
time.
Another factor which could be responsible for the retention of particulates
in the filter beds was the recycling of particulates from bed to bed during the
blow back step. The shrouds which surround each element may impede the settling
into the lock hoppers of the particulates blown from the elements during blow
back. That is, if a new filtration step began before the particulates fell
into the hoppers, the particulates could get carried back into the filter ele-
ments. This could result in high internal recirculation rates between elements,
causing the particulates to build up in the beds and thereby causing an
increase in the outlet particulate concentration.
In discussions with representatives of Ducon Company, the original sup-
plier of the filters, Ducon claimed that the filtration velocities used in
these tests were excessively high and could have promoted dust penetration
(12). Ducon suggested a velocity of about 0.2 m/s.
Regardless of the reasons, the outlet concentrations measured in these
tests were consistently higher than those measured when a third stage cyclone
was used instead of the filter. Therefore, the filter showed no efficiency
advantage over the cyclone. It should also be mentioned that the best run in
the series of six shown in Table IV-10, run 65, was made with a cyclone
installed between the second stage cyclone and the filter. The cyclone was
not designed for the flow rate used in run 65 and was probably not too effi-
cient, but 1t still probably contributed to the improved performance observed
in that run. The cyclone was later removed and replaced with a more efficient
cyclone which was used for the tests described in the earlier section on
cyclone performance.
136
-------
Filtration efficiencies were not measured during the test program since
Inlet partlculate loadings could not be measured.
Typical particle size distributions of the dust entering and leaving the
granular bed filter are given in Table IV-11. Additional particle size data
after the filter for some of the filter runs are presented 1n Appendix M-6.
TABLE IV-11. PARTICLE SIZE DISTRIBUTIONS
Percent Less Than Particle Size (ym)
5% 10% 25% 50% 75% 90% 95%
Filter Inlet 1.5 2.3 3.5 6.9 13 32 45
Filter Outlet 1.1 1.4 2.0 3.0 5.1 12 15
The outlet size distribution is relatively coarse when compared to typical
distributions measured from the third stage cyclone. The cyclone consistently
produced participates with a median size of 1 to 2 ym compared to 3 ym shown
in Table IV-11. The coarser particulate distribution 1n the filter outlet was
also probably caused by dust penetration resulting from poor bed cleaning.
Loss of Filter Media--
The loss of filter media during blow back continued throughout the program
and was never satisfactorily prevented. Screens were used in most runs to pre-
vent loss, but the screens consistely plugged with flue gas partlculate even
though screen opening sizes as large as 10 mesh were used. The material
plugging the screens was a loose agglomerate of dry, fine particles which
could be removed by brushing. A photograph of a plugged screen was given in
the previous report (1). Small openings above the screens and below the top
flanges of each filter bed were used to allow entry of dirty gas and exit of
blow back gas. This arrangement minimized but did not prevent loss of filter
media. Alumina, 850 to 1400 ym, was used in all runs after 60 except 63 and
64 in which 425 to 2000 ym speculite was used. The screens were removed in
runs 63 and 64 since it was believed that the very dense and coarse speculite
would not be lost from the beds. However, loss of the filter media stm
occurred. In addition, 1n run 64, a temporary upset caused an Increase in the
CO concentration in the flue gas. This apparently reduced the speculite (iron
oxide) to Fe° which was then oxidized after the CO concentration decreased.
This is a highly exothermic reaction. As a result, the filter was severely
damaged. Two filter beds could not be repaired. Fine Iron oxide dust was
also found 4n the filter outlet piping sections. The use of speculite was
discontinued.
Testing of the two dimensional transparent model then began 1n an attempt
to determine the cause and possible cure for the filter media loss problem
A series of tests were conducted in which various internal baffles and inlet
gas slot sizes were used. The most significant finding from these tests was
that most of the filter media loss occurred at the beginning of the blow back
The sudden opening of the blow back valve caused a surge 1n gas flow which blew
filter media from the vessel. The baffles did not prevent loss caused by the
137
-------
flow surge although some designs did decrease losses caused by high blow back
velocities following the initial surge. Baffles were installed in one filter
element and tested in run 65. Screens were removed from this filter element
and retained on the other element. Both elements showed loss of filter media,
the baffled element without screens showed the greater loss, 32% of the start-
ing media bed compared to 18% from the second element. The use of unscreened
baffled filters was then discontinued.
The plexiglas model tests would indicate that slower acting blow back
valves or surge capacity between the valve and the filter would decrease the
loss of filter media caused by the flow surge. Ducon representatives also
claimed that filter bed freeboards 50% larger than those used in these tests
are needed to prevent loss during blow back (12).
Other Operating Problems--
In addition to the problems described above, other operating problems
occurred which persisted until the end of the program despite attempts to
improve operations. One problem was an increase in the combustor pressure
which occurred during filter blow back. At times, the increase was large
enough to cause the combustor pressure to exceed the pressure in the coal
injection vessel. When this happened, the hot solids in the combustor would
back up into the coal injector vessel, causing a fire in the vessel. A
nitrogen purge system on the coal injector was used to extinguish the fires,
but these incidents always caused the termination of the run and removal of
some charred coal from the Injector vessel.
Temperature drops across the filter was another problem. Although this
did not affect performance of the filter, it did cause problems for the gas
turbine materials test. Flue gas temperatures were required to stay above
840°C at all times to prevent condensation of alkali sulfates, the material
which could cause hot corrosion of the turbine parts. Injection of cold blow
back air and heat losses from the filter elements to the interior of the pres-
sure vessel resulted in temperature drops in excess of those required. Injec-
tion of natural gas into the filter outlets, insulation of the filter shrouds
and piping of the clean flue gas from the filter outlet directly to the pressure
vessel outlet decreased the temperature drop to an acceptable level.
An attempt was also made to replace the blow back nozzle and seal plate
with an ejector. This would decrease the amount of cold blow back air needed
and also replace the seal plate which occasionally did not work satisfactorily.
An ejector was installed on one filter element and tested in runs 60 and 61.
The second element was equipped with the nozzle and seal plate arrangement.
In the tests with the ejector, it was found that the pressure drop through the
ejector during the filtration step was high, giving rise to unequal flow rates
between the two filter elements. The element with the ejector was not cleaned
as well during blow back and the outlet loading from the test (run 61) was
very high, 1.0 g/Nm3. The ejector was removed and not tested further.
Poor bed cleaning also led to high pressure drops across the filter. At
the beginning of a run, the pressure drops were usually around 14 to 20 kPa
after bed blow back, increasing to about 28 to 35 kPa before blow back. The
138
-------
pressure drops usually increased as the run progressed, indicating poor bed
cleaning. Pressun drops of 28 to 35 kPa after blow back and 40 to 60 kPa
before blow back were measured after 8 to 15 hours of filter operation.
Another intrinsic problem with the filter was the fact that essentially
all problems were non-corrective and as such, required termination of the run,
a long wait to cool down the filter, followed by a long period to remove,
clean and replace the filter elements.
The final run made with the filter was an attempt to run a 100 hour gas
turbine exposure test (run 66). In this case, three filter elements were used
in parallel, each with three filter beds. The filters were charged with
alumina filter media (850 to 2000 ym). The inlet slots were covered with 20
mesh screens except for a 3 mm split above the screens to allow gas flow if the
screens plugged. Filter blow back occurred every 4 min. at 0.3 m/s for 6 s.
These were judged to be the best conditions to assure a successful test. Over
the first four hours of operation, a gradual increase in the filter AP was
noted which then increased sharply to 207 kPa. The blow back velocity and
duration were increased to improve bed cleaning. The AP dropped to 70 kPa, but
a significant amount of filter media was lost. As a result of the unstable
pressure drops and frequent high velocity blow backs, the combustor pressure
control was upset and hot solids back flowed into the coal injector causing a
fire. The run was terminated after 17 hours operation. After the run, it was
found that the high pressure drops were caused by partly blocked filter support
screens. At this point, the granular bed filter program involving the use of
the multi-bed filter elements contained in the large pressure vessel was ter-
minated.
During the test program, a series of tests were also made with the trans-
parent models in an attempt to determine the cause of the poor bed cleaning.
Some of the test results were described in a previous section of this report.
They showed how a portion of the dust particles adhered to the filter media
and were mixed into the filter bed by the motion of the fluidized filter media
granules during blow back. In order to prevent this, a modified blow back
scheme was tested. In this case, most of the blow back air entered the bed
through a ring sparger located at the interface between the filter media and
the dust layer. Blow back air was directed horizontally across the interface.
This resulted in a shearing action which blew off the dust layer without
disturbing the filter media. Blow back upwards through the filter media was
minimized as was the mixing effect caused by it. This technique looked pro-
mising enough to be tested under PFBC conditions.
In order to test this as well as other concepts which could improve the
filter performance, a program was planned to test a modified filter on a flue
gas slip stream. The slip stream tests were to be made with a small, single
bed filter incorporating a number of design modifications. The design of the
new unit would build on the lessons learned in the previous test program and
would attempt simply to show if the granular bed filter concept was feasible
for pressurized fluidized bed combustion applications.
139
-------
Granular Bed Filtration Slip Stream Test
The single element granular bed filter device was tested on a minlplant
slip stream during run 115. The filter element was a further modification
of the "modified Exxon filter element" described 1n Figure IV-18. The loca-
tion and Instrumentation of the test vessel which contained the filter element,
are almost Identical to those of the Acurex high temperature ceramic filter
vessel. This was done to allow a fair comparison of the two high efficiency
hot gas cleanup devices.
The filter element tested was named Exxon Mark IV to distinguish it from
earlier designs. The filter element body was the same one used in full scale
tests. Three things were changed:
1. The fluldizing grid was replaced with a support grid with 1 cm
diameter holes, in order to reduce AP and reduce the velocity
of the grid jets during below-bed blow back.
2. A fluidizing coil was installed just above the support grid.
3. An above-bed blow back coil was installed.
These three modifications were intended to lower the baseline pressure drop,
reduce bed loss during blow back and improve the effectiveness of the blow
back. Both coils have 0.84 mm holes every 1.6 cm along the coll. In the
below-bed coil, the holes are oriented approximately 45° from the vertical
to provide upward fluidizing jets. In the above-bed coil, the holes are hori-
zontal to blow the dust cake off the surface of the filter media by a shearing
action without disturbing the filter media granules to a great extent. In
this way the superficial blow back velocity also increases just above the bed
to blow more particulate and less filter media out of the bed. Air could
still be added below the support grid to aid in fluidization of the bed. A
schematic of the element 1s shown in Figure IV-23.
The test vessel which contained the Exxon Mark IV element, was installed
on an isokinetlc slip stream taken from a point between the second and third
cyclones. A schematic of the installation is shown in Figure IV-24. The
valves and electronic timers are the same ones used for the Acurex ceramic
filter tests. The high pressure (1300 kPa) air supply is used on the above-
bed coil for a pulsed blow back. The lower pressure air supply (1000 kPa)
1s used 1n the in-bed coll and for the below-bed blow back air. Particulate
samples were taken before and after the ganular bed filter using the Balston
filter method.
Results and Dlscussions--
The Exxon Mark IV filter was run for 4 hours during run 115. Before
filtration was begun, the inlet lines were preheated with flue gas to a tem-
perature above 400°C to prevent condensation 1n the line. The filter could
not be preheated. The filter element was charged to a 7 cm bed depth with
-7 kg of -8+14 U.S. mesh alumina. The blow back cycle was program 1 in Table
IV-12. The filtration period was 5 m1n with a 20 sec cleaning cycle. After
the first 45 m1n it became apparent that the pressure drop could not be con-
trolled with this blow back program. At this time, the cleaning cycle was
140
-------
FIGURE IV-23
EXXON MARK IV GRANULAR BED FILTER
(SINGLE BED)
DIRTY
GAS
INLET
IN-BED
BLOW BACK-
COIL
CLEAN GAS OUTLET
t
ABOVE-BED
BLOW BACK
COIL
SUPPORT
GRID
BELOW-BED
BLOW BACK
141
-------
FIGURE IV-24
GRANULAR BED FILTER INSTALLATION SCHEMATIC
FROM 2ND_
CYCLONE
TO
3RD CYCLONE
-^-PREHEAT VENT
•—t><-j -J<
1300 kPa AIR
950 kPa AIR
-•<
BLOW BACK
SHUTOFF
VALVE
WATER
KNOCKOUT
-------
TABLE IV-12. GRANULAR BED FILTER CLEANING PROGRAM
Program 1
Event Time (m:s)
0 Forward Flow Off, Reverse Flow On 0:00
1 Above Bed Pulse (1.5 sec) 0:02
2 Above Bed Pulse (1.5 sec) 0:07
3 Above Bed Pulse (1.5 sec) 0:14
4 Reverse Flow Off, Forward Flow On 0:19
5 Reset To Start Cleaning Cycle 5:19
Program 2
Event Time (m:s)
0 Forward Flow Off, Reverse Flow On 0:00
1 Above Bed Pulse (1.75 sec) 0:01
2 Above Bed Pulse (1.75 sec) 0:05
3 Above Bed Pulse (1.75 sec) 0:10
4 Above Bed Pulse (1.75 sec) 0:14
5 Above Bed Pulse (1.75 sec) 0:17
6 Forward Flow On, Reverse Flow Off (Filtration) 0:19
7 Reset To Start Cleaning Cycle 5:19
TABLE IV-13. COMPARISON OF PARTICULATE SIZE DISTRIBUTION
OF MATERIAL BEFORE AND AFTER FILTER TEST ELEMENT
Sample
Before Filter
After Filter
Test 1
Test 2
Parti cl
e SI
Vol . % Finer
5id
1.4
1.4
1.4
10%
1.7
1.7
1.7
25%
2.4
2.2
2.4
bu%
3.1
2.8
3.0
ze (ym)
Than Size
75%
4.2
3.8
4.0
90%
5.6
5.1
5.4
95%
6.4
6.4
6.2
143
-------
changed to program 2 1n Table IV-12. After 3 hours, the filter was isolated
from the miniplant and the Balston total filter which follows the granular
bed filter test vessel was replaced. This Balston sampling filter was com-
pletely destroyed by the high temperature of the gas exiting the filter
vessel. A large amount of particulate matter was found in the off gas
piping before and after the Balston total filter. This was evidence of a
large amount of particulate penetration. A second Balston filter was also
destroyed. After the removal of the second filter, the Balston filter vessel
failed and the test was terminated. The size distribution of the particulates
removed from the filter support screen is shown in Table IV-13. The average
particle size of 3 microns is almost identical to the average size of
material before the filter. Therefore it appears that very little filtra-
tion took place.
The filtration velocity that was used during these tests was higher than
any tested before on the miniplant. This happened because the AP cell which
measures the pressure drop across the measuring orifice was miscalibrated.
It was calibrated for 0-30 kPa rather than 0-30 in We. This error resulted
in actual filtration face velocities much higher than expected.
During the second Balston total filter sampling test, the pressure drop
across the filter was constant throughout the filtration cycle. This indi-
cated a complete loss of filter media. This, along with the failure of the
sampling filter vessel, caused the termination of the test.
After the run, the plenum of the test vessel was opened and approximately
7 kg of material was collected. This material was alumina bed material, which
had been blown out of the filter bed, containing less than 10% flyash.
A summary of granular bed filter test conditions and results for Run 115
is shown in Table IV-14. Even with all of the things that went wrong during
the test, some conclusions are possible. With a single element and controlled
blow backs, the maintenance of a proper AP was very difficult. The pressure
drop continued to rise even during periods when some bed material was lost.
The pattern of pressure buildup with time can be seen in Figure IV-25. Per-
haps lower face velocities would have kept the pressure drop from rising as
quickly. However, the fact remains there was a continued failure to clean
the filter properly and check the increased pressure drop even with signifi-
cant loss of filter media. The operating problems seen in this run, poor
filter cleaning and loss of filter media, are the same problems that termi*
nated the earlier program.
Unfortunately, the testing of the slip stream filter was terminated
before all operability questions could be answered. An unresolved issue was
whether the loss of filter media could have been prevented by modifying the
blow back procedure, possibly by introducing blow back air more gradually
instead of in short pulses. The effect of lowering the filter face velocity
on filtration performance was also unresolved.
144
-------
TABLE IV-14. RUN 115 FILTER TEST SUMMARY
Run Length (total)
Number of Beds
Design
Filter Medium
Bed Depth (cm)
Filter Media Particle Size (ym)
Vessel Preheat
4 Hours
1
Exxon Mark IV
Alumina
7
1410-2380
Inlet Pipe Only
Run Breakdown
Segment
Run Length (hrs)
Operating Conditions
Filter Inlet Temperature (°C)
Filter Outlet Temperature (°C)
Pressure (kPa)
Baseline AP
Clean Bed (kPa)
AP Before Blow Back (kPa)
Superficial Velocity (m/min)
Duration: Filtration (min)
Blow Back Conditions
Superficial Velocity
Duration: Blow Back
Blow Back Program
(s)
0.75
400
350
900
28
35
38
5
NA
20
1
2.25
NA
630
900
38
43
52
5
NA
20
2
NA
630
900
23
23
45
5
NA
20
2
NA = Not Available
145
-------
FIGURE IV-25
GRANULAR BED FILTER FLOW AND PRESSURE DROP VS. TIME
3.7
3.1
°°E 2.5|-
o
II I
g
«O
v-i
0
5 MIN
FLOW
TIME-
60
40
70
-o
20
146
-------
CONVENTIONAL PARTICULATE CONTROL
A series of long term (4 to 10 day) continuous tests were carried out in
which a conventional low temperature, low pressure electrostatic precipitator
(ESP) and bag house were tested on the flue gas from the miniplant. These
tests were conducted during minlplant runs 103 through 108, with the three
stages of high temperature/pressure cyclones in operation; the low tempera-
ture/pressure devices thus represented a sourth stage of particle cleanup.
The purpose of the tests was to determine if an ESP or bag house could be
used, after expansion of the flue gas through the gas turbine, to meet partl-
culate emission standards. The possibility being considered was that cyclones
may be sufficient to protect the gas turbine from excessive wear but may not
be sufficient to meet environmental standards. The tests were performed
using mobile, trailer mounted control devices operated by Acurex Corporation
for the EPA. These units received a stream of expanded, diluted miniplant
flue gas during 3 runs each for the ESP (runs 103, 104, 105) and the bag
house (runs 106, 107, 108). Due to the system configuration, the flue gas
was diluted by 50 to 70% with air used for pressure control. A sketch of
the test configuration is shown in Figure 111-10.
Results
Preliminary results with the EPA mobile ESP indicate the applicability
of conventional electrostatic precipitation for control of cyclone-cleaned
PFBC particulate emissions. Overall results from 17 days of operation (runs
103, 104 and 105) reflect 87% efficiency, corresponding to an emission level
of 8.6 ng/J (0.02 g/Nm3).
Preliminary results with the EPA mobile bag house also indicated the
applicability of conventional fabric filtration for control of cyclone-cleaned
PFBC particulate emissions. Overall results from 15 days of operation (runs
106, 107 and 108) reflect 99.3% efficiency, corresponding to an emission level
of 0.46 ng/J (0.001 g/Nm3).
A more detailed and comprehensive accounting of these tests will be
contained in a report to be published in late 1979 by Acurex.
147
-------
SECTION V
REGENERATION STUDIES
In 1976, the combined operation of the mini pi ant combustor and regenerator
sections was demonstrated. That demonstration run, reported in the previous
annual report, illustrated the operability of the system and demonstrated that
sorbent regeneration does reduce the amount of makeup sorbent required. The
next step in the regeneration study, and the primary objective of the combined
combustor-regenerator test program conducted in 1979, was to quantify the
effect of certain key variables on the performance of the combined combustor/
regenerator system. The variables studied were makeup Ca/S ratio and sorbent
recirculation rate.
EQUIPMENT AND PROCEDURES
Equipment
The equipment and materials were described in detail in the previous
two annual reports (1,2). No major changes have occurred since, and only
brief summaries are included in this report.
Air System--
The two separate air systems are burner air and supplementary air. All
air is supplied by the main air compressor. Automatic control systems, con-
sisting of control valves, flow measuring orifices, and electronic control-
lers, are used to regulate air flow. Burner air is supplied to the burner,
located beneath the fluldizing grid, in sufficient quantity to completely
burn the fuel (natural gas), Supplementary air is added about halfway up the
bed in order to create an oxidizing zone in the upper portion of the bed.
Fuel System—
The two fuel systems are burner fuel and supplementary fuel. Automatic
control systems, similar to those used for air flows, are used to regulate
the flow of natural gas. Burner fuel is supplied to the burner where it is
burned with an approximately stoichiometric amount of air. Supplementary
fuel is added directly to the regenerator column just above the fluidizing
grid in order to produce reducing gases (CO, H2).
Off Gas Handling--
Hot pressurized gases leaving the regenerator are cooled in a single
pass double pipe heat exchanger and expanded to nearly atmospheric pressure
across a control valve. Dust is removed from the gas upstream of the cooler
by a cyclone and upstream of the pressure control valve by a stainless steel
knockout vessel.
Off gases from the regenerator are sent to a Research-Cottrell scrubber
for cleanup before venting. Ammonia is injected to neutralize the scrubber
water.
148
-------
Gas Sampling System--
A slipstream of the off gas is taken downstream of the pressure reducing
valve. The gas is filtered (Balston Model 33 filter) and dried (Perma-Pure
Model PD-1000-24S self-regenerative membrane-type dryer) before entering the
gas analyzers.
Fluidizing Grid--
The fluidizing grid has 88 holes of 3.6 mm (9/64 in) diameter for passage
of the fluidizing gases (from the burner located beneath grid) and 14 water
cooling channels of 4.8 mm (3/16 in) diameter. The cooling water flow is
controlled independently through six separate cooling zones. The 14 channels
were separated into groups of 3, 2, 2, 2, 2, and 3, each group or zone having
its own water supply.
Burner--
No changes were made to the regenerator burner since the last report.
This unit is identical to that used in the miniplant combustor and is described
in a previous annual report (2).
Sorbent Transfer System--
The solids transfer system used to circulate solids between the combustor
and regenerator is shown schematically in Figure V-l. Pressure in the regen-
erator is maintained slightly higher than that in the combustor. Solids in
the regenerator-to-combustor transfer line move into the combustor when a
pulse of nitrogen is applied to the lower end of the transfer line. The flow
rate of solids is controlled by adjusting the frequency, duration, and inten-
sity of the pulse. Two slide valves are used in the combustor-to-regenerator
transfer line in order to prevent back flow of gas from the regenerator up the
line. These automatic valves trap solids in the piping between them. Solids
are discharged into the regenerator when the bottom valve is opened. The two
solids' take-off plugs shown in Figure V-l are inserted into the ports during
startup to prevent solids from entering the lines. Plugging can occur if the
solids become wet due to water condensation during startup. The manual slide
valve in the regenerator-to-combustor line is also closed during startup and
during upsets.
The components of the sorbent transfer system (valves, expansion joints,
etc.) are described in detail in a previous annual report (2). The transfer*
lines themselves were fabricated from 6 inch schedule 40 carbon steel pipe
and refractory lined to an inside diameter of 7.6 cm. The sloping portions
of the lines were sleeved with 2-1/2 inch Schedule 10 316 stainless steel
pipe, which has an inside diameter of 6.7 cm.
Operating 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
149
-------
FIGURE V-l
MINIPLANT SOLIDS TRANSFER SYSTEM (SCHEMATIC)
REGENERATOR
COMBUSTOR
SOLIDS TAKE
OFF PLUG
AUTO
SLIDE
VALVE
NITROGEN
PULSE
150
-------
1n order to prevent solids from backing up Into the lines. The slide valves
are also cycled occasionally 1n order to dislodge any sol Ids that may have
passed by the takeoff plugs. Pressure 1n the regenerator 1s set slightly
above that 1n the combustor (usually 2.5-10 kPa) and the beds are heated to
near operating temperatures. Transfer of sol Ids Is started by pulling the
plugs out of the takeoff ports. The transfer rate 1s controlled by setting
the cycle time of the slide valves 1n the combustor-to-regenerator line.
Starting Reducing Conditions--
The regenerator 1s operated under net oxidizing conditions during heatup.
Under these conditions, no regeneration of sulfated sorbent occurs. When the*
combustor and regenerator bed temperatures are uniform and close to the
desired operating temperatures, the switchover to reducing conditions 1n the
regenerator 1s made. This 1s accomplished by increasing flow of supplementary
air to the required value and then increasing the flow of supplementary fuel
Supplementary air flow 1s 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, supple-
mentary air, and supplementary fuel are all adjusted to yield the desired bed
temperature. Oxygen and CO concentrations 1n 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 1n the oxidizing zone.
Shutdown--
It is important during shutdown, to empty the transfer lines of sol Ids;
otherwise plugging of the lines may occur when the unit is restarted. Hence
the first step 1s to shut the plugs 1n the sol Ids takeoff ports, thereby '
preventing solids from entering the lines. Cycling of the slide valves 1n
the combustor-to-regenerator Hne and pulsing of nitrogen in the regenerator-
to-combustor line is continued until the mlniplant is shut down. This assures
that the lines are emptied.
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.
EXPERIMENTAL RESULTS AND DISCUSSION
Shakedown
After some cold flow tests, a hot test, run 101, was made to insure that
the solids transfer lines and control systems were operating smoothly and to
determine the operating pressure for the test program (7 atm). Shakedown went
151
-------
very smoothly with the combined units operating as well as they had during the
first 100 hour demonstration run In 1976.
Operating Performance
During the regenerator test series (runs 102, 103 and 105) over 400 hours
of operation were logged with solids circulating between the regenerator and
combustor, 370 of these hours were with sorbent regeneration. Including the
100 hour demonstration run in 1976, the combined combustor-regenerator system
was operated for more than 500 hours. Overall, operations of the two units
was smooth. Combustor problems were limited to infrequent interruptions in
coal feed due to coal flow problems which were quickly remedied. Problems in
regenerator operation, which were basically limited to two areas, required
maintenance during and between runs but did not result in forced shutdown.
The two problem areas were off gas handling and solids transfer,
Regenerator off gas passes through a cyclone before being cooled and
pressure reduced. The cyclone was operational during run 102, but during the
second day of run 103, the downcomer became plugged with fines, probably a
result of wall buildup from the previous run coupled with moist particulates
from the startup of 103. A large, semi-batch granular bed sand filter was
put on stream to replace the cyclone rather than shutting down to clear the
cyclone. The high dust loading 1n the regenerator off gas prior to the
installation of the sand filter and during the filter blow back intervals
caused appreciable erosion of the pressure control valve stem and housing.
The major problem with the solids transfer system was a blockage in the
combustor to regenerator line that prevented sol Ids movement at the start of
run 103. It was found that a piece of refractory wall in the 45° angle sec-
tion just above the pulse pot had fallen into the opening. The section was
removed, the refractory recast, and the run resumed.
The solids transfer system also proved to be a limiting factor on the
solids circulation rate. As solids are transferred from the combustor to
regenerator, they are trapped and held between the two automatic knife valves.
While the solids are trapped between cycles, they cool. If the transfer leg
temperature drops too low, water vapor may condense making the sol Ids mushy
and preventing solids flow. The lowest solids circulation rate which seemed
safe for the existing system was 10-12 kg/hr.
During the regeneration test series, additional testing was done by
outside contractors. Acurex Company tested a mobile electrostatic precipita-
tor using a slip stream of combustor flue gas during runs 103 and 105. GCA
Technology Company performed Level I comprehensive analysis tests on both
the combustor and regenerator during run 105. These test programs are
discussed elsewhere in this report.
Run Conditions and Summaries
Run summaries giving operating conditions for both the combustor and
regenerator are presented in Tables V-l and V-2. During the 3 runs, six
steady state conditions were reached. Temperatures, pressures, superficial
152
-------
in
CO
TABLE V-l. MINIPLANT REGENERATOR RUN SUMMARY
Regenerator Operating Conditions
Run No. 102 103.0 103.1 103.2 103.3 105
Run Length, (reducing) hrs 97.5 54 48 20 56.5 80.7
Pressure, kPa 710 710 710 710 710 705
Air Flow Rate, m3/min 2.86 3.08 2.72 2.52 2.31 2.3
Fuel Flow Rate, m3/m1n 0.26 0.30 0.27 0.23 0.20 0.23
Average Bed Temp - Reducing Zone 1025 1004 1003 1008 974 1010
°C - Oxidizing Zone 1053 1024 1021 1025 985 1031
Superficial Gas Velocity, m/s
- Reducing Zone 0.64 0.78 0.67 0.50 0.58 0.60
- Oxidizing Zone 0.91 1.08 0.98 0.87 0.78 0.77
Expanded Bed Height, m 1.65 1.37 1.38 1.38 1.37 1.85
Air to Fuel Ratio, 4>
- Reducing Zone 1.36 1.31 1.37 1,35 1.32 1.27
- Oxidizing Zone 0.92 0,95 0.94 0.91 0.96 1.02
Solids Recirculation Rate, kg/hr 88.5 22.7 22.7 12.2 12.2 15.4
Average SOg Emissions, % 0.32 0.40 0.46 0.25 0.16 0.51
Average CO Emissions, ppm 1610 934 484 231 996 1438
Average C0£ Emissions, % 9.8 9.44 9.42 8.38 8.69 10.9
Average 02 Emissions, % 0.17 0.21 0.31 0.42 0.18 0.26
-------
TABLE V-2. MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Combustor
Operating Conditions
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m3/min
Temperature Gradient, °C/m
Average 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 Ratio-Equivalent**
Excess Air, %
Sorbent
Coal
102
3/13-17/79
107
700
12.4
-15,2
908
1.43
2.06
0.96
2.17
69.5
1.5
54
20.3
GL
CH
103.0
3/29-4/1/79
81
700
13.3
-12.2
913
1.52
2.11
2.65
76.6
1.35
12
16.8
GL
CH
103.1
1/4-3/79
49
700
13.4
-14.0
901
1.52
__
._
2.83
77.7
0.68
11
17.7
GL
CH
103.2
4/3-4/79
22
700
14.1
-12.3
901
1.59
__
__
1.97
79.4
0.68
6.1
13.0
GL
CH
103.3
4/4-6/79
59
700
14.4
-17.7
902
1.62
-_
1.42
2.24
79.9
0.93
6.4
18.9
GL
CH
105
4/30-5/5/79
99
700
14.0
-55.9
894
1.46
2.23
1.71
3.10
77
1 .29
8.3
45.8
GL
CH
Flue Gas Emissions - Combustor
S02, ppm
NOX, ppm
CO, ppm
C02, %
021 *
Results - Combustor
Retention, %
Ca Sulfation, %
Lb SO?/M BTU
Lb NOx/M BTU
GL = Grove Limestone
CH = Champion Coal
27
126
90
14.0
4.0
97.1
5.1
0.06
0.19
28
77
71
13.7
3.3
41
56
62
14.1
3.4
22
65
64
14.1
2.6
77
55
56
13.7
3.6
97.5
18.2
0.06
0.12
96.4
23.7
0.09
0.09
98.0
22.2
0.05
0.10
93.0
28.0
0.17
0.09
70
59
77
12.4
4.1
93.6
23.4
0.15
0.09
**
Fresh sorbent feed only.
Includes recirculating sorbent from regenerator, as well as fresh
sorbent feed; assumes 100% regeneration of the recirculated sorbent,
-------
gas velocities, bed heights, coal feed rates, sorbent type, coal type, and
reducing atmosphere composition were held fairly constant throughout the runs.
Solids redrculation rate and Ca/S ratio fed to the combustor were the vari-
ables studied for their effects on combustor and regenerator $03 emissions,
sulfur retention, and makeup sorbent rate needed to meet EPA emission stand-
ards.
Results
In all cases, sulfur retention exceeded 90% with makeup Ca/S ratios rang-
ing between 0.68 and 1.5, compared to a Ca/S ratio of 3 to 4 which is required
to achieve the same results for once through operation with limestone. The
significant reduction in fresh (makeup) sorbent requirements is due to the
regenerated sorbent which circulates from the regenerator to the combustor.
The
Ca/s
The
r /c - 0.32 x SCA x (SFR x SCR + 1.79 RSR x (1 - SSL) x (1 + SCR}}
Ca/SEQ - CS x SFR
where SCA = % CaC03 in sorbent, for limestone ~100%
SFR = coal + fresh sorbent feed rate, kg/hr
SCR = sorbent to coal ratio, kg/kg
RSR = recirculating sorbent rate, kg/hr
SSL - sorbent sulfation level of regenerated sorbent, wt. fraction
CS = coal sulfur content, %
Comparison of the fresh sorbent Ca/S ratio with the Ca/S equivalent shows that
the rate of circulating regenerated sorbent is much larger than the fresh sor-
bent rate and therefore, probably plays a larger role in controlling S0£ emis-
sions.
Combustor SOg Retention—
As mentioned previously, S0£ retention exceeded 90% for the runs made,
covering a range of Ca/S ratios. An effective method to evaluate sorbent
regeneration is to compare SOg retention as a function of CB/SEQ for the
regeneration runs to the data obtained for once through operation. Figure
V-2 shows such a comparison. All SOg retentions were adjusted to a 2 second
gas residence time using the first order rate expression.
Figure V-2 shows that the data points generated by the regeneration runs
fall around the curve for once through operation with limestone, The values
for Ca/Sfcq are based on an assumption of 100% sorbent regeneration due to
difficulty in obtaining regenerated bed samples. If the samples were not
fully regenerated, i.e., the assumption was invalid, the data points would
shift to the left toward a lower CB/SEQ. For example, one regenerator bed
sample was obtained during run 105 when a pressure upset caused some regen-
erator bed to be blown out of the column into the cyclone. The sample was
analyzed and found to contain 14.6% CaS04 by weight. The equivalent Ca/S
molar ratio dropped from 8.3 to 6.0 when the sorbent sulfation level was
taken into account in the Ca/Sgq expression.
155
-------
FIGURE V-2
SO2 RETENTION ADJUSTED TO 2 s RESIDENCE TIME VS Ca/S RATIO
(N
z
o
I—
Z
UJ
LU
oe.
CN
100
90
80
70
60
50
40
'LIMESTONE
DOLOMITE
ONCE-THROUGH OPERATION
• WITH LIMESTONE
• REGENERATION RUNS
0 1
3456789
Ca/S MOLAR RATIO*
10 11
* FRESH SORBENT ONLY FOR ONCE-THROUGH RUNS
FRESH PLUS RECIRCULATED SORBENT FOR REGENERATION
RUNS
156
-------
The scatter in the data is not appreciable and from this treatment of
the data it can be concluded that the regenerated sorbent behaves as if it
were fresh sorbent.
Sorbent Activity Loss-
Previous work done in TGAs and batch units indicated a decline in sorbent
activity as the sorbent was cycled between sulfations and regenerations (13),
In a continuous unit such as the miniplant, the sorbent does cycle a number of
times but the cycling is not comparable with batch cycling experiments. Sorbent
lost by attrition must be replaced with makeup (fresh) sorbent. The fresh
sorbent feed serves to maintain the age distribution of the sorbent in the
system, and therefore the activity of the bed, at a steady state value. From
the outset of a run, the average age of the sorbent bed changes exponentially
and the steady state bed age distribution is approached asymptotically. The
rate at which the bed steady state is reached is determined by the system time
constant, t, where t = bed inventory/feed rate. (Based upon the range of
values for t in these miniplant runs, the exponential expression would suggest
that the combustor bed should approach within 90% of its steady state level in
about 90 hours of operation.) Therefore, unless the sorbent suffers rapid
and severe deactlvation, sorbent deactivation would not be obvious in a con-
tinuous system once a steady state bed age distribution has been reached.
Figures V-3, v-4 and V-5 illustrate the change in average bed age with
time as well as the combustor S02 emissions for the regeneration test runs.
When the regenerator was brought into reducing conditions and regenerated
sorbent began to enter the combustor at high rates, combustor SOg emissions
dropped markedly. As seen in the figures, the combustor S02 emissions leveled
out at a steady state value a long time before the bed reached its steady
state age distribution. If the sorbent experienced deactlvation, this would
have been detectable as an increase in combustor SOp emissions during the
initial line-out period.
Based on the results exhibited in Figures V-3 through V-5, it can be
concluded that during the residence time of a sorbent particle, minimal or no
sorbent deactlvation occurs or if deactlvation does occur, it does not greatly
alter the overall activity of the bed.
The results of the regeneration test series which are reported here con-
tradict the findings of the regenerator demonstration run reported previously
(1). In the earlier run, after about 50 hours combustor S02 emissions began
to increase rapidly for a period of 50 hours, then leveled out at 550 ppm.
The increase in emissions was assumed to have been caused by a gradual decline
in the activity of the regenerated sorbent. This hypothesis 1s not supported
by the most recent miniplant regeneration studies where there was no sign of
sorbent deactlvation. It 1s also not consistent with the analysis of the
recent runs described above. This analysis points out that 1f deactlvation
occurs, its effect would only be seen during the initial part of the run and
not as far into the run as 50 hours. The increase in combustor S02 emissions
during the previous run may have been a result of operational difficulties
rather than sorbent deactlvation. If sorbent regeneration or solids transfer
was inhibited for any reason, such as a bed agglomerate, sulfur retention in
the combustor would be adversely affected.
157
-------
en
TO
I
Q_
Z
O
on
to
cs
Qi
O
fc
Z)
CO
FIGURE V-3
COMBUSTOR BED AGE AND SCX EMISSIONS VS. TIME (RUN 102)
0 ' 30
START
REGENERATION
SO EMISSIONS
60
90 120 150
HOURS INTO RUN
180
210
100
80
60
40
20
o
>
O
>
O
m
O
m
CO
C
;H
O
z
-------
FIGURE V-4
COMBUSTOR BED AGE AND SO2 EMISSIONS VS. TIME (RUN 103)
oo
Z
o
OO
10
UJ
O
CO
Z)
CD
5
O
u
REGENERATION
90 120 150
HOURS INTO RUN
03
rn
O
o
m
^
ff-
-T3
?0
O
o
O
O
m
TO
55
c
O
z
-------
en
o
O
CQ
O
FIGURE V-5
COMBUSTOR BED AGE AND SO2 EMISSIONS VS. TIME (RUN 105)
OO
z
o
8 100
CM
SO2 EMISSIONS
100
80
60
40
20
oo
m
D
O
-o
-o
•yo
o
3
5
S
oo
O
m
oo
C
O
z
0 I 30
START
REGENERATION
90 120 150
HOURS INTO RUN
180
210
-------
Sorbent Elutriation Losses--
During the test series, the minimum Ca/S ratio fed to the combustor was
set by sorbent attrition rather than by sorbent desulfurization activity. To
maintain bed levels in the combustor and regenerator, bed lost through attri-
tion was replaced by the incoming makeup sorbent. Sorbent losses represent
elutriation of the fines formed by attrition.
Since sorbent elutriation was the factor limiting the Ca/S ratio, it is
important to compare elutriation losses for the once through process and the
regenerative process to see the effect regeneration has on attrition and
subsequent elutriation. Sorbent elutriation losses for the once through
process (miniplant combustor only) for both Grove limestone and Pfizer dolomite
were presented in a previous report (2). Sorbent losses for the regeneration
test series are presented in Table V-3. Losses for the once through process
using both uncalcined and precalcined limestone are included for comparison.
The sorbent loss calculations are based on the calcium contents of the
materials entering and leaving the system and are essentially calcium losses.
Combustor sorbent losses expressed as the Ca/S ratio needed to offset the
loss, and as percent of fresh sorbent, are extremely high. However, combustor
elutriation losses expressed as a percent of bed inventory for the regenera-
tion runs do not differ greatly from the values for once through operation.
With the combined combustor-regenerator unit, the regenerator off gas provides
an additional source of sorbent elutriation, Regenerator sorbent losses
expressed as a percent of bed inventory tend to be higher than percent of bed
inventory combustor losses. A possible explanation for higher regenerator
elutriation losses is that the regenerated sorbent (CaO) is less attrition
resistant than sulfated sorbent (CaSCty). The sorbent is also subject to
thermal shock as it enters the regenerator. This may also contribute to
higher attrition rates.
Combined losses from both the combustor and regenerator reflect the fact
that the miniplant, under the conditions of the regeneration test series, was
operating in the range of the minimum Ca/S ratio needed to maintain bed levels.
In a number of the regenerative cases, elutriation losses exceeded the feed
rate of fresh sorbent. Regeneration therefore can reduce the Ca/S ratio
needed to meet EPA emission standards but not beyond the point were elutria-
tion losses are larger than the makeup sorbent ratio. To specify the minimum
Ca/S ratio necessary to provide 90% or less S02 removal in the miniplant, a
high sulfur coal would have to be burnt so that bed activity, rather than
attrition, would be the limiting factor at those lower percentage S02
removals.
S02 Content in Regenerator Off Gas--
Miniplant regenerator S02 emissions are not limited by thermodynamics.
For the operating conditions used in the test series, thermodynamics predicts
an equilibrium S02 concentration of -"3%. In all runs, S0£ emissions were
much lower than the equilibrium S02 concentration. Average SO^ emissions
ranged from 0.16 to 0.51%, equivalent to 6 to 16% of the equilibrium concen-
tration (see Table V-4),
161
-------
TABLE V-3. SORBENT ELUTRIATION LOSSES
ro
Sorbent
Run No.
UncaUlnedW
GL
Precalcined ' '
GL
102
GL
103.0
GL
103.1
GL
103.2
GL
103.3
GL
105
Superficial
Gas
Velocity
m/s
i 1.4-2.2
) 1.6-2.5
1.43
1.52
1.52
1.59
1.62
1.46
Combustor
Feed Rate
Makeup
Ca/S
1.5-2.8
2.5-4.0
1.5
1.35
0.68
0.68
0.93
1.29
Ca/S Loss
Equivalent
0.2
0.2
2.4
0.80
0.58
0.85
0.45
0.55
Fresh
Feed
12
8
160
58
86
126
48
43
Losses
Vol. %
Bed/Hr
1.1
0.8
3.2
1.3
0.9
2.1
1.0
0.9
Total
Feed
12
8
6
11
9
23
11
11
Reg.
Losses
Vol . %
Bed/Hr
—
1.6
6.2
1.6
4.3
1.6
0.7
Combined Losses
Ca/S Loss
Equivalent
0.2
0.2
2.8
1.5
0.78
1.4
0.63
0.67
Fresh
Feed
12
8
185
113
114
200
68
51
Vol. %
Bed/Hr
1.1
0.8
2.8
2.1
1.1
2.6
1.1
0.8
GL = Grove Limestone
(1) Once-through operation.
-------
(71
CO
TABLE V-4. COMPOSITION OF REGENERATOR OFF GAS
Equilibrium
Avg. S02 S02 Cone. @ S02 Emissions
Run No. Emissions (%) Conditions (%) (% of Equilibrium Cone.)
102
103.0
103.1
103.2
103.3
105
105 Peak
0.32
0.40
0.46
0.25
0.16
0.51
1.0
3.4
3.0
3.0
3.1
2.6
3.2
3.5
9
13
15
8
6
16
29
-------
The low S02 concentrations are caused by limitations imposed by the size
of the minlplant regenerator and combustor. S02 concentration in the regen-
erator off gas 1s determined by mass and energy balance constraints rather
than chemical equilibria or reaction rates. Heat losses from the regenerator
are high and require the addition of more hot gas than is needed to satisfy
the requirements of the regeneration chemical reactions. Also, the amount of
CaS04 (sulfated limestone) fed to the regenerator 1s set by the sulfur content
of the coal and the size of the combustor. The regenerator is actually over-
sized for the CaS04 rates possible in the recent test series. Reducing the
size of the regenerator or increasing the size of the combustor is not prac-
tical and coals with a higher sulfur content were not available for this
program.
In addition to these constraints, the superficial gas velocity 1n the
regenerator also had to be high enough to exceed the minimum fluidization
velocity and promote vigorous sol Ids mixing. Operating pressure had to be
lowered from 900 kPa to 700 kPa in order to achieve the proper superficial
velocity at the desired operating temperature without diluting the off gas
even further.
Therefore, the SOg levels measured 1n this test program are set by the
limitations imposed by the minlplant system and do not represent typical
results from a larger facility which would not have the same limitations.
The larger facility would be designed to allow the S02 levels to approach
those predicted at chemical equilibrium.
The last portion of run 105 was a test to determine the maximum $03 con-
centration attainable in the regenerator off gas and to compare this with
thermodynamic predictions. The test was carried out by establishing oxidizing
conditions 1n the regenerator to stop further sorbent regeneration and allow
the sulfation level of the circulating sorbent to Increase. It was Intended
to do this for a period of time until the combustor $03 emissions increased
to a fairly high level. This would indicate the sorbent was highly sulfated.
When regeneration conditions were re-established, the initial S02 concentra-
tion from the regenerator would approach an equilibrium concentration.
After 6 hours of oxidizing conditions, SO? emissions from the combustor
had not increased, Indicating that before oxidizing conditions had been
established, the sorbent had been regenerated to a high degree and was still
very active. Due to time limitations, reducing conditions had to be established
1n the regenerator following the 6 hour oxidizing period but before combus-
tor emissions had shown any increase. Regenerator and combustor S02 emissions
following the start of reducing conditions are plotted as a function of time
in Figure V-6. The cyclic nature of the regenerator S02 emissions 1s due to
sorbent transfer from the combustor which cools the regenerator bed and
decreases the regeneration reaction. The Initial increase in combustor S02
emissions was probably caused by regenerator gas passing into the combustor
since the regenerator operates at a slightly higher pressure to facilitate
solids transfer. Regenerator S02 emissions seemed to peak and level out at
1.0%. At the existing conditions, thermodynamics predicts an equilibrium con-
centration of 3.5%. Therefore the peak emission was only 29% of the equilib-
rium value (see Table V-4).
164
-------
en
en
z
cs
ex.
o
LLJ
z
LLJ
O
LLJ
Qi
FIGURE V-6
REGENERATOR AND COMBUSTOR SO2 EMISSIONS
RUN 105
0
OXIDIZING
60 120
MINUTES INTO REDUCING CONDITIONS
O
O
OO
C
CO
o
•70
O
co
°2
O
Z
CO
-------
Thermodynamics was obviously not the limiting factor in determining regen-
erator off gas SOg concentrations. The Champion coal used in the test series
had a low sulfur content (< 2%) which led to a low sulfur load on the unit,
thereby limiting regenerator SOg emissions due to mass and heat balance con-
siderations as discussed previously.
Mass Balances--
Table V-5 summarizes steady state sulfur mass balances for the regenera-
tor runs. Also included are total mass balances for each run. Total mass
balance closure was 99+% in all cases. Sulfur balances for three of the runs
(102, 103.1, and 105) were very good, just slightly over 100%. Sulfur recovery
for run 103.0 was 120%, indicating that the initial sulfur inventory (from the
initial bed charge of sul fated sorbent) was being depleted by regeneration and
was affecting the mass balance.
The last two segments of run 103 (103.2 and 103.3) had low sulfur recov-
eries, 74% and 48%, respectively. Since the total mass balances were 100%,
the cause of the deficient sulfur balances is not obvious, A possible explana-
tion is that the regeneration process was not keeping up with the sulfation
process, increasing the sulfur inventory of the bed. Combustor bed probe
samples exhibit a slight increase in sulfur level toward the end of the run,
supporting this possibility. Another explanation would be a sudden change in
sulfur load (coal sulfur content) which would not be accounted for in the
sulfur balance. During run 103, an average coal sulfur content of 1.61% was
used. In addition, two coal samples were taken during run 103.3 to determine
the consistency of the coal sulfur content. The samples were taken 15 hours
apart; the first had a sulfur content of 1.84%, the second dropped to 1.37%.
If this drastic change in coal sulfur content was not just transient, the sul-
fur mass balance would be in error.
Conclusions
Continuous operation of the miniplant provided a realistic way of measur-
ing the potential benefits of a regenerative system compared to a once
through system. Regeneration can reduce the makeup sorbent requirement needed
to meet EPA emission standards by a factor of 3 to 4 but not below the makeup
sorbent ratio needed to compensate for attrition losses and thus maintain bed
inventory. Sorbent activity loss, if it occurs, was not apparent from the
performance of the combined combustor-regenerator system on the basis of
retention or combustor SOg emissions. Results of this study may have been
affected by the fact that all work fell in the high SOg retention range.
If a higher sulfur coal (~4%) would have been available for the test
series, perhaps more conclusive results would have been obtained. High sulfur
coal would have allowed lower makeup Ca/S ratios to be tested without attri-
tion being the limiting factor. The increased sulfur load would have also
improved the SOg concentration in the regenerator off gas, an important factor
in downstream sulfur recovery.
166
-------
TABLE V-5. MINIPLANT REGENERATOR MASS BALANCES
en
--J
Run No.
102
103.0
103.1
103.2
103,3
105
Sulfur Mass Balance
Sulfur Input
Coal
TOTAL
Sulfur Output
Regenerator Off Gas
Combustor Off Gas
Combustor Overhead Solids
2° Cyclone
3° Cyclone
Regenerator Overhead Solids
% Sulfur Recovery
% of Sulfur Entering
100
100
69.4
2.9
22.8
2.9
3.8
101.7
100
100
81 .9
2.4
19.5
3.2
13.8
120.2
100
100
82.3
3.2
13.6
2.4
4.0
105.9
100
100
40.7
1.6
19.6
3.1
9.4
74.3
100
100
24.1
7
10.9
2.3
3.9
47.7
100
100
78.4
6.3
13.5
1.6
2.4
101.4
Total Mass Balance
99.88%
99.86%
99.89%
99.79%
99.87%
100.6%
-------
SECTION VI
COMPREHENSIVE ANALYSIS OF EMISSIONS
LEVEL I AND II COMPREHENSIVE ANALYSIS TESTS
The U.S. Environmental Protection Agency has developed a phased approach
to assess the environmental impact of solid, liquid and gaseous emissions from
a process (26). This approach includes sampling and analysis for a wide range
of organic and inorganic species. The first phase of the assessment (Level I)
is intended to provide preliminary environmental assessment data and identify
principal problem areas. The objective of the second phase (Level II) is to
obtain more detailed and accurate data than is available from a Level I study.
Level II studies are intended to identify and quantify specific compounds
whose presence could be inferred from the results of Level I tests. Level II
studies may also be initiated on the basis of results of the biotests
employed as part of Level I (27). Exxon has participated in Level I and
Level II testing programs using the miniplant facility. This work was done
in cooperation with two other EPA Contracts, Battelle Columbus Laboratories
and GCA/Technology Division.
Level I Emission Measurements
Level I comprehensive emission measurements from the PFBC miniplant unit
without the regenerator unit operating were conducted in cooperation with
Battelle Columbus Laboratories during run series 50 in early 1977 (14). These
results will be reported in detail in a separate report prepared by Battelle
and Exxon. In this test series, an eastern (Champion) coal was burned with
Pfizer dolomite sorbent. In a subsequent test (run 69), made with Illinois
coal and Pfizer dolomite, samples of solid materials were obtained and sent
to Battelle for analysis. The samples were analyzed for inorganic elements
using spark source mass spectroscopy (SSMS), the same method used on the
solids samples from run series 50. The purpose of this supplementary test
program was to provide a data base for Illinois coal use, to add to that
already established for Champion coal. The data will be incorporated by
Battelle in the interpretation of the Level I results. The results of the
analyses are reported in Appendix I.
A second series of Level I tests was conducted during run 105 on the
miniplant with the regenerator also in operation. Champion coal and Grove
limestone were used in run 105 which was conducted in cooperation with GCA/
Technology Division. The principal effect of the GCA program was the collec-
tion of samples from the combustor flue gas and the regenerator flue gas
using the Source Assessment Sampling System (SASS). Two SASS tests were con-
ducted on both the combustor flue gas after expansion, diluted with air used
for pressure control, and on the undiluted regenerator off gas. Two SASS
train tests were also conducted on a filtered, undiluted slip stream of the
combustor flue gas. Other gas samples were taken by GCA for analysis of $03,
HC1, nitrogen species and Ci to Cj hydrocarbons. Gas samples were also
collected by Exxon personnel for analysis of the volatile sulfur compounds.
168
-------
In support of the comprehensive analysis, solids samples and scrubber slurry
samples were also collected by Exxon. The details of the sampling analysis
effort were presented In a test plan prepared by GCA prior to sampling (15).
GCA is responsible for analysis of the samples, and will report the results of
the Level I tests from run 105.
Level II Emission Measurements
An evaluation of the survey data of run 50 series by Battelle Columbus
Laboratory (14) indicated the need for a Level II characterization of the
emissions from the miniplant unit. A series of Level II tests was conducted
in cooperation with GCA during run 107 on the miniplant without the regenera-
tor unit operating. Operating conditions were similar to those used in the
run 50 series. Champion coal was burned with Pfizer dolomite sorbent.
The gas and solid sampling techniques were the same as in run 105. A
total of 5 SASS train tests were completed by GCA. Three SASS train tests
were completed on the combustor flue gas after expansion, diluted with pres-
sure control air. Particulates from the first SASS test on the diluted
combustor flue gas were collected at about 100°C instead of 200°C, the normal
sampling temperature, to check for organic deposition on the particulates at
lower temperature. Two other SASS train tests were completed on a filtered
undiluted slip stream of the combustor flue gas for analysis of gaseous
emissions. Other gas samples were taken by GCA and Exxon personnel for
analysis of $03, C] to Cj hydrocarbons and other sulfur compounds. Parallel
solid samples were also collected by Exxon personnel. Details of the sampling
analysis effort are included in a test plan prepared by GCA prior to sampling
(16). The Level II analytical results of run 107 will be reported by GCA.
PRESENCE OF Mg3(CaS04)4 IN 3RD CYCLONE FLYASH
As was reported by Argonne (17), the formation of Mg3Ca(S04)4 or MgSO/i
should be avoided when dolomite is used in PFBC if the sulfated solid waste*
is to be disposed of as landfill. In the presence of water, Mg3Ca(S04)4 1s
unstable, decomposing to MgS04 which is soluble in water, and CaS04 which is
relatively insoluble in water. The chemistry related to the formation of the
binary salt Mg3Ca(S04)4 is not known to date. However, it had been reported
(17,18,19,20) that Mg3Ca(S04)4 can form during the sulfation of dolomite under
conditions similar to those in PFBC. Therefore, a short study was undertaken
to determine if the binary salt is formed in the waste solids produced in the
miniplant. Samples from run 67 of the bed overflow solids, partlculate
captured in the second and third stage cyclones and solids found in the tur-
bine test section downstream of the third cyclone were analyzed by X-ray dif-
fraction. In run 67, the flue gas contained about 700 ppm S02 and 3 to 5%
02. At the flue gas pressure, 9 atm, MgS04 was thermodynamically possible at
temperatures below about 840°C. No evidence of the binary salt or MgS04 was
found in the second cyclone particulates even though the temperature was
between 770 and 800°C. However, third cyclone material and turbine test
section deposits did contain Mg3Ca(S04)4 but no MgS04. The temperature in
the third cyclone was 700 to 820°C, the turbine test section 660 to 820°C.
The absence of Mg3Ca(S04)4 in the second cyclone material but its presence
169
-------
in the third cyclone and turbine test section material could be due to a
residence time effect. The residence time in the piping between the two
cyclones is greater than the residence time between the combustor and the
second cyclone. The temperature is also lower in the third cyclone and tur-
bine section thus favoring the formation to a greater extent.
The extent of magnesium conversion in the third cyclone and turbine sec-
tion solids was estimated by a sulfate balance to be from 30 to 100%. How-
ever, the 700 ppm $03 level during run 67 was higher than the S02 levels which
will be experienced under the New Source Performance Standard (1.2 Ib SOg/lO^
BTU or less, depending upon the percentage reduction required in a particular
case). Furthermore, the residence time of the flyash in the piping from a
PFBC combustor to an efficient third stage particulate removal device should
be shorter than in the miniplant and the gas temperatures higher. Therefore,
formation of soluble Mg303(804)4 will be less under higher S02 retention
conditions, shorter residence times and higher temperatures between the com-
bustor and the particulate removal devices. Additional tests should be con-
ducted to determine the extent of the binary salt formation under more
realistic conditions.
170
-------
SECTION VII
BENCH UNIT STUDIES
Programs were carried out 1n both the bench combustor and regenerator
sections. The combustor had been modified, as described in the previous
report (1), to permit continuous solids feeding and removal. After an initial
check out period, the combustor was used to evaluate three NOX control methods,
two stage combustion, NH3 Injection and simulated flue gas redrculation. The
regenerator was used 1n a series of tests using sulfated sorbent produced in
the miniplant, studying the use of natural gas and coal to fuel the regenera-
tion section. The following sections describe the results of the bench unit
combustion and regeneration tests.
COMBUSTION STUDIES
The bench combustion unit consists of a refractory lined combustor vessel
which normally operates at temperatures of 840 to 950°C, pressures of five to
eight atm, superficial velocities of 1 to 2 m/s and coal feed rates of 1 to
12 kg/hr.
Equipment
The unit was described in detail in previous reports (1,2,3). Additional
modifications were made to permit the NOX control studies to be carried out.
A brief description of the combustor section and the recent modifications are
described below.
Combustor Vessel —
The combustor (shown 1n Figure VII-1) consists of four sections of 25 cm
diameter standard wall carbon steel pipe, lined with Grefco 75-28 refractory,
to an inside diameter of 11.4 cm. The height of the vessel above the fluid-
Izing grid is about 4.9 m. Below the grid is a 61 cm burner section lined
with Grefco Bubblite refractory. The preheat burner employs a mixture of
natural gas and air to heat (and fluidize) the bed up to temperatures suf-
ficiently high to ignite the coal.
Three sets of vertically mounted 316 SS water cooled coils are located
inside the combustor to assist in temperature control. These coils remove
50-60% of the heat of combustion.
Two flanged Inlet lines are welded into the final section above the grid
at angles of 60° to the horizontal. These permit the charging of fresh sor-
bent and the return of the solids from the primary cyclone.
The second and third sections of the combustor have solids draw off lines
Inclined at 60° to the horizontal; these are used to control bed height and
volume. Solids drawn off through them are discharged to lock hoppers. The
lower draw off line is located 1.09 m above the grid and provides a bed volume
of 0.0111 m3. It was used for all tests described in this report. The higher
line is located 1.85 m above the grid and provides a bed volume of 0.0189 m3.
171
-------
FIGURE Vll-l
SCHEMATIC OF BATCH COMBUSTION UNIT
COMBUSTOR
SHELL
SORBENT
FEED
HOPPER X/
Refractory
Lining
SUPPLEMENTARY AIR1
TRANSPORT
AIR
V HOPPER
COOLING WATER OUT
OFF GAS PRECOOLER
COOLING WATER IN
OFF GAS COOLER
COOLING WATER OUT
TO SCRUBBER
TO ANALYTICAL
TRAIN
BACK PRESSURE
SOLIDS REGULATOR
OVERFLOW
BURNER
-------
The use of these draw off facilities and variations In operating para-
meters permit operating over a wide range of gas residence times.
Ancillary Systems--
Coal is fed to the combustor by a means of a pneumatic transport coal
injector capable of feeding up to 14 kg/hr for periods of up to 8 hours
before refilling. Sorbent is charged through a separate system. The system
consists of two cycling valves which trap a small amount of sorbent between
them and discharge the contained sorbent into the combustor intermittently.
Flue gas leaving the combustor passes through two cyclones, a cooler and
filter before expansion across a control valve.
Combustor System Modifications—
The combustor was modified to permit the study of the effects of staged
combustion, ammonia injection and simulated flue gas recirculation on NOX
emission levels (see Figure VII-1). Modifications were made to introduce
supplementary combustion air (for two stage combustion) and ammonia directly
into the fluidized bed, or above the bed, during combustor operation. Nitrogen
(for simulated flue gas recirculation) and ammonia could also be injected into
the inlet air, underneath the grid plate.
Probes used to inject supplementary air or ammonia into the combustor
were constructed of 3/8 inch 316 SS tubing sealed off at one end. The probes
extended horizontally across the diameter of the combustion zone to the far
wall, and contain three or four holes drilled horizontally. The holes were
sized and located so that high velocity gas streams could be obtained and
directed to impinge upon the combustor walls. This provided adequate mixing
of the incoming gases with the contents of the fluidized bed. Injection
locations were varied, depending upon the test conditions; they are summarized
in Table VII-1.
The reasons for the various probe locations are discussed in detail in
those parts of the text relating to the specific program.
The NH3 addition system consisted of a cylinder of NHs maintained at
about 40°C in a heated water bath. NH3 gas was transferred through heated
lines to prevent condensation and metered through a rotameter. N2 and H? can
also be metered and mixed with
Other than for these modifications, the bench unit combustor remains
essentially as described in the previous report (1).
Two Stage Combustion
The program to study the effects of staged combustion upon NOX emission
was developed as a 2 X 3 X 2 factorial design and included the effects of
overall excess air, primary air to fuel ratio (before injection of secondary
air), and bed temperature (see Table VII-2).
Initially, it was intended to conduct the designed study at a combustor
pressure of 8 atmospheres, but operational difficulties prevented this. At 8
atmospheres, superficial gas velocities 1n the primary combustion (reducing)
173
-------
TABLE VII-1. LOCATION OF GAS INJECTION
POINTS BENCH UNIT NOX CONTROL STUDIES
Program
-• Staged Combustion
Ammonia Injection
Simulated Flue Gas
Recirculation
Injector
No.
1
2
3
1
2
3
Injector Location
Distance Above Grid
200
15
33
Distance Relative to Top of Expanded
(cm)
78 above
140 below
89 below
Bed
168 46 above
Ammonia mixed with combustion air - no injectors used
290 168 above
N added to combustion air - no injectors used
-------
TABLE VI1-2. DESIGNED EXPERIMENT FOR
STUDY OF STAGED COMBUSTION IN BENCH UNIT
Bench Unit Combustor Run Number
Primary
Average
Overall
Air (% of Stoich) •>
Bed Temp. (°C) ->
Percent
Excess Air
Unstaged Staged Combustion
Combustion 90% 75%
840 930 840 930 840 930
15
21.1 20.1
21.2 20.2
21.3 20.3
30
17A1 22.1
17A2 22.2
17A3 22.3
-------
zone were too low for good fluidization. This resulted In repeated episodes
of bed agglomeration. As a consequence, all of the runs in the designed
experiment were made at 5 atmospheres. Champion coal (-16 mesh, 2.19% sulfur)
was used in all runs, and the Ca/S ratio was set at 3.0.
Originally, it was planned to use the entire bed as the reducing zone
and to inject the supplementary air above the bed to oxidize the CO formed.
In the Initial runs at 8 atm, the supplementary air probe was located 200 cm
above the grid, or about 45 cm above the top of the bed. Pfizer dolomite was
selected as the sorbent. This scheme was not practical. The dolomite
attrited rapidly and was elutriated from the bed early in the run. This
resulted in smaller (and shorter) beds, and greater distances between the top
of the bed and the point of air injection. As a consequence, the off gas
temperature at the point of air injection was on the order of 600-650°C rather
than at the bed temperature of 870-930°C. This resulted in Incomplete burnout
of CO. The ignition point of CO in air at atmospheric pressure is about 610°C.
Occasionally, there were brief temperature rises above the air injection point
indicating that combustion of CO was taking place sporadically. In light of
the above, 1t was decided to place the probe in the bed at 15 cm above the
grid and to use Grove limestone (-8+25 mesh) as the sorbent to reduce attri-
tion losses.
These changes provided adequate CO burnout, but resulted 1n a limited and
unstable reducing zone. During several runs at 8 atm, fusion of the combustor
bed occurred as a consequence of erratic coal feed rates and poor mixing in
the bed, particularly at low primary air flow rates. Partial disassembly of
the combustor revealed that the refractory lining immeidately above the grid
had been eroded. The diameter was 14 cm to 15 cm rather than the original dia-
meter of 11.5 cm. This resulted in low gas velocities 1n this area and
inadequate mixing in the bed. As a consequence, the low velocities and poor
solids mixing were probably responsible for fusion of the bed. Another pos-
sibility was that CaS was formed and, upon occasion, was carried into the
oxidizing zone. The oxidation of CaS to CaS04 is highly exothermic, and would
result in excessive bed temperatures and fusion of the bed material.
The refractory lining was repaired and the diameter of the combustor
restored to the original 11.5 cm. At this point, the combustor pressure was
reduced from 8 atmospheres to 5 atmospheres for subsequent runs, These
changes resulted 1n higher gas velocities and improved mixing of the bed
solids. The supplementary air probe was relocated to a height of 33 cm above
the grid in order to provide a larger reducing zone than was obtainable when
the probe was located 15 cm above the grid.
The results of the study were evaluated by Analysis of Variance (ANOVA)
techniques for both main effects and interactions. If an effect could not be
demonstrated at the 90% confidence level (or higher), it was assumed not to
exist.
Emission levels were characterized as "Emission Indices" 1n terms of
pounds of the components in the flue gas per MBTU of fuel supplied. This
approach is preferable to using the component concentrations in the flue gas,
as the latter is affected by excess air levels.
176
-------
Detailed results of the test program are given 1n Appendix P-2.
Effect of Operating Parameters Upon NOx Emlssions--
Data obtained at 8 atm_pressure"—The results of the present study were
compared with those from earlier work to measure NOx emissions during unstaged
combustion. The earlier results were reported 1n a previous report (2). The
results of the earlier work, at excess air levels of 0% to 40%, were combined
with results of four new unstaged runs (runs 12.1, 13.1, 14.1, 17.1) made at
8 atm. A "t Test" with both paired and unpaired comparisons was used to deter-
mine the effect of staging. The staged tests (runs 10.1, 11.1, 11.2, 12.2,
13.2, 14.2, 17.2) were carried out at conditions similar to those in Table
VII-2. The test indicated that staged combustion reduces NOX emissions about
30% below those resulting from unstaged combustion (see Table VI1-3). The
"t Test" Indicated that the confidence level was greater than 90% when the
unpaired comparison was made and over 98% when the paired comparison was made.
Data obtained at 5 atm pressure—The data obtained at 5 atmospheres —
through the completion of the test matrix in Table VII-2 — are summarized in
Table VI1-4. Staged combustion reduced the emission Index (Ibs N02/MBTU) from
0.272 (Ibs/MBTU) (without staging) to 0.245 (Ibs/MBTU) with the primary air
at 90% of the stoichlometric amount required to burn the coal. This is a
reduction of only 10%. A reduction of 20% to 0.220 (Ibs/MBTU) was achieved
by reducing the primary air to 75% of stoichlometric. The effect was observed
at the 90% confidence level. These reductions are lower than those observed
at 8 atm pressure.
The staged and unstaged data obtained in this series were also analyzed
to determine the effects of variables other than staging on NOX emissions.
For all runs, staged and unstaged, increasing the overall excess air levels
from 15% to 30% increased the NOX emissions from 0.222 (Ibs/MBTU) to 0.270
(Ibs/MBTU), an increase of 22%, Increasing the combustor temperature from
843°C to 927°C increased emission Indices from 0,218 (lb/MBTU) to 0.273 (lb/
MBTU), a 25% Increase with increasing temperature.
These results are generally consistent with previous findings from the
bench and minlplant units (1,2,3).
No statistically significant Interactions were found.
Effects of Operating Parameters on S02 Emissions—
The S02 emissions observed during the staged combustion matrix at 5 atm,
are shown 1n Table VII-5. Staged combustion appears to Increase the $63 emis-
sion levels. The emission index increased from 1.66 for all unstaged runs to
1.90 (lb/MBTU) for staged runs, an Increase of 16%. Effects of temperature
and excess air on S02 emission levels were also found.
High combustion temperatures, regardless of staging or excess air levels,
decreased S02 emission levels. Increasing the combustor's operating tempera-
ture from 840°C to 930°C reduced S02 emission by 21%.
177
-------
TABLE VI1-3. EFFECTS OF STAGED COMBUSTION
ON N0 LEVELS AT 8 ATMOSPHERES
Overall
NQX (Ib/M BTU) Excess Air (EA)
7
s2
Run No.
12.2
13.2
14.2
17.2
—
«•
Staged
0.14
0.19
0.19
0.21
0.182
0.030
Unstaged(l)
0.192
0.235
0.256
0.317
0.250
0.052
A
0.052
0.045
0.066
0.107
0.067
0.028
(%)
7.8
15.7
19.6
30.6
--
— ^
(1) Calculated from regression equation tNOx (lb/M BTU) -
0.00548 (% EA) + 0.149], for all unstaged combustor runs.
v~
-£• = TfTJfr = 1.374 or 37% increase in NOX levels with unstaged
X~ ' combustion over the NOX levels with
staged combustion.
r
1 - 37- » 1 - 0.728 » 0.272 or 27% reduction in NOX levels
X~c when using staged combustion.
178
-------
TABLE VI1-4. EFFECTS OF STAGED COMBUSTION ON N0y EMISSIONS AT 5 ATMOSPHERES
Emission Indices for NOX (lb N02/M BTU)
10
Unstaged
Primary Air (% of Stolen) -»- Combustion
Average Bed Temp. (°C) -»• 840 930
Overall Percent Excess Air
15 0.20 0.32
30 0.26 0.31
Sub Column Means = 0.230 0.315
Column Means = 0.272
Group Means
Group 1 (840°C) = 0.218
Group 2 (930°C) = 0.273
Grand Mean = 0.246
Staged Combustion
90%
840 930 840
0.22 0.23 0.15
0.23 0.30 0.25
0.225 0.265 0.200
0.245
Significant
Effect
Temperature
Staging
Excess Air
75%
930
0.21
0.27
0.240
0.220
Effects
Conf, Limits
95%
90%
95%
Row
Means
0.222
0.270
-------
TABLE VII-5. EFFECT OF THE BENCH UNIT'S OPERATING PARAMETERS
UPON S02 EMISSIONS AT 5 ATMOSPHERES
00
o
Emission Indices for SOz (Ibs/M BTU)
Primary
Average
Overall
Air (% of Stoich) -»•
Bed Temp. (°C) -»•
Percent Excess Air
15
30
Unstaged
Combustion
840
2.25
1.48
930
2.02
0.890
Staged Combustion
90%
840
2.23
2.16
930
1.79
1.63
75%
840
1.56
2.53
930
1.78
1.56
Row
Means
1.94
1.71
Sub Column Means =
Column Means =
1.86 1.45
1.66
2.20 1.71
1.95
2.04 1.67
1.86
Group Means
Group 1 (840°C) = 2.04
Group 2 (930°C) = 1.61
Grand Mean = 1.82
Significant Effects
Effect
Temperature
Excess Air-Temperature Interaction
Excess Air-Staging Interaction
Conf. Limits
95
90
95
-------
Unstaged combustion at high excess air levels (30%) produced the lowest
S02 emission levels. The mean emission level under these conditions was 1.18
(lb/MBTU) (the average of 1.48 and 0.890), while the mean emission level for
all other twelve conditions was 1.97 (lb/MBTU), a 40% difference 1n emission
levels. Combustion under the condition of high excess air levels (30%) and
high temperature (930°C), regardless of staging, materially decreased S02
levels. The mean SO? emission level under these conditions was 1.36 (lb/MBTU),
while the mean for all other eleven conditions was 1.82 (Ib/MBTU) or a reduc-
tion in emission levels of 25%.
The effects of the combustor operating parameters upon SOg emission
levels may be interpreted In light of the following equations.
CaC03 - — » CaO + C02 (1)
(2)
The detrimental effect of staging is probably due to the need to have a
sufficient oxygen level to promote equation 2. Since oxygen is depleted in
the reducing zone, equation 2 is hindered there. Equation 2 will occur in the
oxidizing zone, but because the gas residence time in the oxidizing zone is
less in the staged configuration, equation 2 is further hindered,
The effects of the higher combustor operating temperatures upon SO? emis-
sion levels stems from the fact that the sorbent 1s more completely calcined
(Eq. 1) at930°C than 1t is at840°C. The higher the degree of calcination,
the greater the degree of porosity and the more reactive the sorbent.
The apparent effect of excess air 1n the unstaged runs is not under-
stood. Although the increase in excess air could promote equation 2, it is
believed that at excess levels of 15% or so, oxygen 1s present in sufficient
excess to have no further effect.
Overall, the SO?. emissions measured in this test series were fairly high,
compared to results obtained 1n the m1n1plant. The difference 1s due to a
much lower gas phase residence time in the bench studies (-0.5 s) compared to
normal mlniplant operations (2 to 3 s).
Effects of Operating Parameters Upon CO Emissions--
Staged combustion Increases CO emission moderately, and then only at low
levels of primary air (see Table VII-6). The mean value of the CO emission
Index of 0.292 (Ibs/MBTU) for a primary air rate of 90% of stolchiometrlc
1s not significantly different from the mean value of 0.277 (Ibs/MBTU) for
unstaged combustion. However, the mean value of the CO emission indices of
0.335 (Ibs/MBTU) for a primary air rate of 75% of stoichlometric Is 20% higher,
which is significantly different from the value of unstaged combustion.
CO emission indices are markedly Increased by low combustion temperatures
an effect seen 1n previous studies (2). Decreasing the combustor 's operating
temperature from 930 °c to 840 °C Increased the mean CO emission indices from
181
-------
TABLE VI1-6. CO EMISSIONS FOR STAGED AND UNSTAGED COMBUSTION AT 5 ATMOSPHERES
Emission Indices for CO (Ibs/M BTU)
oo
Unstaged
mary Air (% of Stoich) ->• Combustion
rage Bed Temp. (°C) -> 843 927
Percent Excess Air
15 0.35 0.25
30 0.33 0.18
Sub Column Means = 0.340 0.215
Column Means = 0.277
Group Means
Group 1 (843°C) = 0.355
Group 2 (927°C) = 0.248
Grand Mean = 0.302
90%
843 927
0.29 0.22
0.38 0.28
0.335 0.250
0.292
Temperature
Staging
75%
843 927
0.34 0.29
0.44 0.27
0.390 0.280
0.335
Sianificant Effect
Effect
Temperature, Excess Air Interaction
Staging, Excess
Air Interaction
Row
Means
0.290
0.313
Conf. Limit
99%
95%
95%
95%
-------
0.248 (Ibs/MBTU) to 0.355 (Ibs/MBTU), an increase of 43%. Over the range
of 15 to 30% excess air, the level of excess air has no significant effect
upon the CO emission indices. This finding was unexpected.
High excess air levels (30%) and low combustion temperatures (840=0)
Increase CO emission levels. At these operating conditions the mean CO emis-
sion index was 0.38 (Ib/MBTU) while the mean value for other operation con-
ditions was 0.30 (Ib/MBTU), or an increase of 27%. This may be the result
of the combined effect of higher levels of CO formation at the lower tempera-
ture and the introduction of large amounts of cold secondary air into the bed,
which has the effect of causing local cooling, thereby preventing CO burnout.
The combination of high excess air levels (30%) and low primary air rates
(75%) also increases CO emission levels. This effect may be due to the higher
levels of CO formed in the bed at low primary air rates and the introduction
of large volumes of cold secondary air, which would retard CO burnout.
Conclusions--
Two stage combustion was shown to reduce NOX emissions by about 20%.
However, S02 and CO emissions were both increased about 20% by staging.
Increasing the gas phase residence time in the oxidizing section of the com-
bustor slightly should offset the increased $03 and CO emissions. Two stage
combustion could also create a boiler tube materials problem if cooling coils
were subjected to alternate low and high oxygen concentrations. This was not
addressed in this study but must be at some point, if two stage combustion is
to be considered further.
NH3 Injection
Exxon Research and Engineering Company has developed a process based on
the selective, homogeneous, gas phase reduction of NO by NH3. The amount of
NH3 needed is comparable to the amount of NO reduced. Temperature has an
effect on the effectiveness of the reaction. When the temperature is too low,
NHa and NO tend to remain unreacted, when the temperature is too high, NHs
tends to form additional NO. Thus, 1t 1s possible to achieve an efficient
reduction of NO with little NH3 remaining, but only within a narrow temperature
range. The NOX destruction and production reactions are:
4NO + 4NH3 + 02 + 4N2 + 6H20 (1)
4NH3 + 502 •»> 4NO + 6H20 (2)
Reaction (1) dominates at temperatures around 950°C whereas reaction (2)
dominates above 1100°C.
R. K. Lyon (21) proposed a free radical chain mechanism for the reaction
of NH3, 02 and NO to reduce NOX:
NH2 + NO * N2 + H + OH (1)
NH2 + NO * N2 + H20 (2)
183
-------
H + 02 + OH + 0
0 + NH3 + OH + NH2
OH + NH3 -»• H20 + NH2
H + NH3 -> H2 + NH2
The addition of a third component such as hydrogen, carbon monoxide or various
hydrocarbons which form reactive radical intermediates, reduces the optimal
temperature for the reaction (2). Hydrogen is preferred since it is itself
not an air pollutant. Although hydrogen reduces the optimal temperature for
the reaction, it has the disadvantage of decreasing the selectivity with which
NH3 reduces the NO. Hence, if too much hydrogen is added, NH3 may react to
form NO rather than to reduce NO. The preferred levels of the gases based on
tests at low pressure are:
NH3 NH3/NO = 0.5-3.0
H2 H2/NH3 < 3
02 0.1-20.0 volume %
At H2/NH3 ratios around 2/1, the NOX reduction reaction can proceed at
700°C. By selecting the proper H2/NH3 ratio, the reaction can be accomplished
at any temperature between 700°C and 950°C (22).
Laboratory data has shown that reductions in NOX emissions up to 70% are
possible. The process has been commercially demonstrated in gas and oil-fired
steam boilers and process furnaces. Until this work, no tests have been made
on a pressurized fluldized bed coal combustion system.
Experimental Conditions—
A series of eight runs covering 27 conditions were made to test the effect
of NHs injection on NOX emissions. The NH3/NOX and H2/NH3 ratios were varied
and 3 injection levels were tested. Table VII-7 lists the runs which were
made. Figure VII-2 shows the location of the 3 injection levels and Table
VII-8 indicates the port (and height above the grid) which was used at each
level.
The injection level, or probe location, was set prior to the start of a
run. In each experimental day, a series of steady state periods was obtained.
Each series began with a steady state period at baseline conditions (no NHi
injection) and usually ended at baseline conditions to see if the emissions
levels returned to the same base level.
Constant operating conditions were as follows:
Champion Coal (-16 mesh, 2.19% S)
Grove Limestone (-8 + 25 mesh)
Ca/S mole ratio: 3.0
184
-------
TABLE VII-7. AMMONIA INJECTION RUN CONDITIONS
NH3 Injection
Run Number
1.1
1.2
1.4
3.1
3.2
4.1
4.2
4.3
5.1
5.2
5.3
5.4
5.5
5.5
5.7
6.1
6.2
6.3
7.1
7.2
7.3
7.4
8.1
8.2
8.3
8.4
8.5
Locat1on
LI
LI
LI
L2
(1)
Approx.
Temp.
816
816
816
954
Conditions
L3
L3
LI
704
704
816
NWNOx
No Injection
0.88 0
0.88 2.00
No Injection
2.27 1.48
No Injection
2.00 1.00
No Injection
No Injection
2.18 2.94
1.27 2.50
No Injection
3.41 0
8.05 0
No Injection
No Injection
0.82 0
No Injection
No Injection
1.36 0
2.25 0
No Injection
No Injection
1.24 0
2.29 0
1.39 1.46
No Injection
(1) See Figure VII-2 and Table VII-8 for description of
each location.
185
-------
FIGURE VII-2
BENCH COMBUSTOR SHOWING LOCATIONS
FOR AMMONIA INJECTION
COMBUSTOR
SHELL
REFRACTORY
LINING
AMMONIA
•>
(L3)
AMMONIA
(LI)
COILS
AMMONIA (L2)
1st
SOLIDS
OVERFLOW
GRID
MAIN AIR
186
-------
TABLE VII-8. AMMONIA INJECTION LOCATION
Probe Height Above
Location Description Grid (cm)
L3 Inject ammonia near top 290
of column
Ll Inject ammonia above bed 168
(at overflow port)
12 Mix ammonia with main air Below Grid
stream
187
-------
Pressure: 657 to 758 kPaa
Bed Temperature: 900°C
Excess Air: 15%
Results and Discussions--
The operating conditions for each of the runs is summarized in Appendix
Table P-3. In all, 27 steady state periods were run which included 13 periods
at baseline conditions (with no ammonia injection). Figure VII-3 shows the
emission indices for NOX, S0£ and CO, respectively. These bar graphs show
the 95% confidence interval for each steady state level. Standard deviations
for NOX, S02 and CO emission index, used to draw the confidence interval (or
2 standard deviation bars), were calculated from the 13 base level steady
states.
NOx Em is sums--In the 5 and 7 run series a statistically significant
change in the NOX emissions index occurred. Within each of the run series 1
3,4, 6, and 8 the confidence intervals overlap, suggesting that a real change
in the level may not have occurred. Significant changes in the NOX emissions
from the baseline level occurred in runs 5.5, 5.6, 7.2, and 7.3.
Table VII-9 summarizes the results. The NOX emission index is a strong
function of the NH3 injection location and NH3/NOX ratio. With injection into
the main air stream (run 5) the NOX level increased 50%. With injection above
the bed (runs 1, 3, 4, and 8) the NOX level remained unchanged. With the
injection near the top (runs 6 and 7), the NOX level decreased 30 to 50%.
Table VII-10 shows that with injection into the main air (run 5) the NOX'
level increased in proportion to the amount of NH3 injected (NH3/NOX ratio)
and with injection near the top (run 7) the NOX level decreased in proportion
to the amount of NH3 injected.
Because changing the injection location changes the injection temperature
the effect of injection location on the NOX emissions may be simply a tempera- '
ture effect. If this is true, then it would appear that in this combustion
system, the NH3-NO reaction is faster at 700°C, the NH3-02 reaction is faster
at 950°C, and at 820°C the rates of the two reactions are equal. However
oxygen partial pressures and mixing may be different, as well as temperature
at each injection location. High oxygen partial pressures, which probably *
exist immediately above the grid, favor NOX production. This may explain NO
production at injection location 12. Slugging may cause a high oxygen levelX
above the bed (LI) also.
Table VII-11 compares the results of this work to the laboratory data
obtained in another study (21). The optimal temperati-e for the NOX reduc-
tion reaction without hydrogen addition appears to have shifted down from
950°C in the earlier laboratory study to 700°C in this work with the bench
combustor. This shift is possibly due, in part, to the presence of CO or
unburned hydrocarbons in the flue gas. CO and hydrocarbons may act in place
of hydrogen as a source of radicals for the NH3-NO reaction.
Effect on SQj) and CO Emissionsj--Examination of the bar graphs for SO?
and CO emission indices (Figure VII-3) reveals that no significant change
occurred due to ammonia injection location or NH^/NO ratio.
0 X
188
-------
FIGURE VII-3
EMISSION INDICES FOR AMMONIA INJECTION PROGRAM
SHOWING 95% CONFIDENCE INTERVALS
0.'
X
LLJ
II
00 \
oo .of
o
z
0.3
0.2
0.1
NO EMISSIONS
x
X
LLJ
o
Z
LO \
LO -Q
CN
O
«, ,
X
LLJ
Q
ll
o
u
.11.21.4 3.13.2 4.1 4^4.3 5.1 5.25.3 5.4 5^5j5 5J
AMMONIA INJECTION RUN NUMBER
INJECT ABOVE BED *h— INJECT INTO *\
MAIN AIR
INJECTION LOCATION
189
-------
FIGURE VII-3 (CONT'D)
EMISSION INDICES FOR AMMONIA INJECTION PROGRAM
SHOWING 95% CONFIDENCE INTERVALS
Q
z
£. D
O £
Q
Z
Of
00
CN
Q
i
00
O
u
0.3
0.2
0.1
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
NO EMISSIONS
x
QLJ 1 1 1 1 1 1 1 1 L
SO2 EMISSIONS
CO EMISSIONS
8
1
6.1 62 6.3 7.1 72 7.3 7A 8.1 82 8.3 8.4 8.5
AMMONIA INJECTION RUN NUMBER
INJECT >f< INJECT
NEAR TOP ABOVE BED
INJECTION LOCATION
190
-------
TABLE VII-9. RESULTS OF PROGRAM TO
REDUCE NOX BY AMMONIA INJECTION
Temperature
Signi ficant
Run
Series No.
1
3
4
5
6
7
8
Injection
Level
LI
LI
LI
L2
L3
L3
LI
at Injection
Point (°C)
816
816
816
954
704
704
816
Change in NOX
Emission Index?
No
No
No
Yes
Yes
Yes
No
Increased/Decreased
NOX Levels
—
—
—
Increased
Decreased
Decreased
_ __
Maximum
Change (%)
—
—
—
50
30
50
__
-------
TABLE VI1-10. RESULTS OF AMMONIA INJECTION RUNS 5 AND 7
NH3 Injection
Run Conditions
No. NH3/NOX H2/NH3
Base (No NH3
Injection)
5.2
5.3
5.5
5.6
2.18
1.27
3.41
8.05
2.94
2.50
0.0
0.0
(Ib NQ2/MBTU)
Base
Level
0.21
Level with
NH3 Injection
NOX Emissions
Change
From Base
Level (%)
Significant
Change
0.24
0.23
0.27
0.32
+14
+10
+29
+52
No
No
Yes
Yes
7.2
7.3
Base (No NH3
Injection)
1.36
2.25
0.0
0.0
0.18
0.12
0.09
-33
-50
Yes
Yes
192
-------
TABLE VII-11 . COMPARISON OF THE RESULTS OF THIS WORK AND AN
EARLIER WORK (21) ON AMMONIA INJECTION TO REDUCE NOU
Temperature (°C)
Added
NH3
"-
Yes
Yes
Yes
Yes
Yes
Yes
H2
No
No
No
No
Yes
Yes
Earlier
Work (21) This Work (PFBCC -
(Lab Data) Bench Combustor)
1093
954
816
949 704
871
704
Change in NOX
( Increase/Decrease)
Increase i
Increase *
No Change
Decrease 4-
Decrease 4-
Decrease 4-
193
-------
Effect of Excess Air Level on N0y-F1gure VII-4 shows that the excess air
level varied only slightly from a mean of 17.2%. The two steady states for
which the NOX emissions Increased significantly (run 5} and the two steady
states for which the NOX emissions decreased significantly (run 7) are shown
1n addition to the 13 base level steady state periods. Figure VII-4 shows
that the change in NOX emissions due to NH3 Injection 1s not due to a change
In the excess air levels. The uncontrolled NOX emissions 1n this study are
generally lower than In earlier bench combustor studies (2). This 1s pre-
sently unexplained.
Response Time—Figure VII-5 shows graphically how response times were
determined for run 5. The response time 1s the time required for the system
to reach a new steady state NOX level once a change 1n the ammonia Injection
conditions (NHs/NOx and Hg/NHs) is made. Table VII-12 shows the response
times for run 5. The system response is very rapid, averaging 10 minutes.
Ammonia Material Balance—
NHs material balances were attempted by analyzing all solid and gaseous
product streams. The balances were very low and variable, ranging from 14 to
69%. Analysis of sol Ids collected by the cyclones and by the final filter
Indicated that ammonium salts were forming and precipitating on the collected
sol Ids. The compounds were most likely NfyHSOs and/or NfyHSO^ The mass of
ammonia which would have to have been lost to account for the low mass
balances was very little, ranging from 4 to 130 gm. Therefore, the poor
balances could have been caused by the formation and deposition of ammonium
salts, possibly 1n the flue gas ducting. Other sampling and analytical
problems may have also occurred which contributed to the low balances.
Conclusions—
NHj injection at temperatures around 700°C in PFBC flue gas can reduce
NOX emissions 30 to BQ%. The temperature level 1s lower than expected and
further work may be needed to determine 1f it is the true optimum. If such
a temperature is required, NH3 could be Injected ahead of the gas turbine.
The desired temperature would occur within the turbine Itself, where the NO
removal reactions would take place.
Simulated Flue Gas Recirculation
In the simulated flue gas redrculatlon (SFGR) program, nitrogen was
mixed with the main air stream to simulate flue gas reclrculatlon. The
variables studied were:
Recirculation Ratio, R * N2/A1r * 0.10, 0.20
Average bed temperature = 840, 900°C
Excess A1r = 15, 30%
Constant operating conditions were:
Pressure = 808 kPaa
Ca/S mole ratio =3.0
Champion coal (-16 mesh, 2.19 wt. % S)
Grove limestone (-8 + 25 mesh)
194
-------
FIGURE VII-4
AMMONIA INJECTION PROGRAM:
EFFECT OF EXCESS AIR LEVEL ON NOx EMISSIONS
U.H
0.3
S
Y~
co
0 n o
Z O'2
_Q
X
O
Z
0.1
0
1 1 1
A
A
• •
••
•»o»-
4h V ^k
v^
•*A\
A \
O
MEAN OF THE BASE LEVEL
NH3 DATA:
(EXCESS AIR: 17.2%)(NOX: 0.17 Ib/MBTU)
i i i
AMMONIA
INJECTION
10 20
EXCESS AIR(%)
30
40
BASE LEVEL (No NH3 INJECTION)
WITH NH3 INJECTION (RUN 7)
WITH NH3 INJECTION (RUN 5)
195
-------
a.
a.
LU
o
z
FIGURE VII-5
RESPONSE TIME
CHANGE IN
OPERATING VARIABLE
STEADY-STATE NO. 1
STEADY-STATE NO. 2
RESPONSE
TIME
TIME (MIN)
196
-------
TABLE VII-12. RESPONSE TIME FOR AMMONIA INJECTION RUN 5
Run
No.
5.1
5.2
5.3
5.4
5.5
5.6
5.7
Operating Variables
Subject to Change
NH3/NOX
H2/NH3
Response
Time
(min)
Base Level (No Injection)
2.18
1.27
Base
3.41
8.05
Base
2.94
2.50
Level (No Injection)
0.0
0.0
Level (No Injection)
10
10
20
7.5
5
7.5
NOX Level
(ppm)
145
210
200
160
210
250
170
Avg. = 10
197
-------
The N£ flow system used to cool down the combustor during an emergency shut-
down was used to inject N£ to simulate flue gas recirculation. The No is
mixed with the main air stream far upstream of the injection point and there-
fore, perfect mixing may be assumed.
Results and Discussion--
The run conditions for the nine steady state periods are summarized in
Appendix Table P-4. In each run, a steady state period was maintained without
nitrogen addition and then in a second period, nitrogen was added without
adding additional air. However, additional coal had to be burned to heat the
nitrogen which was injected cold. Therefore, the excess air levels reported
for the second period are all lower than for the first. Figure VII-6 shows
the bar graphs for each emission index (E.I.) for N02, S02, and CO, and the
95% confidence interval for the emission levels. The error bars shown on
the graphs were drawn using the standard deviations derived from a statistical
analysis of the ammonia injection program data.
Table VII-13 summarizes the results of the SFGR program, Table VII-14
indicates the significance of changes in the emissions for each run. Signi-
ficant reductions in the NOX emissions, observed in runs 2 and 4, are cor-
related with changes in the excess air level which accompanied N2 addition
during these tests. Increases in both the S02 emission index and CO emission
index are correlated to changes in the excess air level and the gas residence
time. These changes in the NOX, SC^ and CO emissions are discussed below.
NOy Emissi'ons--Figure VII-7 compares the NOX emissions vs. excess air
results for the SFGR program with the baseline NOX emissions from the ammonia
injection program (no NH3 injection). Table VII-15 shows that the mean of the
SFGR data (all data) nearly coincides with the mean of the ammonia injection
data (base levels only). Also, because NOX emissions decrease with decreasing
excess air levels, it is believed that the lower NOX emissions which resulted
from N2 addition in these tests are really the result of lowering the excess
air level. If the test had been conducted in a way which would have prevented
an accompanying decrease in excess air with N^ addition (e.g., by increasing
the air rate), it is expected that no change in NOX emissions would have been
observed. Therefore, flue gas recirculation should not have an effect on
reducing NOX emissions.
SO? and CO Emissions—As seen in Table VII-13, both S02 and CO emissions
increase as a result of N2 addition. Although the increase in SO? emissions
could not be shown unequivocally to be statistically significant (Table VII-14),
it was consistent.
Adding nitrogen to the main air has two effects:
1. The partial pressure of oxygen is lowered.
2. The gas velocity is increased, which lowers the residence
time of the combined gas stream.
198
-------
FIGURE VII-6
EMISSION INDICES FOR SIMULATED FLUE GAS RECIRCULATION (SFGR)
PROGRAM SHOWING 95% CONFIDENCE INTERVALS
X
LLJ
Q
Z
Z S
O 2
x
O
Z
Z =
LLJ
CN
Q
Z
o..
0.3
ca
CO
00
0.1
2.4
2.0
0.8
0.7
0.6
0.5
0.4
o 0.3
0.2
8
NO EMISSIONS
X
_| I | | i
SO2 EMISSIONS
m
P
i
CO EMISSIONS
^
^i
g
VA
i
i
1A.1 1A.3 2.1 3.1 4.1
1A.2 2.2 3.2 4.2
SFGR RUN NUMBER
Note: The standard deviation values used to draw the error bars are from
a statistical analysis of the Ammonia Injection Program data.
199
-------
TABLE VII-I3. SIMULATED FLUE GAS RECIRCULATION
PROGRAM - SUMMARY OF RESULTS
SFGR
Run
No.
1A.1
1A.2
1A.3
2.1
2.2
3.1
3.2
4.1
4.2
Independent Variables
R
(%) Bed Temp. (°C)
874
11.2 886
S82
Average = 881
874
23.4 887
Average = 881
906
11.2 897
Average = 901
819
11.3 817
Average =818
Excess
Air (%)
18.6
12.9
17.2
18.4
7.4
26.5
25.0
16.7
13.0
Dependent Variables
Emission
NOx
0.19
0.16
0.19
0.22
0.07
0.20
0.17
0.29
0.13
Avg =
0.18
Index (1
S02
0.89
1.17
0.72
0.81
1.09
1.01
1.07
1.18
1.72
Avg =
1.07
b/MBTU)
CO
0.14
0.34
0.33
0.17
0.27
0.14
0.45
0.29
0.56
Avg =
0.30
200
-------
TABLE VI1-14. SIGNIFICANCE OF CHANGES IN EMISSIONS -
SIMULATED FLUE GAS RECIRCULATION PROGRAM
Emission Index
(Ib/MBTU)
Run No.
-v
lA.ll
1A.3J
1A.2
2.1
2.2
3.1
3.2
4.1
4.2
1A.2
2.1
2.2
3.1
3.2
4.1
4.2
lA.fl
1A.3J
1A.2
2.1
2.2
3.1
3.2
4.1
4.2
Base
Level
0.190
0,216
0.203
0.293
0.804
0.813
1.014
1.177
0.237
0.170
0.135
0.292
Level Change
With N? From Base
Addition Level (%)
NOX Emissions
0.164
0.068
0.173
0.129
S02 Emissions
1.168
1.094
1.066
1.722
CO Emissions
0.341
0.274
0.446
0.562
201
-14
-67
-15
-56
+45
+35
+5
+46
+44
+61
+230
+93
Significant
Change?
Maybe
Yes
Maybe
Yes
Maybe
Maybe
No
Maybe
No
No
Yes
Yes
-------
FIGURE VII-7
COMPARISON OF THE SIMULATED FLUE GAS RECIRCULATION (SFGR)
PROGRAM WITH THE AMMONIA INJECTION PROGRAM
i—
CQ
o
z
0.4
0.3
O 0.2
0.1 -
O Mean of all of the
SFGR data: (Excess Air:
17%)(NO : 0.18
lb/MBTU)X
I
I
SFGR PROGRAM
SFGR PROGRAM
AMMONIA
INJECTION PROGRAM
10 20 30
EXCESS AIR (%)
(BASE LEVEL)
(WITH N2 ADDITION)
(BASE LEVEL)
40
202
-------
TABLE VI 1-1.5. COMPARISON OF AMMONIA INJECTION AND
SIMULATED FLUE GAS RECIRCULATION RESULTS
Mean Value
Excess Air (%) NOX (lb NOT/MBTU
Ammonia Injection 17 9 n 17
(Base Levels Only) "'*
Simulated Flue Gas
Recirculation 17.3 0.18
(All Data)
203
-------
The first effect, the lowering of the oxygen partial pressure, may reduce
both the rate of reaction of SOg to form CaS04 and the rate of reaction of CO
to form C02- As was mentioned earlier, some excess oxygen is required for
the reaction of S02 with CaO according to the reaction:
CaO + S02 + 1/202 -»• CaS04
Therefore, reducing the oxygen partial pressure to low levels, may decrease
the S02 retention and increase the S02 emissions. CO reacts with 02 in the
gas phase according to the reaction:
CO + 1/202 -*• C02
Similarly, decreasing the oxygen partial pressure substantially will decrease
the rate of CO combustion and increase CO emissions.
The second effect, the lowering of the gas residence time, means that
there is less time for both of the above reactions to occur. Hence, adding
Ng to the gas stream means that the rate of each reaction and the time
available for each reaction to occur are less. Therefore, both the S02 and
CO emissions increase with N2 addition as shown by these runs.
For each steady state period, the sorbent feed rate was adjusted to
maintain the Ca/S mole ratio at a constant value of 3.0. However, Table
VII-16 shows that the Ca/S ratio varied about the average of 3.0, ranging
from 2.65 to 3.36. In two runs, Ca/S increased when N2 was injected and in
two runs Ca/S decreased when N2 was injected. This variation was not system-
atic; Ca/S did not change in proportion to the amount of N2 injected into the
main air. Therefore, any systematic changes in S02 emissions would not be
expected to be caused by changes in the Ca/S ratio.
Conclusions—-N? addition to the main air stream has no direct effect on
NOX emissions. The observed decrease in the NOX emissions in this program
can be explained by changes 1n the excess air level; No addition lowered the
percent excess air which in turn lowered the NOX emissions. If the excess
air level had been held constant, no change in NOX emissions would have been
expected.
N2 addition results in increased S02 and CO emissions probably because
of lowered oxygen partial pressure and the gas residence time.
Combined Techniques
A series of tests was conducted to determine if a combination of NOX
control techniques could result in further reductions in NOX emissions.
Since only staged combustion and ammonia injection caused a reduction in NOX
emissions, only the combination of these two methods was studied.
Two runs were completed and are discussed below. In run 1, two steady
states were run at base level conditions, two with NH3 Injection, and then
one with staged combustion. No sulfur retention data were obtained because
the S02 analyzer was inoperable during this run. In run 2 one steady state
204
-------
TABLE VII-16. VARIATION IN Ca/S MOLE RATIO
Run No.
1A.1
1A.2
1A.3
2.1
2.2
3.1
3.2
4.1
4.2
Without
No Addition
(Base Level)
2.97
3.10
3.36
3.05
2.55
With
N2 Addition
3.10
2.85
2.87
2.87
Change in
Ca/S (%)
+2.1
-15.2
-5.9
+8.3
Average =2.98
205
-------
was run at base level conditions, one with staged combustion, one with both
NH3 injection and staged combustion and then one more with staged combustion
only. The results are summarized in Appendix Table P-5. The constant opera-
ting conditions for both runs were:
Pressure 505 kPa
Temperature 900°C
Ca/S 3.0
Excess Air 25%
Coal Champion, -16 mesh, 2.19% S
Limestone Grove, -8 + 25 mesh
Ammonia injection probe location 290 cm above the grid
Supplementary air probe location 33 cm above the grid
Primary Air as % of Stoich 75% (2-stage combustion)
NH3/NOX 1.5 (ammonia injection)
The values of the staged combustion and the NHs injection variables are those
found to be promising in the previous testing.
Table VII-17 shows the changes in NOX levels in each run as changes
were made from base level to various NOX control and combination of NOX con-
trol techniques.
Table VII-17 shows that in this work, either ammonia injection or staged
combustion or a combination of the two techniques tend to lower NOX emissions
by 15 to 50% below the base level. The reduction in NOX emissions by staged
combustion 1s reasonably consistent with the earlier study at 505 kPa in
which reductions of 10 to 20% were observed. The reduction in NOX emissions
by ammonia Injection at 505 kPa with an NH3/NOX ratio of 0.89 to 1.43 is
consistent with the earlier ammonia injection program at 808 kPa in which a
reduction of 33% was observed with an NH3/NOX ratio of 1.36. Figure VII-8
shows that the reduction 1n NOx emissions from the base level when both
techniques were used simultaneously was not greater than the reduction when
either technique was used alone. However, since only one data point was
obtained when using the two techniques simultaneously, 1t cannot be stated
with any certainty whether an interaction does or does not exist. Perhaps,
if a larger NH3/NOX ratio had been used, a greater reduction in NOX may have
been achieved.
S02 and CO emission data were obtained for run 2. Table VII-18 shows
that the S02 emissions Increased from 0.87 to 1.03 Ib/MBTU going from base
level conditions (run 2.1) to staged combustion (run 2.2) and the CO emis-
sions were unchanged at 0.26 Ib/MBTU. However, over the entire run both the
S02 and CO emissions increased; runs 2.2 and 2.4 were both with staged com-
bustion only, but the S02 emissions Increased from 1.03 to 1.22 Ib/MBTU and
the CO emissions Increased from 0.26 to 0.35 Ib/MBTU. These upward drifts in
the emission levels are unexplained.
BENCH REGENERATION STUDIES
The bench unit regenerator, because of its relatively small size, pro-
vides a means to experiment with operating conditions and methods quickly
206
-------
TABLE VII-17.
COMBINED NOX-CONTROL TECHNIQUES PROGRAM.
CHANGES IN NOX LEVELS.
Run
No.
1.1
1.2
1.3
1.4
1.5
2.1
2.2
2.3
2.4
Excess
Air («)
25.1
22.6
25.2
23.3
22.0
31.3
24.5
25.3
24.2
Average Bed
Temp. (°C)
891
896
894
900
901
Avg. - 896
876
883
887
883
Avg. = 882
Control
NFh Injection
( Yes/No J(NH3/NOX
No
Yes 0.89
No
Yes 1 .43
No
No
No
Yes 1 .61
No
Technique
Staged
Combustion
) (Yes/No)
No
No
No
No
Yes
No
Yes
Yes
Yes
NOX
Mb/MBTU)
0.21
0.15
0.15
0.09
0.10
0.17
0.14
0.13
0.15
Change in NOX
from Base
Level (1) (%)
-15
-49
-43
-21
-26
-15
(1) Base level NOX emission = 0.18 Ib/MBTU
(Average of Runs 1.1, 1.3, 2.1)
-------
Z
g
fc
^>
CO
5
o
(J
Q
LLJ
O
YES
NO
FIGURE VII-8
REDUCTION IN NOX EMISSIONS BY USING
TWO CONTROL TECHNIQUES SIMULTANEOUSLY
0.21 )
— 0.15>
0.17/
AVG. =0.13
NOX
(Ib NO2/MBTU)
AVG. =0.18
(BASE LEVEL)
NO
0.13
O.is)
0.09)
AVG. =0.12
YES
NH3 INJECTION
208
-------
TABLE VII-18. S02 AND CO EMISSIONS FOR
COMBINED NO -CONTROL TECHNIQUES RUN 2
Run Emission Index (Ib/MBTU)
No. Control Technique S02CO
2.1 Base level 0.87 0.26
2.2 Staged-combustion 1.03 0.26
2.3 Staged-combustion 1.15 0.32
with ammonia injection
2.4 Staged-combustion 1.22 0.35
209
-------
and Inexpensively. Promising results obtained in the bench regenerator can
then be scaled up to miniplant regenerator operation,
The test program in the bench regenerator was divided into two segments,
natural gas fueled regeneration and coal fueled regeneration. Most of the
work concentrated on developing satisfactory operating techniques. The work
done with natural gas-fueled regeneration included a study of the most
effective way to introduce fuel to the regenerator and a limited study of
operating variables and their effect on sorbent regeneration. Studies made
with coal-fueled regeneration were preliminary 1n nature and were intended
to determine the feasibility of operating a regeneration unit with coal.
The incentive to use coal instead of natural gas is to reduce operating costs
and to reduce reliance upon gaseous (or liquid) fuels.
The results of the bench unit regenerator test program will be discussed
in the following sections.
Equipment
The batch regenerator unit was modified to permit semi-continuous opera-
tion and renamed the bench regenerator unit. The modified facilities were
discussed in the previous annual reports (1,2) and will not be described in
detail.
The regenerator vessel is 4.57 m high and.1s constructed of 12 inch car-
bon steel pipe and lined with Grefco 75-28 refractory to an inside diameter
of 9.52 cm. The burner plenum chamber below the grid 1s 0.69 m high and is
lined with Grefco Bubblelite refractory.
Sulfated sorbent is intermittently charged to the regenerator from a
lock hopper via a line equipped with two cycling valves to meter in the sor-
bent (see Figure VII-9). The sorbent is Introduced near the bottom of the
fluid bed (9.5 cm above the grid) to insure complete mixing. Sul fated sor-
bent from miniplant operations is used. Regenerated sorbent exits the bench
regenerator through a bed overflow line which leads to an overflow lock
hopper. The below-bed burner is used to heat up the bed and provides some
of the energy required by the regeneration reactions. A reducing zone is
established immediately above the grid by adding supplementary fuel directly
to the bed. This creates the CO and H2 rich zone needed to carry out the
regeneration reactions. Supplementary air is then added further up the bed
This creates a mildly oxidizing zone at the top of the bed which oxidizes
any undesirable CaS formed in the reducing zone. All remaining fuel is also
burned 1n the oxidizing zone, to maximize fuel u^e efficiency.
Natural Gas-Fueled Regeneration
Operating Procedures--
The Initial charge of bed material is heated to the desired temperature
under oxidizing conditions by operation of the below-bed burner. Sometimes,
small amounts of supplementary fuel are added directly to the bed to Increase
bed temperature, but not enough is added to produce reducing conditions 1n
the bed. Supplementary fuel can only be added when bed temperature is above
650°C, the temperature at which natural gas will ignite easily.
210
-------
FIGURE VII-9
BENCH REGENERATION UNIT
SORBENT
FEED
HOPPER
AUTOMATIC
VALVES
FLUIDIZING GRIDH-
SUPPLEMENTARY
AIR
COOLING
WATER IN
f
BURNER
COOLING WATER OUT
OFF GAS PRECOOLER
COOLING WATER IN
COOLING
WATER OUT
-OFF GAS COOLER
BACK PRESSURE Tr.
REGULATOR IU
SCRUBBER
OFF
GAS
FILTER
TO
ANALYTICAL
TRAIN
SORBENT
OVERFLOW
HOPPER
-------
When the bed temperature 1s uniform and close to the desired run tempera-
ture, the switchover to reducing conditions is made. This 1s 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 to avoid adding air 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. Oxygen and CO con-
centrations 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 1s the minimum value which will just produce CO 1n
the oxidizing zone.
Shakedown-
Shakedown of the bench regenerator was completed 1n two runs. During the
first shakedown run, all systems except the sorbent feeder operated success-
fully. The opening of the feed line in the regenerator was found to be par-
tially blocked with fused bed and refractory, which had been left from an
earlier batch run. The next run proved to be more successful. The sorbent
feed system functioned properly and the high temperature necessary for regen-
eration was achieved using the air preheater.
The main operational difficulty during shakedown runs was maintaining a
satisfactory bed height. Although the solids overflow port is 1.0m above
the grid, the highest expanded bed height obtained was under 0.5 m. It is
believed that the bed slugged badly and the solids were moving, piston-Uke,
up the regenerator and out the sorbent overflow due to high superficial gas
velocities. This was a continuing problem during the test program. The
effect was partly offset by minimizing the gas velocity.
Effect of Fuel Injection Mode-
One objective of the work done with natural gas-fueled regeneration was
to determine the best way of introducing the natural gas. The energy
required to maintain the temperature of the unit can be supplied in two ways.
The burner below the fluidizing grid can be operated at oxidizing conditions
and supplementary fuel added to the bed to create the reducing zone (Mode 1),
or all fuel can be added directly to the bed where it would partially burn In
the fluidizlng air (Mode 2) (see Figure VII-10). In both modes of opera-
tions, the reducing gas 1s produced by the Incomplete combustion of methane
which 1s added directly to the bed through the supplementary fuel probe.
Adding all the fuel to the bed eliminates the heat losses associated with
operation of the burner, thereby decreasing the fuel requirements and pos-
sibly increasing the S02 content in the off gas. One drawback to adding all
the fuel directly to the bed is that the entering cold fuel must heat up
to Ignition temperature before it burns. The heat up requires a certain
residence time in the bed, causing the methane to move further up the bed
before it Ignites. As the level of combustion varies, the location of the
reducing zone also varies which might have a detrimental effect on tempera-
ture control and on regeneration.
212
-------
FIGURE VII-10
FUEL INJECTION MODE
BED
FLUIDIZING
GRID
SUPPLEMENTARY
AIR
SUPPLEMENTARY
FUEL
BURNER (j-*
AIR —
FUEL
SUPPLEMENTARY
AIR
FUEL
MODE NO. 1
MODE NO. 2
213
-------
Burner section heat 1ogses--In determining the best mode of Introducing
fuel Into the regenerator, it was observed that a significantly smaller
amount of fuel was required to maintain bed temperature when all the fuel
was added directly to the bed as opposed to operation with the burner (e.g.,
0.079 compared to 0.045 NnvVmin). To account for this discrepancy in energy
requirements, the heat losses associated with burner operation were deter-
mined. Losses Include radiation and convection of heat from the burner
plenum, and heat loss to the cooling water in both the fluidizlng grid and
the burner head. These heat losses were calculated and are summarized below:
Radiation and convection from plenum:
Q1 = AkAT = 0.63 m2 x 21.8 JM'W1 (533-300) K
= 3200 Js"1
Flu1d1zat1on grid cooling water:
Q2 = mCpAT = 4 x 3.17 x 10"2kg sec"1 (4.187 x 103Jkg~1K~1)(306-294)K
= 6364 Js"1
Burner cooling water:
Q3 = mCpAT = 3.17 x lO'^gs"1 (4,187 x 103Jkg~V1)(328-294)K
= 4508 Js"1
Total heat loss = Q1 + Q2 + Q3 - 14072 Js"1
where Q = heat loss
A = area
k = combined heat transfer coefficient for radiation and
convection
AT s temperature differential
m = mass flow rate
Cp = heat capacity
The heating value of methane is 3.73 x 10 Jm" . Therefore, the heat
loss associated with burner operation 1s equivalent to the heating value of a
flow rate of 3.78 x lO-^nrs-"1 of methane. Thus, the difference 1n fuel
requirements between operation with and without the burner can be largely
explained by the heat losses associated with burner operation. Since about
0.079 Nm3/m1n of methane are required to operate the regenerator with the
burner, a significant portion (about 30 percent) of the fuel requirement Is
used to overcome heat losses.
Results and Discussion—
Mode No. 2 operations—Six runs were made 1n the bench regenerator to
evaluate operation in which all the fuel (methane) is added directly into the
bed. In five of the runs, supplementary fuel was injected at the wall of the
reactor. In the sixth run (R-8), a new supplementary fuel probe designed to
give better fuel distribution across the bed was used. Operation with all of
the fuel added to the bed generally resulted in an uniform temperature pro-
214
-------
file throughout the bed but controlling the temperature to the desired level
proved difficult. During the series of six runs, only one steady state regen-
eration condition was reached. The steady state was a 15 minute segment of
run R-6. Operating conditions for run R-6 are presented in Table VII-19.
Many of the regenerator runs in which all the fuel was added directly
Into the bed were ended by high temperature shutdowns. Controlling the tem-
perature proved difficult when mode No. 2 was used. The factors which might
cause a temperature runaway are:
1. While in oxidizing conditions during heat up:
a. an increase in fuel flow, thereby adding additional energy input
b. a decrease in air flow thereby increasing the temperature of
the preheated incoming gas
2. While in reducing conditions:
a. a decrease in fuel flow causing oxidizing conditions and
oxidation of CaS (an exothermic reaction)
b. an increase in air flow causing oxidizing conditions
3. In either reducing or oxidizing conditions:
a. a decrease or stoppage in solids feeding
b. poor mixing, including fuel bypassing, air bypassing, and
poor solids circulation, causing hot spots and agglomeration
Possible reasons for the problems of 1n-bed fuel injection into the regenerator
were examined and it is believed that the cause is related to mixing. In
order for fuel addition into the bed to work, the fuel must be heated above
Its ignition temperature shortly after injection into the bed. If the cold
fuel entering the regenerator through the probe does not mix with the hot bed
solids and is not brought up to Ignition temperature, the "flame" may be
blown out. It may be that the large, sudden, changes in temperature were
caused by ignition or extinction of the methane flame. These changes could
have been brought about by shifts 1n the pattern of solids mixing within the
fluidized bed.
The hypothesis that poor solids and gas mixing results in "flame blow
outs" is supported by experimental evidence. During many of the runs, it was
observed that the Og concentration in the off gas Increased suddenly, although
no change in operating conditions (fuel or air flow rates) had occurred. The
sudden increase in the 02 concentration in the off gas would be characteristic
of the effects of a flame blow out. The increase of 02 concentration, since
the air was no longer used 1n combustion, led to oxidation of the CaS present.
an exothermic reaction. The exothermic reaction resulted 1n high bed tempera-
tures and the subsequent high temperature shutdowns.
It 1s possible that the problem of a flame out could have been lessened,
or even eliminated, by heating to higher temperatures the air entering the
regenerator through the fluidizlng grid. This would have reduced the heat
transfer requirement produced by solids backmixing. Unfortunately, with the
equipment at hand, it was possible to heat the incoming air only to 150-
200°C.
215
-------
(M
TABLE VI1-19. BENCH REGENERATOR RUN SUMMARIES
Avg. Sup. Gas /?\ Solid Extent
Bed Expanded Gas Res. ^ ' Res. S02; Mole % of Solids
Run
Number
R-eO)
R-9
R-10
R-ll.l
R-11.2
R-12
R-13
R-14
Temp.
K
1370
1340
1350
1350
1350
1345
1395
1360
(1) Mode No.
(9\ * = <;tr
P
Bed
atm Height m
6.9
7.5
7.6
7.6
7.6
7.6
7.6
7.5
0.9
0.6
0.7
0.6
0.6
0.8
0.6
0.6
2 Operation (al
lirMnmetr-ir air
Vel.
m/s
0.7
1.1
1.0
0.9
0.9
0.9
1.2
1.1
1 fuel
tn fm
Time
sec.
1.34
0.56
0.70
0.66
0.70
0.89
0.50
0.55
added
al vati
Equiv Time
Ratio min.
1.43
1.42
1.64
1.85
1.83
1.50
1.57
1.45
to bed,
irt v ^US
29
48
35
43
29
51
22
37
burner
Average
Peak
(Range) Equil
1.17
(0.24-1.74)
0.62
(0.18-0.99)
0.58
(0.18-1.17)
0.59
(0.09-1.32)
0.24
(0.06-0.92)
0.27
(0.03-0.66)
0.53
(0.1-1.35)
0.43
(0.15-0.84)
not used).
0
0
0
0
0
0
0
0
% S02
. % S02
.25
.25
.25
.29
.20
.15
.15
.15
Regeneration
%
NA
NA
44
71
71
48
83
79
NA = Data not available.
-------
Also, 1n a larger bed with improved solids mixing, the difficulty in
maintaining the flame should be lessened and in-bed injection of the fuel
should be feasible.
Mode No. 1 operations—The alternative to in-bed fuel injection, as
described previously, is operation of the burner supplemented by in-bed fuel
injection. The disadvantage of this method of operation is the increased
heat losses which accompany burner operation; increased heat losses require
additional air and fuel inputs thereby diluting the off gas SOg concentra-
tion if thermodynamics is not limiting. An advantage of this method is that
the burner head serves as a flame holder, an element that was evidently
lacking in the previous method described. The combustion gases rising from
the burner also provide the heat necessary to ignite the incoming supple-
mentary fuel, eliminating the problem of flame blow outs.
Six successful runs were made with the burner in operation while supple-
mentary fuel is added above the fluidlzing grid to produce a reducing zone.
The results of these six runs, which represent the variable study, are pre-
sented in the next section.
Results of process variable study--0perat1on with both burner and supple-
mentary fuel resulted in a steadier and more easily controlled bed tempera-
ture. The important operating conditions for runs R-9 through R-14 are sum-
marized in Table VII-19. It was intended to determine the effect of operating
variables on SOg concentration in the off gas and extent of sorbent regenera-
tion. However, because of the data scatter and the limited amount of data
obtained, it was not possible to draw conclusions about the relationships
between operating variables and sorbent regeneration.
S02 concentration in the regenerator off gas can be limited by thermo-
dynamics, kinetics or the heat and mass balance. The SOg concentration from
the bench unit, as in the case of the miniplant, is limited by the heat and
mass balance. The operating variables in Table VII-19 (temperature, pressure,
gas residence time, solids residence time, and equivalence ratios) are either
thermodynamic or kinetic variables used to define the system. Since the
system is heat and mass balance limited, no conclusive relationships between
the variables and the S02 concentration in the off gas were expected. Indeed,
from the data in Table VII-19, no such relationships were apparent, confirm-
ing that the S02 levels were not kinetically or thermodynamics11y controlled.
A further consequence of the heat and mass balance limitation is very low SOg
concentrations in the regenerator off gas. As seen, the concentrations ranged
from 0.2 to 0.6%. The approach to thermodynamic equilibrium of the system can
be expressed as the ratio of peak S02 concentration to the S02 concentration
at equilibrium for the prevailing temperature and pressure. From Table
VII-19, it is seen that this ratio was below 0.3 (or 3Q% of equilibrium) for
all runs.
Coal Fueled Regeneration
The objective of the test series using coal in the bench regenerator was
to evaluate the feasibility of coal-fueled regeneration in a PFB. Factors
which needed to be evaluated were: uniformity of temperature in the fluidized
217
-------
bed, temperature control, tendency for bed agglomeration, S02 levels pro-
duced, and degree of reduction of CaSCty to CaO. As can be seen, these are
primarily operability questions.
Operating Procedures--
Coal is fed to the regenerator by using the same coal feeding and con-
trol system used for the bench combustor. The system was described in the
previous reports (1,2). A coal probe, similar to the design used for the
bench combustor, is directed downward at about a 45° angle and delivers coal
at the inner regenerator wall about 8 cm above the fluidizing grid. Coal is
fed to the regenerator after the bed temperature is heated above 650°C with
the natural gas burner.
Experimental Results and Discussions--
Nine experimental runs (Runs R-15 through R-23) were made with coal-
fueled regeneration. Four runs, R-19 to R-22, were shut down soon after coal
injection was started because of high and erratic bed temperatures which could
have resulted in serious bed agglomeration. The other five runs were more
successful; about ten hours of operation on coal were accumulated. Operating
conditions for the steady state periods are reported in Table VII-20. The
major operational problems during the runs were plugging of the off gas filter
maintaining proper superficial gas velocities, and bed temperature control. '
It was necessary to end three of the runs, R-15, 16 and 18, because the
sintered metal off gas filter became plugged, resulting in loss of column
pressure control. Filter plugging might have been caused by an increased
solids loading in the regenerator off gas due to the carryover of unburnt
coal and ash particles. Another factor which might have contributed to the
plugging of the filter was increased bed solids carryover due to excessive
superficial gas velocities.
There is a problem in maintaining proper superficial gas velocities when
fueling the regenerator with coal. When operating the regenerator with coal,
combustion air is introduced to the regenerator from two sources: primary
combustion air enters through the grid, and coal transport air enters through
the coal probe, which is positioned about 8 cm above the fluidizing grid and
aimed downward. Transport air flow is approximately 0.26 m^/min, which is
almost half of the total air flow. Exactly how the air entering via the coal
probe contributes to bed fluidization is unknown. Therefore, to insure good
bed fluidization, a minimum superficial gas velocity of 0.61 m/s was main-
tained just above the fluidizing grid by adjusting the primary air flow.
Operating pressure was varied during the coal run series to adjust the super-
ficial gas velocity to a level which would insure good bed fluidization and
avoid slugging of the bed.
The ability to control bed temperature is crucial to the operation of a
coal-fueled regenerator. Temperature runaways to 1100°C can result in exten-
sive bed agglomeration because of the presence of coal and ash in the bed.
During the coal runs, a few temperature runaways occurred, resulting in a
high temperature shutdown of the unit and a limited amount of bed fusion.
218
-------
no
H-•
10
TABLE VII-20. COAL FUELED BENCH REGENERATOR RUN SUMMARIES
Avg. Sup. Gas m Solid Extent
Bed Expanded Gas Res. 4>v' Res, SOe; Mole % Peak % S02 ofSollds
Run Temp. P Bed Vel. Time Equiv Time Average TZ7T\—j <-A Regeneration
Number K atm Height m m/s sec. Ratio min. (Range) tqui . » au2 %
R-17 1422 4.6 2.1 1.6 1.32 1.52 62 1.32 0.16 40
(0-3.5)
R-18 1422 4.7 0.7 1.5 0.44 1.54 16 2.1 0.17 38
(0.1-3.7)
R-23 1338 6.1 0.8 1.5 0.51 1.56 31 0.9 — 40
A A1r Required for Complete Combustion
<(> Mr Supplied
Assumes 100% Combustion Efficiency
Sorbent: Sulfated Dolomite
Coal: Champion
-------
When the unit was operated in oxidizing conditions, temperature control
was not a problem. However, temperature control was difficult under reducing
conditions. To maintain reducing conditions in the unit, the coal feed rate
must be such that the air-to-fuel ratio is substoichiometric. With the large
amount of transport and primary air required, the coal feed rate needed to
produce reducing conditions exceeded that needed to satisfy the energy
requirements of the unit. That is, operating temperatures were too high and
resulted in high temperature shutdowns. To reduce the coal feed rate, com-
bustion air flow rates also had to be reduced.
To achieve this reduction in air rate, the coal feeding system was mod-
ified to use nitrogen, instead of air, as the transport medium, making the
primary air the only source of combustion air. Only one run, R-23, was made
with the modified coal transport system. During the run bed temperatures
were more easily controlled than in previous runs. However, after 25 minutes
of steady state regeneration, bed temperatures again began to rise. At this
point, the run was ended to avoid agglomeration of the bed. The tendency for
the bed temperature to rise after a given period of steady state regeneration
has been noted not only in the bench regenerator studies but also in the
batch studies carried out in the batch and miniplant regenerator units. The
most probable reason for the temperature increase is that the bench unit
operates as a "batch regenerator" during startup because the unit was charged
with sulfated sorbent prior to the run. Initially, the energy released by
the combustion of the coal is absorbed in the endothermic regeneration of
this large mass of startup bed material. When the entire bed inventory is
regenerated, the endothermic regeneration reaction takes place at a lower
rate, since only the incoming sorbent is being regenerated. Unless the
coal feed rate is reduced to compensate for the lower heat requirement, a
temperature increase could result after the initial charge is regenerated.
Changes in the off gas composition which occurred concurrently with the
temperature rise support this explanation. The S02 level dropped even
though the regenerator remained in reducing conditions.
The results of the few runs in the bech regenerator are promising. No
problems were encountered which would rule out the use of coal. Maintaining
good bed fluidization while still maintaining a constant bed height was a
problem as it was with natural gas fueled regeneration. The partial bed
agglomerations that did occur were probably a result of poor fluidization.
The small diameter (9.5 cm) of the regenerator makes good gas and solids
mixing and fluidization difficult. Temperature control was a problem at
times but again, the cause could have been poor gas and solids mixing in the
small regenerator vessel. Regeneration levels for the coal runs (-40$) were
not as high as the natural gas runs. A possible explanation is that the
concentration of CO and H2 in the reducing zone was lower than in the pre-
vious runs. Obviously, more work must be done in the area of coal-fueled PFB
sorbent regeneration on both small and larger scales to make a complete
evaluation. Such questions as where and how to add supplementary air and
coal to maintain proper fluidization, mixing and temperature control must be
answered. Modifications to the regenerator vessel, such as a tapered cross
section in the lower part of the vessel, should also be considered.
220
-------
SECTION VIII
REFERENCES
1 Hoke, R. C., et al., "Miniplant Studies of Pressurized FTuidized-Bed
Coal Combustion: Third Annual Report," EPA-600/7-78-069, April 1978.
2. Hoke, R. C., et al., "Studies of the Pressurized Fluidized-Bed Coal
Combustion Process," EPA-600/7-77-107, September 1977.
3. Hoke, R. C., et al., "Studies of the Pressurized Fluidized-Bed Coal
Combustion Process," EPA-600/7-76-011, September 1976.
4. Hoke, R. C., et al., "A Regenerative Limestone Process for Fluidized-Bed
Coal Combustion and Desulfurization," EPA-650/2-74-001, January 1974.
5. Skopp, A., et al., "Studies of the Fluidized Lime-Bed Coal Combustion
Desulfurization System," December 31, 1971.
6. Nutkis, M. S., et al., "Hot Corrosion/Erosion Testing of Materials for
Application to Advanced Power Conversion Systems Using Coal-Derived Fuels
Task II - Fluidized Bed Combustion," to be published.
7 Smith, W. B. and Wilson, R. R., "Development and Laboratory Evaluation
of a Five Stage Cyclone System," EPA-600/7-78-008, January 1978.
8. "Pilat (University of Washington) Mark III Source Test Cascade Impactor,"
Pollution Control Systems Corporation, Bulletin 76-3A.
9. Parker, R., et al., "High Temperature and Pressure Particle Collection
Mechanisms," Second Symposium on the Transfer and Utilization of
Particulate Control Technology, Denver, CO, July 23-27, 1979.
10 Koch, W. H. and Licht, W., "New Design Approach Boosts Cyclone
Efficiency," Chem Eng. No. 7, 1977. p 80 ff.
11. Knowlton, T. M. and Bachovchin, D. B., "The Effect of Pressure and
Solids Loading on Cyclone Performance," Presented at AIChE 70th Annual
Meeting, New York, N.Y., November 1977.
12. Zenz, F. A., Private Communication with R. C. Hoke.
13 Jonke, A., et al., "Supportive Studies in Fluidized Bed Combustion,"
' Argonne National Laboratory, EPA-600/7-77-138, December 1977.
14 Murthy, K. S., Howes, J. E., Nack, H. and Hoke, R. C., "Emissions from
Pressurized Fluidized-Bed Combustion Processes," Environ. Sci. and Tech.,
13(2), 197-204, February 1979.
221
-------
15. 6CA Corporation, GCA/Technology Division, "Test Plan for Level I
Characterization of Emissions from Exxon PFBC Miniplant Unit with
Regeneration," 1979.
16. GCA Corporation, GCA/Technology Division, "Test Plan for Level II
Characterization of the Emissions from the Exxon Fluidized-Bed Combustion
Miniplant," 1979.
17. Snyder, R., et al., "Supportive Studies in Fluidized-Bed Combustion,"
Argonne National Laboratory, Quarterly Report, ANL/CEN/FE-77-8,
July-September 1977, p. 31-32.
18. Hubble, B. R., et al., Argonne National Laboratory, "Chemical, Structural,
and Morphological Studies of Dolomite in Sulfation and Regeneration
Reactions," paper presented at the Fourth International Conference on
Fluidized-Bed Combustion held at the Mitre Corporation, McLean, VA,
December 9-11, 1975.
19. Hubble, B. R., et al., "A Development Program on Pressurized-Bed
Combustion," Argonne National Laboratory, Annual Report, ANL/ES-CEN-1011
July 1, 1974-June 30, 1975, p. 84-86.
20. Cunningham, P., et al., "A Development Program on Pressurized Fluidized
Bed Combustion," Argonne National Laboratory, Annual Report,
EPA-600/7-76-019, July 1975-June 1976, p. 134.
21. Lyon, R. K., International Journal of Chemical Kinetics, 8, p. 315-18
1976.
22. U.S. Patent 3900554, "Method for the Reduction of the Concentration of
NO in Combustion Effluents Using Ammonia."
23. Pigford, R. L.^and Sliger, G., "Rate of Diffusion-Controlled Reaction
Between a Gas and a Porous Solid Sphere," Ind. Eng. Chem. Process.
Des. Develop., 12(1), 1973, pp. 85-99.
24. Szekely, J., Evans, J. W. and Sohn, H. Y., Gas-Solid Reactions.
Academic Press, New York, 1976, pp. 125-168.
25. Borgwardt, R. H. and Harvey, R. D., "Properties of Carbonate Rocks
Related to S02 Reactivity," Environ. Sci. and Tech., 6/4), 1972,
pp. 350-360.
26. Lentzen, D. E., et al., "IERL-RTP Procedures Manual Level 1 Environmental
Assessment (Second Edition)," EPA-600/7-78-201, October, 1978.
27. Duke, K. M., et al., "IERL-RTP Procedures Manual: Level 1 Environmental
Assessment Biological Tests for Pilot Studies," EPA-600/7-77-043,
April 1, 1977.
28. Shackleton, M. A., "Extended Tests on Saffil Alumina Filter Media."
EPA-600/7-79-112, May 1979.
222
-------
SECTION IX
LIST OF PUBLICATIONS
1. Hoke, R. C., Gregory, M. W., "Evaluation of a Granular Bed Filter for
Partlculate Control 1n Fluidlzed Bed Combustion," Proceedings of the
EPA/ERDA Symposium on High Temperature/Pressure Particulate Control,
Washington, DC, September 20-21, 1977.
2. Nutkis, M. S., Hoke, R. C., Gregory, M. W,, Bertrand, R. R.,
"Evaluation of a Granular Bed Filter for Particulate Control in Fluidized
Bed Combustion," Proceedings of the Fifth International Conference on
Fluidized Bed Combustion, Vol. Ill p 504, Washington, DC, December
12-14, 1977.
3. Ruth, L. A., Hoke, R. C., Nutkis, M. S., Bertrand, R. R., "Pressurized
Fluidized Bed Coal Combustion and Sorbent Regeneration," Proceedings
of the Fifth International Conference on Fluidized Bed Combustion,
Vol. Ill p 756, Washington, DC, December 12-14, 1977.
4. Ernst, M., Loughnane, M. D., Bertrand, R. R., "Instrumental Methods for
Process Definition in a Pressurized Fluidized Bed Coal Combustion Pilot
Plant," AIChE 85th National Meeting, Philadelphia, PA, June 4-8, 1978.
5 Ruth, L. A., "Regenerate Sorbents for Fluidized Bed Combustion,"
Final Report prepared under NSF RANN Grant AER75-16194, June 1978.
6 Hodges, J. L., Hoke, R. C., Bertrand, R. R., "Prediction of Temperature
Profiles in Fluid Bed Boilers," Journal of Heat Transfer Vol. 100, No. 3,
p 508-513, August 1978.
7. Ruth, L. A., "Regeneration of the Sulfur Acceptor in Fluidized Bed
Combustion," Fluldization, Proceedings of the Second Engineering
Foundation Conference, p 303-313, 1978.
8 Ruth, L. A., Varga, G. M., "Developing Regenerable S02 Sorbents for
Fluidized Bed Coal Combustion Using Thermogravimetrlc Analysis,"
Thermochimica Acta, 25, p 241-55, 1978.
9. Murthy, K. S., Howes, J. E., Nack, H., Hoke, R. C., "Emissions from
Pressurized Fluidized-Bed Combustion Processes," Envir. Sci and Tech.
Vol. 13, No. 2, p 197-204, February.1979.
10 Hoke, R. C., Ruth, L. A., Ernst, M., "Control of Emissions from the
Pressurized Fluidized Bed Combustion of Coal," AIChE 86th National
Meeting, Houston, TX, April 1-5, 1979.
11 Siminski, V. J., Ernst, M., "Operating Experiences at Exxon Research and
'" "-^ ~ its in a Pressurized Fluid1z<
Meeting, Houston, TX, April
3 III! I I O "^ I ) •• **•• *-i ii%* v | ' ' • | wpwi %» v i ny fc./\ f
Engineering with Materials and Components in a Pressurized Fluidized Bed
Combustor (PFBC)," AIChE 86th National Me<
1-5, 1979.
223
-------
12. Hoke, R. C., Ruth, L. A., "Control of Emissions from the Pressurized
Fluidized Bed Combustion of Coal," American Flame Research Committee
Boston, MA, April 30, 1979.
13. Hoke, R. C., Ruth, L. A., Ernst, M., Nutkis, M. S., Garabrant, A. R.,
Goodwin, J. L., Radovsky, I.E., "Miniplant PFBC Studies," Pressurized
Fluidized Bed Combustion Technology Exchange Workshop, Secaucus, NJ
June 6, 1979.
14. Ruth, L. A., Varga, G. M., "New Regenerate Sorbents for Fluidized Bed
Coal Combustion," Environmental Science and Technology. (13) p 715-720
June 1979.
15. Ernst, M., Hoke, R. C., Siminski, V. J., McCain, J. D., Parker, R.,
Drehmel, D. C., "Evaluation of a Cyclonic Type Dust Collector for High
Temperature, High Pressure Particulate Control," Second Symposium on the
Transfer and Utilization of Particulate Control Technology, Denver, CO
July 23-27, 1979.
16. Ernst, M., Shackelton, M. A., Drehmel, 0. C., "Ceramic Filter Tests at
the EPA/Exxon PFBC Miniplant," Second Symposium on the Transfer and
Utilization of Particulate Control Technology, Denver, CO, July 23-27.
1979.
17. Jahnig, C. E., Shaw, H., Hoke, R. C., "Continuous Sorbent Regeneration
in Pressurized Fluidized Bed Combustion," Proceedings of the 14th
Intersociety Energy Conversion Engineering Conference, Vol. I, p 933,
Boston, MA, August 5-10, 1979.
OTHER PRESENTATIONS
Bertrand, R. R., "Pressurized Fluidized Bed Combustion," Fluidized Combustion
of Coal Symposium, MIT Industrial Liaison Program, January 18, 1979.
224
-------
SECTION X
APPENDICES
Page
A Effect of Particle Size on Conversion Rate of Limestone 227
and Dolomite: Thermal Gravimetric Experiments
B Testing of Various Stones for Ability to Absorb S02 230
C Data Management Section 232
D Participate Size Distribution and Concentration 237
Measurement Procedures
E Analytical Techniques 241
F Comparison of 862 Measurements by UV, GC, Wet Chemistry 243
G Results of Flue Gas $03 Analysis 246
H Regenerator GC Analyses 247
I Mass Spectrographic Analyses of Solid Samples of Run 69 248
j Miniplant Component Mass Balances 251
« Miniplant Run Objectives 253
L Mini plant Fluidized Bed Coal Combustion Run Summary 254
M Particle Size Distribution 271
M_l Spent Pfizer 1337 Dolomite Sorbent 271
M-2 First Cyclone Dip! eg 272
M_3 Second Cyclone Capture 273
M-4 Tertiary Cyclone Capture 276
M-5 and Grain Loadings of Flue Gas Participates Before 278
Tertiary Cyclone
M-6 and Grain Loadings of Flue Gas Parti culates After 280
Tertiary Cyclone
N Miniplant Solids Analysis 286
225
-------
APPENDICES (Continued)
Page
0 Mint pi ant Sample Shipments 301
P Bench Combustor Run Summary 305
P-l Initial Checkout 3Q5
P-2 Two Stage Combustion - NOX Control 308
P-3 NH3 Injection - NOX Control 312
P-4 Simulated Flue Gas Recirculation - NOX Control 31g
P-5 Combined NOX Control Methods 318
226
-------
APPENDIX A
EFFECT OF PARTICLE SIZE ON CONVERSION RATE OF
LIMESTONE AND DOLOMITE: THERMAL GRAVIMETRIC EXPERIMENTS
The sulfatlon of limestone and dolomite has been described by models
which are based on the idea that the conversion rate is governed by diffusion.
Several models, called "grain" models, assume that individual limestone par-
ticles are composed of many sub-particles, or grains (23,24). These grains
are imagined to be non-porous. When limestone reacts with SOg, S02 diffuses
through the pores separating the grains (intergranular pores) and reacts on
the grain surfaces. The speed of the reaction is determined by the local S02
concentration inside the pores and by the chemical reaction rate at the sur-
face of the grain. Alternatively, and this is probably the case, when SO?
reacts with limestone or dolomite, the speed of the reaction is determined
by the local S02 concentration inside the pore and by the thickness of the
solid reaction product layer which forms on the surface of each grain.
The grain model implies the primary importance of physical properties
such as porosity, pore size distribution, surface area, and grain size, in
determining the conversion rate of a limestone or dolomite particle. Experi-
mental studies have confirmed that these properties are important.
The effect of particle size on the conversion rate of limestone depends
on which step in the reaction scheme is rate controlling. If diffusion
through intergranular pores is controlling, then a concentration gradient of
SO? could exist from the surface of the particle to its center. Grains near
the surface would sulfate faster because of the higher S02 concentration and
a shell of reaction product (CaSO^ would form on the surface of the particle.
In this situation, particle size has the greatest effect on conversion rate
and the time required to achieve a given fractional conversion of the par-
ticle is proportional to 1/RS where R is the particle radius. At the other
extreme is the situation where S02 is spread uniformly through the pores of
the particle and all of the particle's interior reacts at the same rate. In
this case, particle size will have no effect on conversion rate.
In order to explore the effect of particle size on conversion rate,
limestone and dolomite particles of different sizes were sulfated in a TGA
at conditions of 900°C, 0.25% S02, 5% 02, and balance N2. Four particle
size ranges were investigated: -8 + 25 mesh, -16 + 40 mesh, -100 + 200 mesh,
and -325 mesh. These size ranges correspond approximately to average par-
ticle diameters of about 1500, 700, 100 and 40 ym, respectively. Table A-l
aives data on initial conversion rates, the time to reach 15 percent conver-
sion, and the percentage conversion after 75 and 150 minutes for each of the
four average particle sizes.
A number of significant observations can be made from the data of Table
A-1.
227
-------
TABLE A-l. EFFECT OF PARTICLE SIZE ON THE SULFATION OF DOLOMITE AND LIMESTONE
Dolomite (Pfizer)
ro
ro
oo
Average Particle
Initial Rate,
Time for 15%
(1) Average of 3 runs.
(2) Average of 2 runs.
% Conversion
Diameter, um
1500(1)
700^
100
40<2>
Average Particle
Diameter, ym
1500
700
100<2>
4°(2)
mg S03/(mg CaO)(min.)
0.24
0.044
0.100
0.087
Limestone
Initial Rate,
mq S0.i/(mq CaO)(m1n.)
0.010
0.017
0.056
0.055
Conversion, min.
32
4.0
1.5
2.0
(Grove)
Time for 15%
Conversion, min.
150
150
3.5
3.3
?5 min
28
40
39
75
%
75 min
6.4
5.9
59
79
150 min.
32
44
45
90
Conversion
150 m1n.
7.1
6.7
70
89
-------
1. Particle size has a large effect on the percentage conversion
or utilization of both limestone and dolomite. For example,
the percentage conversion of limestone increases by about a
factor of ten as particle size is reduced from 1500 to 100 urn.
2. Coarse dolomite achieves much higher utilizations than coarse
limestone. However, utilizations for fine (100 and 40 ym)
limestone and dolomite are comparable.
3. The initial rates for dolomite are higher than for limestone
for each of the particle sizes studied.
4. Initial rates, time for 15% conversion, and percentage
conversion after 75 and 150 minutes are all more sensitive
to changes in particle size for the larger size particles.
Each of the above observations can be interpreted in terms of the grain
model. Observation (1), the effect of particle size on utilization, implies
that a concentration gradient of SO? through the particle does exist and
that diffusion of S02 through a sulfate shell surrounding the particle does
influence the conversion rate. Observation (2) suggests that dolomite has
larger intergranular pores than limestone. Indeed, pore diameters for a
limestone and dolomite, different from the stones used in this study, were
measured by Borgwardt and Harvey (25). Their limestone had pores about 1 ym
in diameter whereas their dolomite had 0.5-20 ym pores. Because of the wider
intergranular pores, large dolomite particles achieve higher utilizations
than large limestone particles. However, as particle size is reduced, so is
the influence of Intergranular diffusion. Thus, for small particles, lime-
stone and dolomite give comparable utilization.
Observation (3), that the initial rate for dolomite is higher than for
limestone is consistent with the hypothesis that dolomite has larger pores.
Also, dolomite may have smaller grains than limestone. Observation (4), that
the effects of particle size become less as size 1s reduced, again implies
that intergranular diffusion becomes less important as the particle size
becomes smaller.
There are practical implications for fluidized bed coal combustors from
the above results. Using smaller sorbent particles, particularly when the
sorbent used is limestone, can result in a significant reduction in the
auantity of sorbent required. Although, the use of smaller particles will
Result 1n higher entrapment rates, oversized cyclones can be used to return
sorbent fines to the bed. Lower superficial gas velocities may also be used
to reduce entrapment. Whether or not the use of fine particles is practical
would depend on the trade-off of reduced sorbent and sorbent disposal costs
vs. Increased equipment costs (I.e.. the cyclone recycle) and potentially
lower equipment capacity.
229
-------
APPENDIX B
TESTING OF VARIOUS STONES
FOR ABILITY TO ABSORB S02
Several limestones and dolomites were screened 1n the thermogravlmetrlc
analyzer (TGA) for their ability to absorb S02. Two gas compositions were
used, one to promote calcination and the other to suppress it. Table B-l
gives the reaction conditions and the fraction of each stone sulfated after
60 minutes. The "standard" sorbents, Grove limestone and Pfizer dolomite
are shown for comparison,
The data of Table B-l shows that there 1s considerable variability In
reactivity among the stones, even when testing was done under calcining con-
ditions (condition "A" in Table B-l). The activity of Plum Run dolomite is
fairly high, even under conditions which suppress calcination of CaCOs
(condition "B" 1n Table B-l) because the MgCOj in the dolomite always cal-
cines, Introducing porosity Into the stone. The substantial MgO content of
the low grade limestone 1s probably responsible for its good activity under
condition B. The high grade Mulzer limestone, however, shows fair activity
even under "non-calcining" conditions, even though 1t contains little MgO.
This result was probably caused by partial calcination of the CaCOs in the
stone. The other sorbents showed relatively low activity even under "cal-
cining" conditions. The reasons are not known but could be due to other
structural differences 1n the sorbents.
It should be noted that all stones of Table B-l were tested as powders.
The fractional sulfations observed would have been quite different, probably
much lower, had the stone bed tested 1n the form of 2-3 mm granules, commonly
used 1n fluldized bed combustors.
230
-------
TABLE B-l. REACTIVITY OF VARIOUS STONES UNDERGOING SULFATION
Fraction Sulfated 1n 60 Minutes
ro
to
Sample (Powder) Test
(
Dolomite, Plum Run^ '
Limestone, Mulzer, High Grade' '
Limestone, Consolidation Coal,
Low GradeU/
Limestone, Ste. Genevieve, High Purity^ '
Limestone, Dolomitic^ '
Limestone, Grove, BCR No. 1359
Dolomite. Pfizer, BCR No. 1337
Test Conditions
A - 870-900'C; 0.25% SO?. 5% Op. bal N£
Sample heated to 870 C in N2 prior to sulfation
B - 870°C; 0.25% SO?, 5% 0?. bal CQ2
Sample heated to 870 C in Cp2 prior to sulfation
Conditions "A" Test Conditions "B"
"Calcining") ("Non-Calcining")
0.51 0.29, 0.58(2)
0.36 0.11, 0.13^
0.45 0.28, 0.37^2)
0.17
0.062
0.30
0.39
to suppress calcination
(1) Samples supplied by American Electric Power Service Corp.
(2) Replicate run
(3) Samples supplied by Western Materials Company
-------
APPENDIX C
DATA MANAGEMENT SYSTEM
The data generation and management system for the minlplant combustor
was thoroughly described 1n the previous annual report (1), The entire sys-
tem has been augmented to Include the collection, computation, and summariza-
tion of minlplant regenerator data. Description of the major additions
follow.
HOKE SYSTEM
Figures C-l and C-2 are examples of the tabular form of regenerator data
printout received from a run. This printout 1s then used 1n conjunction with
the combustor printout and the continuous recorder charts to determine the
steady state period.
INPUT/OUTPUT PROGRAMS
Regenerator segments of the computer printout generated by the I/O
program are shown 1n Figures C-3 and C-4. The first figure is an example of
the experimental data inputs used 1n the program and Figure C-4 is an example
of the calculated outputs. Addition regenerator data, such as off gas emis-
sions and solids chemical analyses and particle size distributions, are
Inputed in the appropriate sections. The regenerator effluent streams are
Included 1n the overall mass accountability section,
232
-------
FIGURE C-l
ANALYSIS PROGRAM
4/30/79
REGENERATOR SECTION
ro
CO
CO
TEMPERATURESIDfcG
1
2
3
4
5
6
7
8
9
10
11
TIME
PORT
PORT
PORT
PORT
PORT
PORT
PORT
PORT
PORT
PORT
RUN
5.12
5.13
5.13
5.13
5.12
NO.
NO.
NO.
NO.
NO.
NO.
NO.
NO.
NO.
NO.
18
19
20
21
22
F)
2
3
A
5
7
6
9
10
11
12
1
29
29
29
29
29
(5IN)
( 11IN)
( 17IN)
I23IN)
( 35IN)
(41IN)
(48IN)
(56IN)
(64IN)
(74IN)
23456789 10 11
1820. 1837. 1852. 1859. 1874. 1889. 1885. 1887. 1890. 1893.
1819. 1833. 1849. 1856. 1871. 1886. 1881. 1884. 1887. 1889.
1818. 1832. 1848. 1855. 1869. 1883. 1878. 1880. 1883. 1886.
1815. 1833. 1851. 1857. 1871. 1884. 1879. 1880. 1883. 1888.
1814. 1834. 1852. 1858. 1871. 1885. 1880. 1880. 1884. 1887.
-------
FIGURE C-2
MINIPLANT ANALYSIS PROGRAM
A/30/79
REGENERATOR SECTION
oo
12 TIKE
REDUCTION ZONE
13 SUPERFICIAL VELOCITY(FT/SEC )
14 AVERAGE BED TEMPERATURE(DEC F)
15 AIR-FUEL RATIO(PHI)
OXIDIZING ZONE
16 SUPERFICIAL VELOCITY(FT/SEC )
17 AVERAGE BEL) TEMPERATURE ( DEG F)
18 AIR-FUEL RATIO(PHI)
19 EXPANDED BED HEIGHT(FT)
20 OFFGAS CO LEVEL(PPM)
21 OFFGAS S02 LEVEL(VPCT)
RUN
12
13
15
16
17
18
19
20
21
5.12
5.13
5.13
5.13
5.12
18 29
19 29
20 29
21 29
22 29
2.14 1842.98 1.27
2.14 1840.33 1.26
2.14 1838.95 1.27
2.15 1839.76 1.27
2.15 1840.09 1.27
2.75 1886.46 1.01
2.75 1883.37 1.01
2.75 1880.16 1.01
2.74 1881.18 1.02
2.75 1881.35 1.02
6.20 1274.00 0.55
5.26 1347.48 0.45
7.10 1249.00 0.52
7,56 1385.56 0.34
7.03 1424.60 0.52
-------
FIGURE C-3
ro
CO
en
lFMPFBATUHf(Ci!.LSIUS)
PORT NO. 2 ( 5 IN)
FORT NO. 3 ( 11 IN)
PORT NO. 4 ( 17 IN)
FORT NO. 5 ( 23 IN)
PORT NO. 7 ( 35 IN)
PORT NO. 3 ( 41 IN)
PORT NO, 9 ( 43 IN)
FOiiT NO, 13 ( 56 IN)
FORT NO. 11 ( o4 IN)
FORT MO. 12 ( 74 IN)
fORT NO. 13 ( 83 IN)
PORT NO. 1-1 ( 92 IN)
PORT NO. If (11 PI IN)
PORT NO. 13 (158 IN)
FORT NO. 21 (156 IN)
FORT NO. 34 (194 IN)
FORT NO. 26 (218 IN)
CYCLONE DISCHARGE
COOLER OFFGAS
PRiSSURSS(KPA)
NOMINAL OPERATING
R-C PRESSURE DROP
GRID PRESSURE DROP
BSD PRESSURE DROFS
994.
1334*
1012.
1016.
1025.
1032.
1029.
1230.
1031.
1332.
13c6.
1013.
871.
775-
935.
886.
881.
1443.
420.
735.961
5.961
2.194
SETTLFD BED HFIGHT(M)
INITIAL
FINAL
HalSHI ABOVE :
SUFP AIR PROBE
SUFF FU&L F*OfcE
BED BULK DENSirT(GR/Cl3
INITIAL
FINAL
1.6?
3.74
2.13
******
FORTS 29 TO 371 (5 IN TO 41 IN) 6.573
FORTS 29 TO 34 (5 IN TO 224 IN) 12.413
FLOW HATES(M3/MIN)
BURNER AIR
BURNER FUEL
SUPPLEMtNTART AIR
SUFPLEMENTARlf FUEL
1.85
3.18
2.46
2.2G
C-R TRANSFER RATF(KG/HE)
15.4
AUXILLIAfilf N? FI3»'{13/1IN)
.?. 1
-------
FIGURE C-4
REGENERATOR SECTION
SUPERFICIAL moCITT(M/SKC)
F.EDUCIN3 ZON2
OXIDIZING ZONE
3.64
0.73
SULFATIOM LEVEL DF SORPSNTUPCT)
TO REGENERATOR
FROM REGENERATOR
23.4?
********
EFT TFMPFPATURFS(CKLSIUS)
REDUCING ZONE - PORTS 2 TO !>
OXIL'IZINO ZONE- POSTS 7 TO 12
10B6.79
1339.77
R£SEH«RAriO,M Ll'JZL(FCI) 39.76
COMBUSIOR FEED SULFUR IN R&3KN OfFJAS(rfPCT) 73.CG
ro
CO
at
SETTLED BED HEI3HT(M)
EXFfcKt'ED BED HF.IGHT(M)
GAS RFSIDFNCE TIME(SEO)
SOLIDS RESIDENCE TIME(HRS)
1 .85
2.34
3.23
ENERGK IMFUT(MBTU/HR)
3.51
AIR TO FUSL RflriO(PHI)
HECUCINC ZONF
OXIDIZING ZONE
1.28
1.33
-------
APPENDIX D
PARTICLE SIZE DISTRIBUTION AND
CONCENTRATION MEASUREMENT PROCEDURES
The measurement of PFB generated participates is 1n Its Infancy.
Sophisticated high temperature high pressure 1n-s1tu size distribution measur-
ing devices are being developed and tested to understand better the partl-
culate emissions. While these systems are being developed, simpler systems
are being used to generate data which will give some Indication of the par-
tlculate emissions and particle size distribution. In the roiniplant this is
done by sampling for flue gas particulates with Balston total filters. Some
stream concentrations which cannot be sampled directly are calculated based
on mass balances, Size analysis of all fine particulates (-45 ym) is done
with the Coulter Counter Model TA11. The procedure developed for the mini-
plant will be described in this section. This section assumes a working
knowledge of particulate sampling and Coulter Counter operations. It is meant
to supplement, rather than replace good sampling practice and/or the Coulter
Counter operations manual.
Ralston Filter Sampling
The total filters used on the mlniplant are Type 30/25 Balston filters.
They use a 2.5 cm diameter by 18 cm long Grade BH filter tube. This filter
Is 99.95% efficient for 0.6 ym particles. The filter cartridge is stored in
Us plastic wrapping in a dry box to make sure 1t is dry. Shortly before
sampling, the cartridge 1s unwrapped and weighed on an analytical balance.
The weight is recorded, the filter support screen is Inserted, and the filter
assembled. When everything else 1s ready, the filter is installed for out-
side/In flow. The sampling is started soon after filter installation to pre-
vent unknown quantitites of gas and particulate from leaking through the
valve onto the filter. The combustor side of the valve 1s continually purged
with air or nitrogen to prevent plugging of the probe. If sampling must be
delayed after the filter is installed, the filter 1s pressurized and purged
to minimize leakage further.
To start the sampling, the purge flows are shut off, and the hot isola-
tion valve opened. The flow is set on the flow control system to the iso-
kinetic sampling rate. Temperatures, flow, and flow rate are recorded every
15-20 minutes. Most sampling after the tertiary cyclone requires 2 hours to
build up sufficient filter cake. Samples taken before the cyclone only
require 1 hour. After sampling, the high temperature isolation valve is shut
and both purges are again turned on. The filter 1s allowed to cool and is
removed from the system and taken into the lab.
Note: Coulter Counter is a registered trademark of Coulter Electronics Inc.
Isoton and Accuvette are registered trademarks of Coulter
Diagnostics Inc.
237
-------
In the lab, a one square foot piece of aluminum foil 1s tared and the
filter Is opened over the foil. The filter cartridge 1s carefully removed
and any loose particles are dumped onto the foil. The Inner support screen
1s removed from the filter cartridge and the cartridge 1s wrapped with the
foil. The cartridge 1n foil Is weighed on the analytical balance and the tare
weight subtracted to obtain a partlculate weight. The concentration 1s
obtained by dividing this by the total gas flow and multiplying by a conver-
sion factor. The cartridge with foil wrapping 1s then stored in a dry box.
Unless the filter appeared wet, the filter 1s rewelghed on a random quality
control check basis only,
Fine Particle Sizing by Coulter Counter
Most fine particle sizing 1n the miniplant 1s done with the Coulter
Counter Model TA11 with either 30 ym and/or 100 ym aperture probes. Material
with a fair fraction of particles larger than 45 ym 1s prescreened with a
sonic sifter through the appropriate screens. The two distributions are then
combined.
Equipment--
The Coulter Counter used during these studies is a Model TA11. Modifica-
tions have been made to decrease outside disturbances and increase precision
The instrument rests on a large well grounded metal plate. The sample stand'
next to the Instrument is completely enclosed in a Faraday cage constructed
of a lucite box covered with wire mesh. The sample stand and the Faraday
cage also rest on the grounded plate. The sample vacuum pump is outside
the cage, separately isolated and grounded. The electric stirrer motor sup-
plied with the sample stand was replaced with an pneumatic stirrer. This
equipment configuration has given very repeatable performance and has correla-
ted well with other units in other labs.
The electrolyte use in this study is Isoton II. The Isoton (1% NaCL In
distilled water) 1s continuously filtered through 0.45 and 0.2 ym millipore
filters so that it is ultra clean. This has achieved 12 second background
counts (30 ym aperture) as low as 10.
The sifter used is an ATM sonic sifter with an assortment of sieves
from 5600 to 45 ym (3-1/2 to 325 U.S. Mesh). Various screens are used to
bracket the expected size distribution.
Procedure—
A small piece of the Balston filter cake is carefully removed to avoid
contamination by the filter substrate. The filter cake or the minus 45 ym
(-325 U.S. Mesh) material from the sonic sifter is used as the sample. The
sample (~ 0.1 g) is placed in a small plastic vial of known volume called an
Accuvette. A cationic (Type 3A) dispersant is added (~10 drops) until the
solid is wetted. Ultra clean Isoton II is added to fill the Accuvette half
way. The mixture is placed in an ultrasonic bath for 5-15 minutes. While
the Accuvette is 1n the ultrasonic bath, the Coulter aperture and sample
beaker (500 ml) are rinsed several times with ultra clean Isoton II. Several
rinses are required to reduce the background (no sample) count to less than
100 in 12 second accumulation with the 30 ym aperture (less than 30 with the
238
-------
100 ym aperture). After a low background count 1s achieved, the Isoton II is
retained in the sample beaker and several drops of sample are placed in the
beaker. Sample is added while stirring until the proper sample concentration
1s reached (-2,5 to 55»). Coulter accumulation and operation from this point
are the same as in the operations manual.
Magnetic Particulates—
The above procedure works well for most ordinary PFB particulates.
Occasionally particulate samples are found that are fairly magnetic. The
test for magnetism 1s carried out with a simple pocket magnet. When magnet-
ism is evident, further standardization of technique and care must be exercised
to obtain a true distribution. Failure to dilute the sample properly may
lead to magnetic particle agglomeration and settling. The idea is simply to
deaggl omerate and dilute the sample enough that agglomeration due to magnet-
ism is negligible.
For this procedure, a working sample of 0.100 g is used. This sample
1s wetted with 25 drops of dispersant (Type 3A) in a 500 ml beaker. This
sample is diluted with 300 ml of ultra clean Isoton II. The beaker with its
contents is placed in the ultrasonic bath for 5-15 minutes. Once the
sample is well dispersed, a 25 ml aliquot is taken with vigorous mixing of
the sample. This aliquot 1s added to the 440 ml of ultra clean Isoton II in
the 500 ml beaker on the sample stand. This is sufficient for a good accum-
ulation with the Coulter Counter. The approximate concentration of the
sample is 18 micrograms sample/ml.
This procedure has been found to give repeatable size distribution. It
Is not as sensitive to slight delays in sampling as the other less rigid
procedure. Successive Coulter Counter accumulations do not tend to finer
size distribution as happened with the other procedure.
Balance Calculations,
Some flue gas streams such as the second cyclone inlet cannot be sampled
for particulates directly with a filter. A calculation and sampling procedure
has been developed to obtain particle size and concentration information by
mass balances. These calculations have been used to determine second cyclone
Inlet and outlet concentration as well as second and third cyclone efficien-
cies.
Usually in the operation of the miniplant, Balston filter samples of the
flue gas after the third cyclone are taken at planned intervals. Lock hopper
dumps of the cyclones during or immediately after the filter sampling period
are also sampled. The Balston filter catch, second and third cyclone dump
samples are analyzed for particle size distributions. The amount of material
1n the lock hopper is averaged over 3-5 of the bihourly dumps. In this way,
lock hopper hang up and small inconsistencies are removed. The various steps
of the differential size distribution which is obtained from the size analysis
are multiplied by the average dump weight and divided by the gas flow rate.
The various steps of the Balston filter differential size distribution are
239
-------
multiplied by the participate loading measured. All size distribution steps
are on the same weight/flow basis and a size differentiated mass balance may
be completed starting with the third cyclone. Once concentrations in all
streams are known, cyclone efficiencies may be calculated directly. The only
assumptions in this calculation procedure are that loadings are fairly con-
stant and that there is no change in particle size through the flue gas
system.
240
-------
APPENDIX E
ANALYTICAL TECHNIQUES
analysis of SoTids
Sol Ids from combustion and regeneration runs were analyzed for SO^2,
CO-?-2, Ca+2, Mg+2, Na+, carbon and total sulfur. The analytical techniques
that were used are described below.
SO,
CO
-2
-2
- The sample was treated with acidic
Ca"
Mg
Total
Sulfur
Total
Sulfur
- after Aug. 1978
Total
Carbon
BaCl2 solution.
was weighed.
• before Dec. 1978 -
The BaS04 precipitate
- HC1 was added to an acidified sample.
The solution was stripped with N£ and
the gas passed through drlerlte, CuS04
and ascarite. CO?'2 was determined from
the weight gain of the ascarite.
- after Dec. 1978
- before Aug. 1978 -
- before Aug. 1978 -
The sample was digested by heating
vigorously in a medium of perchloric
add/nitric acid. The determination of
Ca, Mg and Na was made by atomic
absorption.
The sample was fused with NagCOo at
950°C, then dissolved with HC1. The
determination of Ca and Mg was made by
atomic absorption.
(Dletert Sulfur Method) - The sample was
combusted in an oxygen atmosphere at
1250°C. The S02-S03 products in the
effluent gas were analyzed by an automatic
Leco tltrator.
The sample was added to a ^2®$ catalyst
and combusted in an oxygen atmosphere in
an Induction furnace at 1650°C. The
formed 1s selectively measured by an
Infra-red detector. A Leco IR-23 sulfur
analyzer was used (ASTM D-1552).
(Carbon on Catalyst Method) - The sample
was combusted 1n an oxygen atmosphere at
1200°C. The COg evolved was determined
from the weight gain of ascarite.
241
-------
Total - after Aug. 1978
Carbon
Analysis of Flue Gas by
Wet Chemical Analysis
S03 - before July 1978
S03 - after July 1978
SO,
The sample was combusted in an industion
furnace at 1650*C followed by removal of
sulfur, conversion of any CO to C02,
trapping of H£0 and trapping of the C02
on molecular sieves. In the next cycle,
the molecular sieves are heated to 300°C
where the C02 1s expelled and measured by
thermal conductivity detection. A Leco
WR-12 carbon analyzer was used.
The amount absorbed by an 80% Isopropanol
solution was determined titrimetrlcally
using 0.01N barium perchlorate as the
titrant and thorin as the indicator.
SOs was collected as H2S04 by using a
Goksoyr-Ross controlled condensation coil
(maintained at 60°C above the water dew
point). The amount of SO? is determined
by $04 titratlon with O.OlN barium per-
chlorate as the titrant and thorin as
the indicator.
The amount absorbed by a 3% hydrogen
peroxide solution was determined titri-
metr.ically using 0.01N barium perchlorate
as the titrant and thorin as the Indica-
tor.
242
-------
APPENDIX F
COMPARISON OF SOe MEASUREMENTS
BY UV, GC, WET CHEMISTRY
S0£ concentration 1n the mini plant flue gas was continuously analyzed by
a UV Instrument, Periodic measurements of SOe concentration were also made
by analysis of grab samples by the standard wet chemistry method (barium
perchl orate to thorln end point) and by GC. The results from the UV analysis
were consistently lower than the wet chemistry and GC results. The UV results
averaged 14% lower than wet chemistry and 29% lower than GC results. Assuming
the UV measurements were truly low by 14%, the Impact on the measured $03
retention results would only be two percentage points at the 85% S02 retention
level and one percentage point at the 95% S02 retention level. This degree
of bias, if it did occur, would have a negligible effect on the estimate of
dolomite requirements needed to retain 85 to 95% of the S02.
The difference between UV and GC results was not considered Important,
since the wet chemistry method is generally regarded as the standard method
for SOa measurement. Therefore, it would be concluded that the GC results
are 15% higher than the wet chemistry results. A comparison of UV, wet
chemistry and GC results 1s given in Table F-l .
243
-------
TABLE F-l. COMPARISON OF S02 MEASUREMENTS
BY UV, GC, WET CHEMISTRY
UV
72
62
89
131
23
106
90
85
269
279
49
602
201
140
140
82
54
3
5
7
360
150
108
319
220
310
59
322
228
320
10
14
132
28
412
23
98
210
we
57
79
-
136
24
-
-
104
251
315
.
_
-
191
.
_
-
14
.
-
437
195
180
323
256
335
116
324
208
305
8
3
88
146
406
64
327
210
GC
—
.
118
_
.
127
137
.
_
_
307
973
257
_
188
103
55
_
3
7
_
_
-
_
-
_
-
-
-
-
_
-
—
_
_
.
_
_
% Difference
Run S02 (POT)
No,
69
62 79 - -27
-33
70 131 136 - -4
-4
-20
-52
71 85 104 - -22
+7
-13
-527
-62
-28
72 140 191 - -36
-34
-26
-2
73 3 14 -367
+40
0
78 360 437 - -21
-30
-67
79 319 323 - -1
-16
80 310 335 - -8
-100
81 322 324 - -1
99 228 208 - +9
100 320 305 - +5
102 10 8 +20
+78
103 132 88 - +33
-421
+1
-178
-234
0
244
-------
TABLE F-l (CONT'D). COMPARISON OF S02 MEASUREMENTS
BY UV, GC, WET CHEMISTRY
% Difference
Run
No.
104
105
UV_
134
23
44
S02 (DPm)
WC__ GC
158
36
17
UV-WC
UV
-18
-56
+61
UV-SC
u\r
— _
—
--
Averages 14+49(1) 29+20(2)
(1) Three Points Excluded * -367, -421, -234%
(2) One Point Excluded - -527%
245
-------
APPENDIX G
RESULTS OF FLUE GAS S03 ANALYSES
Run
No.
nj^^ ,
62
65
67
v •
69
w*'
70
/ w
71
/ 1
72
73
I **
74
75
/ *»
76
« »
$03 Analytical Concentration
Method (DOTI)
Method 8
Method 8
Method 8
Method 8
Method 8
Method 8
Method 8
Method 8
Method 8
Method 8
Controlled Cond.
1.8
0.9
34.2
0
19.6
26.5
3.6
3.6
24.4
10.6
10.0
30.0
213
2.5
2.5
0
23
0.1
Run
No.
78
79
80
81
99
100
102
103
104
105
503 Analytical
Method
Controlled Cond.
Controlled Cond.
Controlled Cond.
Controlled Cond.
Controlled Cond.
Controlled Cond.
Controlled Cond.
Controlled Cond.
Controlled Cond.
Controlled Cond.
Concentration
(ppm)
0.6
9.9
12.8
0,4
1.1
11.1
3.2
0.3
6.8
28.1
73
0.7
0
0
0
21.3
1.7
6.7
29.8
0.5
0
0
Averages 0)
All Points
Method 8 Results -
Controlled Cond. -
8.6 + 11
12 Tl2
6+9
ppm
ppm
ppm
Two Samples Excluded - Run 71 (213 ppm), Run 100 (73 ppm)
246
-------
APPENDIX H
REGENERATOR GC ANALYSES
Run/
Sample
No.
102 #1
#2
#3
#4
103 #1
#2
#3
#4
#5
#6
#7
#8
#9
105 #1
#2
#3
#4
#5
#6
#7
H2S
(ppm)
15
9
< 1
75
<0.1
230
12
2
2
< 1
< 0.1
< 0.1
< 0.1
< 0.1
115*
77
180
—
100
93
Dr
CS2
(ppm)
< 1
< 1
< 1
-------
APPENDIX I
MASS SPECTROGRAPHIC ANALYSES OF SOLID SAMPLES OF RUN 69
El ement
Li
Be
B
F
Na
Mg
Al
Si
P
S
Cl
K
Ca
Sc
T1
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
Ga
Ge
As
Se
Br
Rb
Sr
Illinois
Coal
Cone.
ug/g
50
3
400
< 2
5000
-1%
-10%
~8%
300
-7%
2000
~1%
3000
< 20
210
130
130
200
-7%
30
200
70
30
30
30
10
< 5
0.3
35
70
Filter
Part
Cone.
ug/g
30
3
125
50
3500
~12%
-2%
~3%
120
~2%
40
-1%
~12%
< 20
3000
200
80
40
6000
5
400
35
10
8
15
3
< 1
0.5
70
2000
Pfizer
Dolomite
Cone.
ug/g
20
0.01
5
34
350
~12%
500
2000
0.2
200
120
300
-12%
< 2
15
2
12
40
180
3
3
4
10
0.2
0.1
8
< QJ
5
2
150
Initial
Bed
Cone.
ug/g
10
0.5
100
17
1000
-12%
-1%
-2%
120
-2%
40
1000
-12%
< 2
300
30
40
40
6000
2
100
20
20
2
15
8
< 1
2
10
150
Final
Bed
Cone.
ug/g
30
0.5
100
30
3500
~12%
-1%
-2%
120
~Z%
40
2000
-12%
< 2
400
30
40
40
6000
1
100
20
10
2
15
2
< 1
2
10
150
Bed
Overflow
Cone.
ug/g
10
0.8
100
17
1000
-12%
500C
-2%
120
-2%
80
500
-20%
<2
300
40
120
120
6000
2
100
35
4
3
40
8
<0.03
2
10
300
2° Cyclone
Cone.
ug/g
10
2
100
17
1000
-12%
-2%
-2%
120
-2%
160
3000
-12%
< 2
1000
100
40
40
6000
3
100
35
10
3
10
8
< 0.03
2
34
300
Bed Probe
Cone.
ug/g
10
0.8
100
17
350
-16%
-1%
-1%
60
-2%
80
2000
-20%
< 2
1000
40
20
40
6000
2
100
35
10
3
60
24
< 0.05
2
7
300
-------
ISi
•>
IO
Element
Sr
Y
Zr
Nb
Mo
Ru
Rh
Pd
Ag
Cd
In
Sn
Sb
Te
I
Cs
Ba
La
Ce
Pr
Nd
Sm
Eu
Gd
Tb
Dy
Ho
Er
Tm
Yb
APPENDIX I (CONT'D)
MASS SPECTROGRAPHIC ANALYSES OF SOLID SAMPLES OF RUN 69
Illinois
Coal
Cone.
yg/g
70
14
20
< 7
12
< 1
< 0.4
< 1
< 0.3
2
< 0.2
1
4
< 0.5
2
2
100
12
20
4
12
1
0.5
< 2
0.4
3
0.3
1
1
1
Filter
Part
Cone.
yg/g
2000
30
140
14
100
< 0.2
< 0.1
< 0.4
< 0.2
< 0.5
< 0.2
0.8
0.4
< 0.1
< 0.07
10
360
20
60
4
25
< 2
1
3
0.5
5
1
1
< 0.2
< 1
Pfizer
Dolomite
Cone.
yg/g
150
0.4
0.1
0.1
2
< 0.2
< 0.04
< 0.4
< 0.1
< 0.2
< 0.2
0.8
0.2
< 0.05
< 0.07
0.1
4
0.4
1
0.05
0,3
< 0.2
< 0.03
< 0.2
< 0.05
< 0.2
< 0.05
< 0.2
< 0.06
< 0.05
Initial
Bed
Cone.
ug/g
150
4
8
1
8
< 0.2
< 0.1
< 0.4
< 0.04
< 0.2
< 0.3
0.8
0.4
< 0.1
< 0.07
2
60
6
8
0.3
0.1
< 1
< 0.2
< 2
< 0.5
< 0.5
< 0.1
< 0.3
< 0.06
<0.2
Final
Bed
Cone .
yg/g
150
4
40
4
8
< 0.3
< 1
< 0.4
< 0.04
< 0.2
< 0.2
2
0.4
< 0.1
< 0.07
3
60
6
8
1
1
< 1
0.3
< 2
0.5
1
0.1
0.4
0.1
0.2
Bed
Overflow
Cone.
yg/g
300
5
14
4
8
< 0.5
< 1
< 1
< 0.1
< 0.3
< 0.5
2
1
< 0.1
< 0.07
3
60
12
28
1
10
< 4
< 0.3
< 3
< 0.05
< 2
< 0.3
< 0.4
< 0.1
< 0.3
2° Cyclone
Cone.
yg/g
300
5
40
4
8
< 0.5
< 0.4
< 1
< 0.2
< 0.7
< 1
2
1
< 0.1
< 0.2
10
140
12
28
3
15
< 3
< 0.5
< 2
< 0.6
< 3
< 0.3
< 0.3
< 0.1
< 0.3
Bed Probe
Cone.
yg/g
300
5
14
4
16
< 0,
< 0,
1
07
< 1
< 0,
< 0,
< 0,
8
1
< 0,
< 0,
1
140
7
14
1
10
< 3
< 0.5
< 2
< 0.2
< 1
,2
,4
,1
< 0.3
-------
ro
en
o
APPENDIX I (CONT'D)
MASS SPECTROGRAPHIC ANALYSES OF SOLID SAMPLES OF RUN 69
El ement
Lu
Hf
Ta
W
Re
Os
Ir
Pt
Au
Hg
Tl
Pb
B1
Th
U
Illinois
Coal
Cone.
yg/g
0.4
< 5
<2
< 2
< 1
< 2
< 1
< 2
< 1
< 10
< 2
< 2
< 1
< 2
< 1
Filter
Part
Cone.
yg/g
< 0.2
< 3
< 0.2
0.3
< 0.2
< 0.2
< 0.1
< 0.2
< 0.07
< 2
14
32
0.2
20
12
Pfizer
Dolomite
Cone.
yg/g
< 0.03
< 0.6
< 0.2
<0.1
< 0.06
< 0.2
< 0.1
< 0.2
< 0.07
<1
< 0.05
0.1
<0.03
0.2
0.2
Initial
Bed
Cone.
yg/g
< 0.06
< 0.6
< 0.2
< 0.2
< 0.2
< 0.2
< 0.1
< 0.2
< 0.07
< 1
<0.05
0.1
< 0.03
1
1
Final
Bed
Cone.
yg/g
0.1
< 0.6
< 0.6
< 0.2
< 0.06
< 0.2
< 0.1
<0.2
< 0.07
< 1
< 0.05
0.1
< 0.03
4
4
Bed
Overf 1 ow
tone.
yg/g
< 0.2
< 0.6
< 0.2
< 0.2
< 0.2
< 0.2
< 0.1
< 0.2
< 0.07
< 1
< 0.3
0.1
< 0.03
4
4
2° Cyclone
Cone.
yg/g
< 0.2
< 0.6
< 0.6
< 0.2
< 0.2
< 0.2
< 0.1
< 0.2
< 0.07
< 1
1
14
< 0.2
5
4
Bed Probe
Cone.
yg/g
< 0.2
< 0.6
< 0.6
<0.2
< 0.2
< 0.2
< 0.1
< 0.2
< 0.07
< 3
< 0.3
0.1
< 0.1
4
4
-------
APPENDIX J
ro
Run No.
61
62.1
62.3
63
67.1
67.2
67.3
67.4
68
69
70
71
72.1
73.1
74.1(1
74.2(1
76
77
99.2
99.3
99.4
99.5
Total Mass
MINI PLANT COMPONENT MASS BALANCES
Weight Percent
S C Total Ca Reactive 02
92.12
91.93
93.39
97,77
91.64
96.05
81.02
91.19
102.83
100.27
111.31
102.69
125.74
105.14
118.07
117.83
100.49
102.17
84.43
80.10
87.93
86.29
100.34
100.97
102.05
101.15
99.52
99.62
99.80
99.71
100.65
99.33
99.39
99.83
99.80
99.38
97.68
98.89
99.59
99.91
91.75
93.17
94.80
96.08
77.07
33.83
293.19
227.28
82.94
102.83
94.56
90.37
155.56
111.35
102.50
123.92
78.03
125.89
66.39
68.20
97.54
128.51
83.26
67.00
62,82
87.18
87.89
88.90
90.96
98.13
109.82
110.60
86.39
98.27
105.17
104.94
126.16
117.37
157.59
122.13
136.44
136.45
119.23
115.26
88.55
92.57
110.97
S3. 77
313.29
173.99
__
__
96.63
95.83
115.42
104.81
123.27
89.09
88.03
122.35
67.59
88.50
68.13
125.22
133.60
126.76
75.73
63.61
65.93
93.43
Mg
191 .19
84.62
88.03
90.05
113.82
100.42
131.83
81.45
83.65
104.39
59.95
75.93
81.24
99.75
84.92
75.12
60.65
86.48
Solids
Inorganics
127
205
264
189
101
101
106
103
146
98,
100,
110,
69,
94,
70,
90.
83.
.87
.09
.21
.20
.50
.98
.55
,27
.42
,93
,43
83
56
95
61
59
08
111.77
-------
ro
in
ro
Run No.
Total Mass
100.2
100.4
102(2)
103.0(2)
103.1(2)
103.2(2)
103.3(2)
105.2(2)
Average
+ Is
93.88
96.58
99.88
99.86
99.89
99.79
99.87
100.60
98.79+
2.48
APPENDIX J (CONT'D)
MINI PLANT COMPONENT MASS BALANCES
Weight Percent
C Total Ca
61.44
62.25
101.70
120.16
105.91
74.32
47.65
101.16
101.16+
51 .32"
81.14
108.30
93.92
97.96
100.51
102.34
101.32
92.30
105.85+
17.33
110.71
44.75
174.92
118.27
118.94
198.85
70.44
65.22
111.90+
53.65"
Reactive 02
82.69
82.41
80.44
75.91
79.88
77.99
81.43
79.99
93.37+
13.18"
Pfizer Dolomite Used for All Runs Except as Noted.
(1) Grove Limestone
(2) Grove Limestone - Combined Combustor-Regenerator Operation
93.35
35.59
91.11 +
31.1
Solids
Inorganics
130.00
90.83
76.62
104.35
81.55
78.41
114.11 +
46.15"
-------
APPENDIX K
MINI PLANT RUN OBJECTIVES
Run N0t _ Major Objective
60 GBF - ejector evaluation (ejector replaced plunger
type blow back nozzle)
61 GBF - ejector evaluation
6& vary filter GBF - coarse alumina filter media/large Inlet
\nedia and retaining screens
Jblow back GBF . Specul1te filter media
6^ conditions
04 GBF - specul He/natural gas Injection used
6c GBF - Internal baffles vs. screens In filter bed
media retention
g6 DOE Fireside Corrosion/Erosion Test - 1st attempt
67 DOE Fireside Corrosion/Erosion Tests • 100 Hr
Shakedown
g3_75 High SOg retention
7g High S02 retention using fine particle dolomite
11 Testing of unit modifications for DOE materials
corrosion test program - bed sol Ids overflow system
7R a! DOE Fireside Corrosion/Erosion Tests, 78 (250 hours),
' 79 (100 hours), 80 (215 hours), 81 (170 hours)
92-96 Test Acurex ceramic bag filter
gy_98 Shakedown of auxiliary coal feed system
gg S02 response to step change 1n coal sulfur content
IQQ S02 response to step change In Ca/S ratio
IQ-J Shakedown of regenerator
102-103 Sorbent regeneration
104 Acurex - electrostatic predpltator
1Q5 Sorbent regeneration; comprehensive analysis
106 DOE Fireside Corrosion/Erosion Tests (265 hours)
107-108 Acurex mobile bag house
10g_il4 3° cyclone efficiency tests
•J15 Slipstream GBF - Exxon Mark IV
253
-------
APPENDIX L. MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
ro
in
Operating Conditions
Run Length, hrs.
Pressure, kPa
A1r Flow Rate, m3/m1n
Temperature Gradient, °C/m
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 UJ.
Ca/S Molar Feed Rat1o-Calc. ^'
Excess Air, %
Sorbent
Coal
Flue Gas Emissions
S0?, ppm
NO', ppm
CO, ppm
C0?, %
027 %
Results
SOg Retention, %
Ca Sulfatlon, %
Lb SOp/M BTU
Lb NO^/M BTU
PD • Pfizer Dolomite
CH • Champion Coal
GL = Grove Limestone
* * No analyzer
61
8/24/77
5.0
795
17.6
4.1
940
1.77
1.4
1.5
3.6
103
0.75
--
36.05
PD
CH
—
115
156
11.6
5.6
__
52
--
0.18
(1) Ca/S
62.1
9/1/77
2.67
795
17.6
8.1
929
1.77
1.8
—
3.7
109
1.15
--
37.7
PD
CH
—
160
*
11.9
5.8
—
29
--
0.24
determined based
feed system.
(2) Ca/S calculated based
bed material .
62.3
9/1/77
5.0
795
17.6
6.5
932
1 .77
--
1.8
4.2
103
0.00
--
38.3
PD
CH
126
89
*
11.7
5.3
87
35
0.28
0.14
upon settings
upon level of
63
9/15/77
10.5
790
17.6
8.2
931
1 .80
1.8
1.7
4.1
102
0.00
--
28.9
PD
CH
632
93
162
12.2
6.3
37
38
1.41
0.15
on coal and
sulfatlon of
64
9/29/77
10.5
850
15.6
5.5
939
1 .48
1.7
2.0
3.6
95.6
0.00
—
14.6
PD
CH
593
87
*
14.7
2.7
43
33
1.25
0.13
sorbent.
spent
-------
APPENDIX L (COHT'D). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
rsi
01
01
Operating Conditions
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m3/rnin
Temperature Gradient, °C/m
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-Calc.
Excess Air, %
Sorbent
Coal
Flue Gas Emissions
S0?, ppm
NO*, ppm
CO, ppm
CO,, %
02; %
Results
Retention, %
Ca Sulfation, %
Lb SO?/M BTU
Lb MT/M BTU
A
65
10/13/77
7
910-960
17.4-20.2
6.52
937
1.45-1.66
2.0
3.4
95
0.00
41.7
PD
CH
80-380
13-98
20-50
1-13
4.6
41
66
11/29/77
17
930
18
36
918
1.6
67.1
12/12-13/77
67.2
12/14/77
.6
.4
2.0
3.2
130
1.25
10
PD
ILL
370
20
205
12
1.9
89
0.68
0.03
28
930
19.8
40.1
915
1.7
3.3
118
1.25
1.8
19
PD
ILL
736
100
300
14
3.0
74
41
1.60
0.16
25
930
19.8
30.1
915
1.7
3.4
130
1.25
1.2
14
PD
ILL
725
65
160
15
2.5
77
66
1.43
0.09
67.3
12/15/77
10
930
18.7-19.3
31.5
875-915
1.6
3.2
125
1.25
1.5
14
PD
ILL
718
65
312
12
2.5
77
50
1.41
0.09
PD = Pfizer Dolomite
ILL = Illinois Coal No.
CH = Champion Coal
-------
APPENDIX I (CONT'D). M1NIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
°C/m
°C
Operating Conditions
Run Length, hrs.
Pressure, kPa
Air Flow Rate,
Temperature Gradient,
Avg. Bed Temperature,
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-Calc.
Excess Air, %
Sorbent
Coal
Flue Gas Emissions
S02, ppm
NO , ppm
CO, ppm
COp, %
fm M
Results
S02 Retention, %
Ca Sulfation, %
Lb SOp/M BTU
Lb NO^/M BTU
ILL » Illinois Coal No. 6
PD - Pfizer Dolomite
* CO Analyzer Malfunctioned
67.4
12/16/77
37
930
19.3
38
875
1.6
1.8
3.0
115
1.25
1.9
31
PD
ILL
739
105
200
12
74
38
1.60
0.61
68
3/8/78
12
930
18.4
16
946
1.7
1.8
2.8
127
2.0
1.6
17
PD
ILL
50
79
*
16
3.2
98
60
0.09
0.11
69
3/16/77
11
930
19.0
17.1
947
1.8
1.8
3.0
132
1.5
1.5
14
PD
ILL
197
56
*
16
2.6
94
64
0.37
0.07
70
3/22/78
11.75
930
20.6
28
936
1.9
1.8
3.0
127
1.5
1.4
23
PD
ILL
94
81
*
17
4.0
96
70
0.20
0.12
71
4/6/78
13
930
19.0
35
933
1.72
2.3
3.S
122
1.52
1.7
12.6
PD
ILL
61
38
*
16.9
2.4
98
56
0.12
0.05
-------
ro
in
APPENDIX L (CONT'D).
Operating Conditions
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m3/m1n
Temperature Gradient, °C/m
Avg. Bed Temperature, °C
Superficial Velocity, m/sec
Settled Bed Height, rr,
Initial
Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Rat1o-Calc.
Excess Air, %
Sorbent
Coal
Flue Gas Emissions
S0
N0
ppm
x. Ppm
CO, ppm
c> !
Jesuits
S02 Retention, %
Ca Sulfation, %
Lb SOo/M BTU
Lb NCT/M BTU
A
MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
-«!;lo 72-2 73-1 73-2
4/13/78 4/13/78 41978 4/19/78
9.3
915
20.4
11
943
1.84
3
116
,7
1.79
1.8
9.37
PD
OH
60
28
*
22.4
2.1
96
55
0.13
0.04
4/13/78
2.66
694
17.0
12
940
2.02
2.2
3.9
112
1.79
1.3
1.02
PD
OH
23
50
*
18.8
0.2
99
76
0.04
0.01
4/19/78
8.5
930
19.6
20
939
1.78
3
117
2
1
,8
14
5
9.12
PD
OH
4.0
51.0
*
17.9
2.0
99
67
0.01
0.07
3
930
19.6
13
946
1.79
1.7
2.7
119
2.14
1.8
5.43
PD
OH
6.0
32.0
*
17.5
1.2
99
55
0.01
0.05
74.1
4/28/78
4.5
930
19.2
40
933
1.68
2.2
3*7
113
7.6
6.8
16.5
GL
OH
1.0
79.6
*
19.5
3.3
99
14
0.002
0.13
OH « Ohio, Valley Camp
PD = Pfizer Dolomite
GL * Grove Limestone
* CO Analyzer Malfunctioned
-------
APPENDIX L (CONT'D). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
ro
VI
co
Operating Conditions
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m^/min
Temperature Gradient, °C/m
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-Calc.
Excess Air, %
Sorbent
Coal
Flue Gas Emissions
S02, ppm
NO , ppm
CO, ppm
C02, %
0 v
2
Results
S02 Retention, %
Ca Sulfation, %
Lb SO./M BTU
Lb N03/M BTU
74.2
4/28/78
5.75
930
19.5
27
937
1.71
2.2
.8
.2
112
4,
8,
18.9
GL
OH
31.4
62.7
*
19.0
3.7
98
11
0.07
0.10
75
5/12/78
9.0
922
19.5
30
936
1.71
2.2
2.5
3.8
115
3.65
4.1
17.0
GL
OH
173
50
*
15.2
3.4
90
22
0.34
0.07
76
6/1/78
77
6/8/78
12.3
922
15
936
1.32
1.5
2.3
2.0
87.3
1,0
1.2
13.3
PD
OH
422
43
*
16,0
2.8
74
60
0.91
0.07
7.5
912
17.5
41
943
1.53
2.2
2.0
2.3
92.0
1.25
1.2
27.3
PD
OH
446
47
*
14.3
5.1
70
60
1.05
0.08
78.1
6/19-20/78
34
925
21.2
60
895
1.81
2.0
3.8
101
1.25
34J
PD
ILL
286
45
*
12.5
5.6
88
0.81
0.03
* CO Analyzer Malfunctioned
GL « Grove Limestone
PD • Pfizer Dolomite
OH - Ohio, Valley Camp
ILL • Illinois Coal No. 6
* CO Analyzer Malfunctioned
-------
APPENDIX L (CONT'D). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
ro
tn
-------
APPENDIX L (CONT'D). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
ro
Operating Conditions
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m^/min
Temperature Gradient, °C/m
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-Calc.
Excess Air, %
Sorbent
Coal
Flue Gas Emissions
SO ppm
NO t ppm
CO, ppm
m t
n 2' I
Wo «*
Results
S02 Retention, %
Ca Sulfatlon, %
Lb SOp/M BTU
Lb N(T/M BTU
/\
78.7
6/26/78
24
925
21.2
35
936
1.77
M •»
_.
3.3
141
1.45
1.7
15
PD
ILL
78.8
6/27/78
24
925
21.2
33
937
1.88
..
w —
2,9
146
1.45
__
11.6
PD
ILL
78.9
6/28/78
24
925
21.2
33
938
1.77
— —
_.
2.9
145
1.45
__
15
PD
ILL
78.10
6/29/78
24
925
21.2
26
895
1.81
— _
2.0
3.8
145
1.45
--
34.21
PD
ILL
79
7/31-8/4/78
100
925
21.4
_-
936
1.88
2.1
1.8
3.1
132
1.45
—
24.2
PD
ILL
113
28
*
14
3.2
95
55
0.27
0.05
363
41
*
15.6
3.2
0.72
0.06
260
36
*
16
2.8
90
0.62
0.06
286
15
*
12.5
5.5
87
0.81
0.03
293
56
*
14.6
4.2
88
0.70
0.10
* CO Analyzer Malfunctioned
ILL - Illinois Coal No. 6
PD - Pfizer Dolomite
-------
APPENDIX L (CONT'D). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
ro
Operating Conditions
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m^/min
Temperature Gradient, °C/m
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-Calc.
Excess Air, %
Sorbent
Coal
Flue Gas Emissions
S02, ppm
NOX, ppm
CO, ppm
rn #
UU« , h
02f %
Results
S02 Retention, %
Ca Sulfation, %
Lb SOp/M BTU
Lb NO^/M BTU
80.1
9/13-16/78
72.3
922
21.4
—
929
1 .77
2*
.1
--
3.9
126
1.41
--
27.2
PD
ILL
217
78
*
13.0
5.0
93
__
0.40
0.10
80.2
9/16-22/78
142.6
922
21.5
--
910
1.76
--
1.3
2.1
130
1.41
1.6
17.8
PD
ILL
152
121
*
12.4
3,5
95
59
0.27
0.16
81
10/9-16/78
171
912
21.4
—
933
1 .80
2.1
1.1
1.9
113
1 .56
1.3
28.5
PD
ILL
296
86
*
13.2
5.1
90
69
0.61
0.13
83
11/28/78
9.9
900
20.6
__
920
1.83
__
1.8
3.0
119
1.45
__
37
PD
ILL
— —
120
__
11.5
6
•» «•
__
0.22
84
11/29/78
6.8
900
20.5
921
1.7
— —
2.8
4.6
128
1 .45
• tm
23
PD
ILL
115
92
14.1
4.2
95
y >j
0.28
0.16
* CO Analyzer Malfunctioned
PD * Pfizer Dolomite
ILL • Illinois Coal No. 6
-------
ro
at
PO
APPENDIX L (CONT'D). MINI PLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Operating Conditions
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m-Ymin
Temperature Gradient, °C/m
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-Calc.
Excess Air, %
Sorbent
Coal
Flue Gas Emissions
S02, ppm
NO , ppm
CO , ppm
C0?, %
02, «
Results
85
11/30/78
6
900
21 .0
--
918
1.7
--
2.8
4.6
132
1.45
--
21
PD
ILL
146
100
__
--
4.1
86
12/1/78
9
900
20.8
--
921
1.7
--
2.9
4.7
125
1.45
--
24
PD
ILL
50
100
--
13
4.3
87
12/4/78
11.3
900
20.8
--
918
1 .7
--
2.9
4.7
122
1.45
— -
18
PD
ILL
150
105
--
16
3.7
88
12/5/78
9.5
900
20.6
~
918
1.7
--
2.0
3.3
115
1.45
--
28
PD
ILL/V
145
105
__
11.4
5.1
89
12/6/78
10.7
900
19.2
-_
917
1.7
--
2.2
3.6
101
1.4
— —
23
PD
V/CH
155
150
—
14
4.8
S02 Retention, %
Ca Sulfatlon, %
Lb SO?/P BTU
Lb NO:/M BTU
94
991
93
PD
ILL
Pfizer Dolomite
Illinois Coal No. 6
93
85
0.35
0.17
ILL/V
V/CH
0.12 0.38
0.18 0.19
« Illinois/Valley Camp
« Valley Camp/Champion
0.39
0.20
0.37
0.26
-------
APPENDIX L (CONT'D). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
-------
APPENDIX L (CONT'D). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
Operating Conditions
Run Length, hrs.
Pressure, kPa
Air Flow Rate, nr/mln
Temperature Gradient, °C/m
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 Rat1o-Calc.
Excess Air, %
5» Sorbent
* Coal
Flue Gas Emissions
SO ppm
NO , ppm
CO , ppm
C02, %
02 , *
95
12/15/78
5.5
900
18.1
__
914
1.6
--
1.8
3.0
98
1.5
__
21
PD
CH
290
102
90
8.9
4.2
96
12/18/78
21.0
900
17.9
927
1.6
_.
2.3
3.8
103
1.5
»_
36
PD
CH
250
100
40
14
6.2
97
1/23/79
5.5
912
17.7
64
902
1.53
2.1
1.8
3.0
116
*
_-
38.4
PD
*CH/ILL
**
124
**
11.7
6.3
98
1/24/79
6.25
922
18.1
64
904
1.55
—
--
3.1
109
*
--
37.3
PD
*CH/ILL
**
111
**
7.8
6.2
99.1
1/25/79
4.0
912
17.4
74
895
1.41
—
--
3.0
90.3
1 .40
—
34.9
PD
CH
203
120
435
9.7
6.0
ResuUs
S0£ Retention, %
Ca Sulfation, %
Lb SO?/M BTU
Lb NCT/M BTU
73
0.69
0.17
78
0.55
0.16
0.18
0.17
PD
CH
*
Pfizer Dolomite
Champion Coal
Switching between Champion
(Ca/S - 1.4) and Illinois (Ca/S
"CH/ILL
83
0.51
0.21
**
0.76) coals
Champion and Illinois Coals used - Champion
Analyses used when necessary.
S02 Response Test - No Steady Emissions
-------
ro
o»
en
APPENDIX L (CONT'D).
Operating Conditions
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m3/niin
Temperature Gradient, °C/m
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 Rat1o-Calc.
Excess Air, %
Sorbent
Coal
Flue Gas Emissions
S02, ppm
NO • ppm
CO; ppm
CO,, %
02, *
Results
S02 Retention, %
Ca Sulfation, %
Lb SO?/M BTU
Lb N(T/M BTU
A
MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
99.2 99,3 99.4 99.5
1/25/79
3.0
912
16.9
62
911
1.38
3,0
93.0
1 .40
1.2
33.9
PD
CH
194
125
61
10.0
5.9
85
70
0.46
0.21
1/25/79
3.67
912
16
56
910
1.37
,9
.2
3.3
99.4
0.76
0.95
30.1
PD
ILL
740
110
34
9.39
5.6
70
73
1.87
0.20
1/25/79
4.0
912
16.9
58
911
1.37
3.7
99.4
0.76
0.93
30.0
PD
ILL
721
113
15
11.3
5.4
70
75
1.82
0.21
1/25-26/79
7.0
912
16.8
63
913
1.37
3.9
82.8
1.40
1.1
33.1
PD
CH
171
121
11
10,
5.8
85
75
0.45
0.23
99.6
1/26/79
6.0
912
16.5
62
913
1.35
3.9
81.8
1.40
32.2
PD
CH
210
118
6
11.2
5.7
82
0.55
0.22
PD « Pfizer Dolomite
CH -- Champion Coal
ILL = Illinois Coal No. 6
-------
APPENDIX L (CONT'D). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
ro
°C/m
°C
Operating Conditions
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m^/min
Temperature Gradient,
Avg. Bed Temperature,
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-Calc.
Excess Air, %
Sorbent
Coal
Flue Gas Emissions
S0
N0
ppm
x, PPm
CO, ppm
C°2« 5
Results
S02 Retention, %
Ca Sulfation, %
Lb SOo/M BTU
Lb NO^/M BTU
99.7
1/26/79
6.83
912
16.5
61
913
1.34
2.4
4.0
81.0
1.40
1.0
30.6
PD
CH
183
120
11
12.24
5.42
84
84
0.48
0.23
100.1
1/30/79
8,0
912
17.6
61
914
1.45
2.1
85.2
1.43
0,90
41.9
PD
CH
360
135
83
14
6.3
67
75
0.93
0.25
100.2
1/30/79
4.0
912
17.6
61
914
1 .45
3.5
80.9
0.38
0.68
38.2
PD
CH
508
100
40
10
5.4
51
76
1.39
0.20
100.3
1/30-31/79
2.0
912
17.6
65
916
1,41
3.5
85.8
0.38
35.3
PD
CH
488
129
61
15.4
6.1
56
1.25
0.24
100.4
1/31/79
7.0
912
17.6
57
913
1.43
3.6
83.9
1.43
0,89
16.1
PD
CH
364
135
76
12.2
3.3
66
75
0.96
0.25
PD » Pfizer Dolomite
CH = Champion Coal
-------
APPENDIX L (CONT'D). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
ro
Operating Conditions
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m3/m1n
Temperature Gradient, °C/m
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 Rat1o-Calc.
Excess A1r, %
Sorbent
Coal
Flue Gas Emissions
SO ppm
NO , ppm
CO, ppm
C0?, %
02f %
Results
S0? Retention, %
Ca Sulfatlon, %
Lb SOo/M BTU
Lb NO^/M BTU
A
100.5
1/31/79
6.0
912
17.6
53
912
1.43
3.8
83.6
1.43
37.5
PD
CH
274
129
86
11.5
5.4
74
0.72
0.24
100.6
1/31/79
4.5
912
17.5
54
913
1.43
2.3
3.8
83.7
1.43
0.96
36.9
PD
CH
311
135
71
10.8
6.2
71
74
0.81
0.25
T02*
3/13-17/79
107
700
12.5
-15
908
1.43
2.1
1.0
2.2
69.5
1.5
19.
20.3
GL
CH
27
126
90
14.1
4.03
97
5
0.06
0.19
103.0*
3/29-4/1/79
81
700
13.4
-12
913
1.52
2.1
2.7
76.6
1.35
5.4
16.9
GL
CH
28
77
71
13.8
3.32
97
18
0.06
0.12
103.1*
4/1-3/79
49
700
13.5
-14
901
1.52
2.8
77.7
0.68
4.1
17.7
GL
CH
41
56
62
14.2
3.5
96
23
0.09
0.09
PD « Pfizer Dolomite
GL = Grove Limestone
CH • Champion Coal
* Combined Combustor-Regenerator Runs;
Combustor Data Only
-------
APPENDIX L (CONT'D). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
r\>
en
CD
Operating Conditions
Run Lenght, hrs.
Pressure, kPa
Air Flow Rate, m3/min
Temperature Gradient, °C/min
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-Calc.
Excess Air, %
Sorbent
Coal
FlMe Gas Emi s sions
S02, ppm
NO^. ppm
CO; ppm
Results
S02 Retention, 5
Ca Sulfation, %
Lb SO-/M BTU
Lb NO^/M BTU
GL a Grove Limestone
PD « Pfizer Dolomite
CH « Champion Coal
103.2*
4/3-4/79
22
700
14,2
-12
901
1.59
2.0
79.4
0.68
4.4
13.1
GL
CH
22
65
64
14.2
2.7
98
22
0.05
0.10
103.3*
4/4-6/79
59
700
14.4
-18
902
1.62
1.4
2.2
79.9
0.93
3.3
19.0
GL
CH
77
55
56
13.8
3.7
93
28
0.17
0.09
104
4/17-20/79
83.5
850
11.5
-42
899
1.62
2.1
1.8
3.0
93
1.49
1.4
42.1
PD
CH
71(5-160)
81
104
13.7
3.8
93(99-84)
66
0.16(0.01-0.37)
0.13
Combined Combustor-Regenerator Runs;
Combustor Data Only
105*
4/30-5/5/79
99
700
14.0
-56
894
1.46
2.2
1.7
3
77
1
29
0
45.8
GL
CH
70
59
77
12.5
4.1
93
23
0.15
0.09
-------
ro
o>
vo
APPENDIX L (CONT'D). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
106.1 106.2 106.3 107 108
Operating Conditions 5/29-6/10/79 5/29-6/10/79 5/29-6/10/79 6/18-23/79 6/26-29/79
Run Length, hrs. 43 5 217 108 71
Pressure, kPa 902 902 912 912 912
Air Flow Rate, m3/m1n 20.6 19.7 20.2 ie.5 16.4
Temperature Gradient, °C/m -64 -69 -" -71 -TOO
Avg. Bed Temperature, °C 914 904 908 872 877
Superficial Velocity, m/sec 1.71 1.62 1.64 1.31 1.31
Settled Bed Height, m
Initial 2.0 - -- 2.1 2.0
Final - -- J-? LI 0.96
Expanded Bed Height, m 2.4 3.3 *•' 1.7 1.6
Coal Feed Rate, kg/hr 139 145 '** 84.4 86.3
Ca/S Molar Feed Ratio-Set L35 1.19 1-25 1.25 1.25
Ca/S Molar Feed Ratfo-Calc. — ~ " ] «4 1 -2
Excess Air, % 15.6 5.8 9.0 31.4 23.0
Sorbent PD PD «> PD PD
Coal ILL ILL ILL CH CH
Flue Gas Emissions
S02, ppm 21.4 37.8 91.4 41.5 27.8
NO , ppm 47.5 42.4 64.1 26.4 44.6
CO, ppm 332 507 540 64.7 130.8
C0?, % 14.8 16.4 13.4 12.8 14.2
02. % 3.1 1.1 2.1 5.1 4.0
Results
S02 Retention, % 93 95 99 99 98
Ca Sulfatlon, % _. __ __ 72 79
Lb S02/M BTU 0.02 0.10 0.11 0.04 0.0
Lb NO^/M BTU 0.06 Q.06 0.08 0.13 0.08
PD « Pffzer Dolomite
CH - Champion Coal
ILL • Zllfnofs Coal No. 6
-------
APPENDIX L (CONT'D). MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
^J
o
Operating Conditions
Run Length, hrs.
Pressure, kPa
A1r Flow Rate, m3/m1n
Temperature. Gradient, °C/m
Average 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 Rat1o-Ca1c.
Excess Air, %
Sorbent
Coal
Flue Gas Emissions
S0?. ppm
N(£, ppm
CO, ppm
CO-. %
n *- #
* Near end of run Ca/S « 0,
109
7/25/79
9.5
915
17.6
—
870
1.65
—
..
2.5
87
1.25
—
19.5
PD
CH
70
140
175
15
3.5
sorbent
110
111
7/26/79 7/27/79
6.5
890
17.6
--
870
1.65
—
.-
2.8
98
1.25
--
39
PD
CH
50
110
175
15.5
6.0
hopper
** Auxiliary feed system used to switch coal
N.A. • Data not available.
PD " Pfizer Dolomite
CH • Champion Coal
ILL » Illinois Coal No. 6
8.0
900
17.6
—
870
1.65
--
-.
2.9
98
1.25
--
16
PD
CH
160
90-175
200
17.5
3.0
empty
feed.
112
7/30/79
6.5
900
13.1
--
870
1.22
—
--
3.4
79
1.25*
—
19.5
PD
CH
60(500)*
175
200
17.5
3.5
113
7/31/79
8.2
915
17.6
--
870
1.65
--
--
3.4
93
1.25(0.76)**
—
23
PD
CH/ULL)**
120(600)**
145
200
16.5
4.0
114
8/1/79
4.6
915
17.6
--
870
1.65
—
--
3.4
95
1.25
--
21.5
PD
CH
no
90
N.A.
16
3.8
115
8/7/79
6.7
915
17.6
--
870
1.65
--
--
3.5
99
1.25
--
25
PD
CH
50
N.A.
200
15
4.0
-------
APPENDIX M-l . PARTICLE SIZE DISTRIBUTION SPENT PFIZER 1337
DOLOMITE SORBENT (EXCEPT AS NOTED)
ro
Particle Size
Run No.
61
62
63
64
65
65
66
67
68
69
70
71
72
73
74(1)
75(1)
76
78
80
81
81
81
99
Material
Initial Bed
Final Bed
Initial Bed
Final Bed
Initial Bed
Final Bed
Final Bed
Final Bed
Final Bed
Final Bed
Final Bed
Final Bed
Final Bed
Final Bed
Final Bed
Final Bed
Final Bed
Bed Overflow
Dump No. 72
Bed Overflow
Initial Bed
Final Bed
Bed Overflow
Bed Probe No. 1
No. 3
No. 5
No. 11
No. 12
No. H
No. 32
5%
Less Than
___
230
800
550
500
420
240
...
300
300
175
96
100
115
600
420
150
105
130
760
620
125
820
600
510
-__
640
700
—
10%
Less Than
• • —
290
920
650
740
560
330
445
460
450
210
150
125
160
700
520
200
135
150
940
760
425
935
890
600
910
750
820
—
25%
Less Than
175
510
1080
820
900
770
700
740
650
850
300
290
185
280
810
780
350
210
205
1100
940
740
1200
970
830
1200
1025
1100
1100
50%
Less Than
1004
880
1360
1075
1200
1075
1325
1755
870
1200
480
640
360
620
1020
1000
690
430
530
1350
1300
1075
1550
1250
1100
1550
1325
1400
1400
(Vim)
75%
Less Than
1520
1230
1800
1400
1500
1450
1875
2150
1150
1630
980
1150
980
1100
1350
1250
1170
1000
1450
1600
1700
1450
1900
1600
1400
1850
1775
1750
1700
90%
Less Than
2000
1700
2040
1850
1900
1900
2200
2330
1560
2000
1380
1600
1400
1400
1700
1600
1500
2000
2000
1900
1950
1850
—
1950
1750
--
2000
--
2000
95%
Less Than
2200
1900
2200
2050
2100
2150
2300
2500
1800
3000
1700
1850
1700
1600
1900
1800
1780
5600
2500
2500
2350
2250
--
--
1950
—
--
--
—
(1) Grove Limestone Sorbent
-------
ro
^j
ro
APPENDIX M-2. PARTICLE SIZE DISTRIBUTION PRIMARY CYCLONE DIPLE6
Particle Size (ym)
Run
No.
61
62
65
67
81
85
88
5%
Less Than
.._
—
—
350
—
—
___
10%
Less Than
180
—
680
33
60
77
25%
Less Than
—
900
no
1375
54
80
97
50%
Less Than
140
1300
272
1950
105
no
140
75%
Less Than
185
1850
880
2175
170
180
190
90%
Less Than
890
2100
1400
2435
345
500
290
95%
Less Than
1250
2300
1750
2600
—
—
610
-------
APPENDIX M-3. PARTICLE SIZE DISTRIBUTION SECONDARY CYCLONE CAPTURE
CJ
Run/Sample No.
61
61
61
62
63
64
65
67
68.1
68.2
69
69.1
69.2
70.1
70.2
71
72.1
73.1
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
6
7
3
9
2
3
6
2
6
7
12
20
25
30
33
44
49
50
6
8.1
8.2
7
5
10
6
10
4
8.1
6
5%
Less Than
4.4
5.0
3.4
4.0
3.7
3.5
4.0
3.5
4.0
4.0
5.0
4.4
5.0
4.9
5.0
4.7
4.5
4.0
4.0
4.8
2.1
8.0
10.0
10.0
2.0
8.6
7.0
8.2
7.6
5.0
3.8
4.8
10%
Less Than
6.4
6.8
4.5
5.6
5.0
5.4
6.0
5.0
5.0
5.6
6.5
5.8
7.1
7.3
7.0
6.9
6.4
5.8
6.1
7.2
2.6
13.0
13.9
14.0
3.0
10.8
10.0
11.3
10.0
6.6
5.0
6.4
Particle Size
25% 50%
Less Than
11.4
13.0
6.9
10.5
9.4
10.0
11.0
9.8
9.8
10.0
11.5
10.0
12.5
13.0
12.0
11.5
10.5
10.5
11.3
13.0
4.3
26.5
35
32
6.5
19
16
30
17.5
10.5
7.6
12.5
Less Than
24
26
11
20
17
19
23
18
16
18
20
17
22
23
20
20
19
18
19
23
9.2
48
64
60
11
50
28
53
29
16
13
25
(ym)
75%
Less Than
42
48
18
50
36
38
49
30
30
35
35
31
39
38
35
34
31
34
33
39
16
84
100
100
17
69
50
59
50
25
25
43
90%
Less Than
60
70
25
70
73
62
120
50
58
56
54
50
57
59
49
50
47
51
47
53
24
—
112
no
24
119
90
86
160
35
40
64
95%
Less Than
90
90
31
105
113
105
230
60
80
86
70
67
100
80
60
60
55
69
63
70
32
--
120
115
30
230
200
130
215
50
54
100
-------
APPENDIX M-3 (CONT'D). PARTICLE SIZE DISTRIBUTION SECONDARY CYCLONE CAPTURE
r\j
Run/Sample No.
74.1 No.
74.2 No.
75
76
78
79
80
81
87
96
99
TOO
102
103
104
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
4
10
6
6
52
85
130
18
64
15
74
86
5
4
13
14
15
21
22
23
36
4
5
6
12
13
14
31
30
33
99
6
30
5%
Less Than
14
8.0
4.3
7.0
4.4
6.0
4.4
7.4
5.9
5.8
5.4
6.4
52
4.4
10.0
13.5
11.0
9.0
10.5
14.0
12.5
10.5
10.5
9.0
9.0
7.4
9.0
8.2
5.2
4.3
4.2
4.4
3.8
10%
Less Than
20
14.0
5.8
9.6
6.0
8.2
6.0
10.0
8.1
8.0
8.0
8.4
70
6.4
15.0
20
15.0
12.4
14.0
19
18
14.0
15.0
13.0
12.5
10.5
13.0
11.5
7.2
5.6
5.8
5.6
5.2
Particle Size (ym)
25% 50% 75%
Less Than Less Than Less Than
72
38
10
15
9.6
13
9.6
15
13
14.5
15
15.5
115
11.5
29
38
27
23.5
26.5
35
38
28
28.5
26
24.5
19.5
27
22.5
10.3
9.4
9.8
9.8
9.0
430
18.5
27
16
23
15.5
28
24
33
36
35
165
21
61
88
56
51
55
68
98
58
68
60
56
40
64
50
32
15
18
16
16
850
35
70
32
43
32
55
47
92
100
86
240
47
135
170
135
130
125
150
215
115
175
155
150
92
155
125
96
26
54
33
32
90%
Less Than
—
150
127
63
74
68
92
92
—
--
155
370
100
270
330
275
275
300
300
370
212
360
310
310
190
300
270
220
50
160
72
66
95%
Less Than
—
--
--
90
100
94
120
120
--
--
—
460
150
385
450
390
410
370
410
470
300
480
425
425
300
420
385
320
94
260
135
105
-------
APPENDIX M-3 (CONT'D). PARTICLE SIZE DISTRIBUTION SECONDARY CYCLONE CAPTURE
Run/Sample No.
105 No. 28
No. 44
106 No. 2
No. 11
No. 23
No. 57
107 No. 15
5%
Less Than
3.7
4.2
5.0
7.0
4.2
4.8
4.6
10%
Less Than
5.0
5.6
6.6
14.0
6.0
6.4
6.0
Particle Size
25% 50%
Less Than Less Than
8.5
10.0
10.5
34
11.5
11.5
10.8
14
17
18
84
23
23
21
(ym)
75%
Less Than
27
32
38
165
52
52
48
90%
Less Than
64
82
83
310
105
105
100
95%
Less Than
125
150
140
430
160
160
150
IM
->J
cn
-------
APPENDIX M-4. PARTICLE SIZE DISTRIBUTION TERTIARY CYCLONE CAPTURE
Run/Sample No.
67
68.1
68.2
71
72.1
73.1
74.1
74.2
5
76
78
79
80
81
86
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
7
12
20
25
30
33
43
48
6
8
4
8
6
4
10
6.1
6.2
6
31
52
85
120
1
18
42
50
64
15
74
5
Less
1
1
1
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
1
1
1
2
2
1
1
1
1
1
5%
Than
.30
.30
.80
.30
.55
--
.60
.60
.50
.50
.35
.15
.30
.20
.50
.25
.85
.60
.40
.80
.10
.65
.75
.60
.40
.20
.35
.20
.80
.60
.80
10%
Less
1
1
2
1
2
1
2
2
1
2
1
1
1
1
1
1
2
2
1
2
2
2
2
2
2
2
1
1
2
2
2
Than
.60
.75
.30
.66
.00
.15
.00
.10
.90
.00
.75
.40
.60
.50
.90
.60
.10
.00
.75
.30
.50
.20
.15
.10
.90
.90
.70
.40
.20
.10
.30
Particle Size (pm)
25% 50% 75%
Less Than
2.3
2.6
4.3
2.4
3.4
1.6
3.5
3.8
3.0
3.4
2.8
2.1
2.3
2.1
3.0
2.3
2.8
3.0
2.5
3.5
3.4
3.1
2.7
2.9
4.2
4.3
2.5
1.9
3.1
3.1
3.2
Less Than
4.0
4.7
7.4
4.2
6.4
2.5
6.6
7.2
5.2
6.1
4.9
3.5
4.0
3.4
5.0
4.2
5.3
5.0
3.5
5.1
4.8
4.5
3.4
4.4
5.9
6.1
3.9
3.3
4.8
4.7
4.7
Less Than
7.0
8.4
12.3
7.3
11.0
4.6
10.5
11.5
8.8
10.5
8.2
6.0
7.6
5.8
8.8
8.4
--
10.0
5.2
7.6
6.6
6.6
4.5
6.4
8.9
8.8
6.0
6.2
7.8
7.1
7.2
90%
Less Than
12.0
12.5
17.5
12.0
18
8.0
13.5
16
17
16
15.0
9.5
30
8.5
23
32
--
44
8.4
10.0
8.8
10.0
8.0
9.6
12.5
11.5
8.6
10.0
12.0
10.0
11.0
95%
Less Than
18
16
25
20
25
11.0
16
19
__
30
44
16
100
10.5
100
__
--
__
11.0
12.5
10.0
13.0
11.8
10.2
16
12.0
9.8
12.7
14.3
12.7
14.0
-------
APPENDIX M-4 (CONT'D). PARTICLE SIZE DISTRIBUTION TERTIARY CYCLONE CAPTURE
f\>
Particle Size (ym)
Run/Sample No.
87 No. 5
96 No . 4
No. 10
99 No. 13
No. 15
No. 21
No. 23
No. 34
No. 36
100 No. 2
No. 4
No. 6
No. 10
No. 12
No. 14
No. 31
102 No. 30
103 No. 33
No. 52
No. 75
No. 99
104 No. 6
No. 18
No. 30
No. 42
105 No. 28
No. 44
108 No. 5
5%
Less Than
1.75
2.10
1.45
1.40
1.42
1.30
1.42
1.30
1.45
1.13
1.42
1.10
1.13
1.30
1.26
1.20
1.40
1.30
1.15
1.05
1.10
1.10
1.35
1.20
1.43
1.60
1.45
1.50
1.59
1.45
1.59
10%
Less Than
2.30
2.70
1.80
1.80
1.75
1.59
1.70
1.59
1.70
1.35
1.65
1.40
1.35
1.70
1.45
1.45
1.70
1.50
1.40
1.36
1.41
1.35
1.70
1.40
1.70
1.95
1.78
1.80
1.90
1.80
1.90
25%
Less Than
3.4
3.7
2.6
2.6
2.4
2.2
2.4
2.2
2.2
1.8
2.2
1.9
1.8
2.2
2.0
1.9
2.2
2.1
1.9
2.1
2.1
1.9
2.5
1.9
2.3
2.6
2.5
2.6
2.6
2.5
2.5
50%
Less Than
5.1
5.3
3.8
3.9
3.3
3.1
3.4
3.1
3.3
2.6
2.9
2.7
2.6
3.1
2.8
2.7
3.0
2.8
2.6
3.4
3.3
2.8
3.6
2.8
3.3
3.9
3.7
3.9
3.4
3.4
3.3
75%
Less Than
7.0
8.0
6.3
6.3
5.2
5.4
5.8
4.9
5.3
3.8
4.2
3.9
3.8
5.0
4.0
3.9
4.4
3.9
3.8
5.5
5.0
4.5
5.4
4.4
5.0
6.3
5.8
6.6
4.8
4.9
4.6
90%
Less Than
11.0
12.0
11.0
10.0
8.4
8.8
9.8
7.6
8.8
5.8
6.3
6.0
6.3
8.0
6.8
6.0
7.0
6.3
5.9
8.5
8.0
6.9
8.0
6.9
7.3
9.8
8.8
9.8
7.1
7.0
7.1
95%
Less Than
13.0
15.0
15.0
12.0
10.0
10.0
12.0
9.6
11.0
7.6
8.4
7.8
9.2
10.0
9.4
7.8
9.2
8.0
7.4
10.0
10.9
9.2
10.0
9.1
8.9
12.0
10.0
12.0
8.8
8.8
9.4
-------
APPENDIX M-5. PARTICLE SIZE DISTRIBUTION AND GRAIN
LOADING - FLUE GAS PARTICULATES BEFORE TERTIARY CYCLONE
ro
~j
oo
/o \Grain Loading
Run/Sample No/ ; gr/SCF
96 BF-2
99 BF-4
99 BF-6
100 BF-1
100 BF-4
100 BF-6
100 BF-8
106 BF-2
106 BF-4
106 BF-1 4
106 BF-18
106 BF-21
106 BF-25
106 BF-35
106 Avg. (1)
106 Avg. (2)
107 BF-5
107 BF-7
107 BF-11
108 BF-2
108 BF-6
109 BF-2
109 BF-4
109 BF-6
110 BF-2
110 BF-5
110 BF-8
0.463
0.335
0.364
0.492
0.476
0.635
0.346
3.451
1.771
0.563
0.718
0.623
0.739
0.212
1.09
0.583
0.402
0.597
0.49
0.336
0.728
0.819
0.540
0.522
0.598
0.546
0.593
5%
Less Than
1.26
1.16
1.15
1.35
1.00
1.00
1.20
1.50
1.65
1.70
1.75
1.70
1.80
1.50
1.70
1.79
1.40
1.35
1.33
1.45
1.50
0.84
1.15
1.35
1.35
1.60
1.50
10%
Less Than
1.59
1.40
1.45
1.75
1.26
1.30
1.45
1.90
2.1
2.2
2.2
2.2
2.4
1.80
2.1
2.2
1.50
1.45
1.45
1.70
1.79
1.08
1.42
1.70
1.60
1.90
1.78
Particle Size (ym)
25% 50% 75%
Less Than Less Than Less Than
2.2
1.9
2.1
2.6
1.7
1.8
2.0
3.0
3.4
3.5
3.4
3.3
3.6
2.6
3.3
3.4
1.9
1.8
1.8
2.5
2.6
1.5
1.9
2.3
2.3
2.6
2.7
3.1
2.9
3.0
3.6
2.4
2.4
2.8
5.0
5.6
6.0
5.6
5.6
6.0
4.2
5.7
5.8
2.8
2.7
?.8
3.8
3.5
2.2
2.6
3.0
3.2
3.3
3.3
4.5
4.8
5.2
5.5
3.9
3.5
4.0
8.6
8.4
10.0
9.1
9.2
9.4
7.0
9.1
9.3
5.8
4.6
6.0
7.0
5.5
3.5
3.7
4.2
4.4
4.4
4.6
90%
Less Than
6.35
9.0
8.6
8.0
7.6
5.7
6.0
14.0
12.5
14.0
12.7
13.5
13.0
10.5
13.1
13.1
12.0
8.4
11.0
11.4
9.0
5.8
5.0
5.8
6.0
5.6
6.0
95%
Less Than
8.4
11.0
11.0
10.0
10.0
8.0
8.0
18
14.8
17.3
15.0
16
16
14.0
16.2
16.2
16
11.8
14.3
14.0
11.6
7.2
6.0
7.0
7.1
6.4
7.1
(1) All
(2) Excluding 2, 4
(3) BF indicates Balston filter samples taken upstream of the tertiary cyclone.
-------
APPENDIX M-5 (CONT'D). PARTICLE SIZE DISTRIBUTION AND GRAIN
LOADING - FLUE GAS PARTICIPATES BEFORE TERTIARY CYCLONE
Run/Sample No.
(2)
Particle Size (ym)
Grain Loading 5% 102 25% 5Q% 75% 90%
gr/SCF Less Than Less Than Less Than Less Than Less Than Less Than Less Than
ro
^J
10
111
111
112
112
112
113
113
113
114
114
115
BF-2
BF-5
BF-2
BF-5
BF-8
BF-4
BF-7
BF-9m
BF-2
BF-4U>
BF-2
0.399
0.494
0.675
0.545
0.500
0.784
0.373
0.760
2.12
2.22
0.237
1.1
i;2
1.6
1.4
1.2
1.1
1.4
1.4
1.3
1.4
1.4
1
1
1
1
1
1
1
1
1
1
1
.37
.5
.9
.7
.4
.4
.7
.7
.7
.8
.75
2.0
2.1
2.6
2.4
2.0
1.9
2.2
2.4
2.7
2.7
2.4
3.0
2.9
3.3
3.0
2.8
2.7
3.2
3.0
4.4
4.0
3.0
4.5
4.2
4.3
4.1
3.9
3.7
4.8
4.2
6.0
5.8
4.2
6.3
5.8
5.8
5.6
5.8
5.5
6.0
6.0
8.0
7.4
5.6
7.6
6.8
6.9
6.6
6.9
7.0
7.1
7.6
9.1
8.2
6.5
(1) Secondary Cyclone Disabled
(2) BF Indicates Balston filter samples taken upstream of the tertiary cyclone
-------
APPENDIX M-6. PARTICLE SIZE DISTRIBUTION AND GRAIN LOADING
FLUE GAS PARTICULATES AFTER TERTIARY CYCLONE
Particle Size (ym)
(3)Grain Loading 5% 10% 25% 50% 75% 90% 95%
Run/Sample No. gr/SCF Less Than Less Than Less Than Less Than Less Than Less Than Less Than
61 (1
62.1
62.3
64
65
67.1
rsj
CO
0 67,2
67.3
68.2
69.1
70.1
70.2
70.2
71
72.1
73
1}
(1)
(1)
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
No.
1
2
3
1
2
• 1
2
1
2
2
1
2
1
2
1
1
1
1
(1)
(1)
(1)
(3)
(3)
(2)
(2)
(2)
0.456
0.026
0.206
0.280
0.290
0.271
0.048
0.061
0.057
0.053
0.081
0.058
0.066
0.033
--
0.416
0.350
__
--
--
0.054
0.040
0.023
2.3
2.2
1.2
1.0
1.1
1.05
1.2
1.0
• V
•1 •
-_
_-
--
__
--
1.7
1.4
1.1
1.2
1.0
—
—
~
3.2
2.6
1.4
1.2
1.3
1.3
1.4
1.2
^^
^—
__
__
—
__
—
2.2
1.8
1.4
1.7
1.1
—
--
--
7.5
5.3
2.1
1.8
1.9
1.8
2.2
1.9
— —
1.0
1.1
1 .0
—
1.1
1.6
3.6
2.9
2.2
2.8
1.6
1.3
1.1
—
16
16
3.1
3.1
2.7
2.6
3.3
2.8
1.3
2.5
1.7
1.8
1.4
1.7
2.1
6.5
5.0
3.8
5.6
2.7
2,1
1,7
1.2
40
30
7.0
5.5
3.9
3.7
6.3
4.6
3.5
9.4
2.9
8.0
4.7
2.9
4.8
9.5
8.5
7.8
9.6
5.3
4.5
2.6
3.2
45
17
10.5
6.3
6.0
16
10.0
12.5
16.7
10.5
14.0
10.7
11.0
20
13.0
13.0
11.0
12.6
8.6
12.5
4.5
13.0
—
23
17
11
9.7
24
16
17
20
17
20
15
20
40
16
18
13
14
11
19
7
18
(1) Runs prior to Installation of 3rd cyclone. (2) Third cyclone bypassed.
(3) Runs prior to run 80, the sample no. refers to the order in which Balston filter samples
were taken with the Old Balston filter sampling system.
-------
APPENDIX M-6 (CONT'D). PARTICLE SIZE DISTRIBUTION AND GRAIN
LOADING - FLUE GAS PARTICULATES AFTER TERTIARY CYCLONE
Particle Size (ym)
Grain Loading 5% 10% 25% 50% 75% 90% 95%
r\>
oo
Run/Sampl
74.1
74.2
75 No.
No.
78 No.
No.
No.
No.
No.
No.
No.
No.
No. 1
No. 1
e No/1'
1
2
2
3
4
5
6
7
8
9
0
1
Avg of Test 2-11
79 No.
No.
No.
on DO i V ^/
oU DD~ 1
80 BA-2
80 BA-3
80 BB-4
80 BA-5
80 BB-6
80 BA-7
1
2
3
gr/SCF
0.024
0.024
0.024
0.031
0.033
0.034
0.028
0.039
--
C.042
0.056
0.045
0.036
0.054
0.041
0.012
0.0086
0.034
0.0143
0.0227
0.0257
0.021
0.023
0.021
0.028
Less Than
__
1.1
_..
—
0.48
0.6
0.62
0.46
0.48
0.48
0.79
0.74
0.6
0.79
0.58
0.54
0.47
0.47
0.52
0.55
0.54
0.53
0.56
0.46
0.58
Less Than
__
1.35
_ _
—
0.56
0.75
0.78
0.54
0.55
0.58
1.00
0.97
0.76
1.10
0.66
0.68
0.53
0.56
0.60
0.61
0.61
0.59
0.64
0.53
0.71
Less Than
1.15
2.0
1.10
1.15
0.74
1.10
1.10
0.74
0.79
0.74
1.49
1.47
1.30
1.75
0.98
1.00
0.69
0.75
0.69
0.76
0.78
0.70
0.91
0.70
1.05
Less Than
1.70
3.8
1.95
1.70
0.96
1.95
3.1
1.10
2.1
1.00
2.5
2.5
3.0
2.5
1.90
1.75
0.98
1.10
0.92
1.06
1.11
0.93
1.30
0.98
1.50
Less Than
2.5
8.0
5.5
9.3
1.4
4.8
6.8
2.8
5.9
1.5
5.5
5.7
5.7
3.7
4.6
3.7
2.4
1.8
1.2
1.9
2.0
1.3
2.0
1.41
2.4
Less Than
8.0
15.0
12.5
14.5
5.0
8.5
10.8
6.1
12.0
3.1
9.2
9.6
8.8
5.7
8.5
6.3
8.8
2.2
1.8
6.0
4.9
1.8
3.4
2.2
4.5
Less Than
13.5
20.0
16.0
18.5
8.7
9.8
12.5
8.4
14.0
5.04
11.5
11.5
11.0
8.0
11.0
8.8
13.5
8.0
2.2
9.0
9.0
2.5
4.9
5.0
6.3
(1) ~Runs~pr1or to run 80 the sample number refers to the order 1n which the Balston filters
were taken with the old Balston filter sampling system.
(2) Runs after run 80 BA refers to Balston filter samples taken with the HTHP sampling system;
BB refers to samples taken with the old Balston filter sampling system.
-------
APPENDIX M-6 (CONT'D). PARTICLE SIZE DISTRIBUTION AND GRAIN
LOADING - FLUE GAS PARTICULATES AFTER TERTIARY CYCLONE
Particle Size (ym)
Grain Loading 5% 10% 25% 50% 75% 90% 95%
ro
oo
ro
Run/Sample No.
80 BB-8
80 BA-9
80 BB-10
81 BA-4
81 BA-10
81 BB-11
81 BA-13
81 BA-16
81 BB-18
81 BB-20
83 BB-1
84 BB-2
86 BA-1
87 BA-1
89 BB-6
91 BB-3
96 BB-3
99 BB-1
99 BB-2
99 BB-3
99 BB-5
L; gr/SCF
0.0145
0.024
0.023
0.063
0.031
0.037
0.037
0.039
0.021
0.018
0.027
0.050
0.028
0.028
(2)
0.050
(2)
0.0459
0.049
0.036
__
0.094
0.055
Less Than
0.47
0.96
0.47
0.56
0.56
0.63
0.56
0.58
0.52
0.58
1.26
1.40
0.88
0.94
1.15
0.60
1.10
1 .20
1.00
1.00
0.74
0.63
0.96
Less Than
0.54
1.18
0.56
0.60
0.60
0.80
0.62
0.70
0.59
0.63
1.70
1.75
1.10
1.15
1.30
0.68
1.20
1.38
1.26
1.25
0.98
0.83
1.20
Less Than
0.73
1.45
0.76
0.79
0.76
1.20
0.77
1.00
0.72
0.80
2.20
2.20
1.50
1.50
1.70
0.92
1.46
1.78
1.65
1.63
1.40
1.20
1.60
Less Than
1.10
1.95
1.13
1.2
1.1
1.7
1.2
1.4
1.6
1.2
2.8
2.8
2.0
2.2
3.2
1.2
2.7
2.3
2.2
2.1
1.9
1.6
2.1
Less Than
1.6
2.8
1.6
2.8
1.7
3.4
3.2
3.1
5.8
2.9
3.5
3.5
2.9
4.6
8.6
2.0
6.8
3.1
3.7
2.5
2.5
2.4
2.8
Less Than
2.5
4.7
2.4
7.5
6.3
6.0
9.0
9.4
11.5
5.8
4.3
4.0
6.3
10.0
19.0
7.1
12.7
7.4
8.0
3.0
3.5
3.9
4.0
Less Than
4.0
6.0
5.0
11.5
13.0
8.0
14.0
16.0
16.0
8.0
5.0
4.9
10.08
13.0
28.5
11.5
20.2
12.7
11.0
3.5
5.0
5.4
7.0
(1) Runs prior to run 80 the sample number refers to the order in which the Balston filters were
taken with the old Balston filter sampling system.
(2) 100 jjm aperture used on Coulter Counter: all others used 30 \im aperture.
-------
APPENDIX M-6 (CONT'D). PARTICLE SIZE DISTRIBUTION AND GRAIN
LOADING - FLUE GAS PARTICULATES AFTER TERTIARY CYCLONE
r\j
oo
/, x Grain Loading 5%
Run/Sample No/' gr/SCF Less Than
100 BB-2
100 BB-3
100 BB-5
100 BB-7
102 BB-5
103 BB-4
103 BB-7
103 BB-11
103 BB-15
104 BB-1
104 BB-2
104 BB-3
104 BB-4
104 BB-5
105 BB-1
105 BB-2
105 BB-3
105 BB-5
106 BA-1
106 BA-3
106 BA-5
106 BA-6
106 BA-7
106 BA-8
106 BA-9
106 BA-1 3
0.059
0.060
0.048
0.049
0.025
0.044
0.075
0.029
0.099
0.057
0.099
0.117
0.084
0.140
0.077
0.074
0.072
0.106
0.083
0.029
0.049
0.037
0.460
0.048
0.094
0.80
1.10
0.63
1.15
0.84
0.63
0.75
1.00
0.60
1.00
1.20
0.79
1.15
1.00
1.00
0.64
1.05
0.88
1.26
0.79
0.74
0.64
0.79
1.10
1.70
0.79
0.79
10%
Less Than
1.35
0.79
1.40
1.10
0.75
0.96
1.26
0.72
1.26
1.40
1.10
1.35
1.17
1.26
0.80
1.30
1.17
1.49
1.05
0.93
0.84
1.05
1.40
2.20
1.15
1.10
Particle Size (ym)
25% 50% 75%
Less Than Less Than Less Than
1.75
1.15
1.85
1.43
1.05
1.3
1.6
1.0
1.6
1.8
1.4
1.8
1.4
1.7
1.1
1.7
1.5
1.9
1.4
1.2
1.1
1.4
2.2
?.1
1.6
1.7
2.2
1.5
2.4
1.8
1.4
1.7
2.1
1.6
2.1
2.2
1.9
2.4
1.9
2.2
1.5
2.2
2.0
2.4
2.1
1.7
1.7
2.0
3.5
4.9
3.1
2.8
2.7
2.1
3.1
2.3
2.1
2.4
2.6
3.2
2.65
2.9
2.5
2.9
2.4
2.8
2.0
2.6
2.5
3.0
3.3
2.4
2.8
2.8
5.6
6.9
4.5
4.3
90% 95%
Less Than Less Than
3.1
3.0
3.9
2.7
5.0
3.5
3.0
5.6
3.1
3.5
3.0
3.5
2.8
3.1
2.5
3.1
3.0
3.5
5.6
4.6
5.0
4.0
7.8
9.2
6.3
6.3
5.0
6.35
4.6
3.0
9.2
5.5
3.5
8.6
3.8
3.9
3.3
3.9
3.1
3.7
3.1
3.6
3.2
3.8
7.4
7.0
8.0
5.0
9.8
10.0
8.0
9.2
(1) Runs prior to run 80 the sample number refers to the order in which the Balston filters were
taken with the old Balston filter sampling systems.
-------
APPENDIX M-6 (CONT'D). PARTICLE SIZE DISTRIBUTION AND GRAIN
LOADING - FLUE GAS PARTICULATES AFTER TERTIARY CYCLONE
ro
oo
Run/Sample No.
106 BA-17
106 BA-20
106 BA-24
106 BA-28
106 BA-32
106 BA-34
Avg. (2)
Avg. (3)
107 BA-4
107 BA-6
107 BA-10
1C8 BA-1
108 BA-5
109 BA-1
109 BA-3
109 BA-5
110 BA-1
110 BA-4
110 BA-7
(1)
Grain Loading
5%
10%
Particle Size (ym)
25% 50% 75%
90%
95%
111
111
BA-1
BA-4
gr/SCF Less Than Less Than Less Than Less Than Less Than Less Than Less Than
0.055
0.063
0.054
0.047
0.043
0.043
0.056
0.050
0.099
0.117
0.121
0.118
0.129
0.118
0.112
0.170
0.105
0.061
0.070
0.084
0.079
0.56
0.64
0.56
0.58
0.59
0.60
0.75
0.65
0.74
1.05
1.00
0.79
1.05
0.93
0.80
1.00
0.64
0.56
0.82
0.69
0.58
0.62
0.79
0.65
0.65
0.68
0.70
0.95
0.80
0.98
1.35
1.30
1.05
1.30
1.26
1.21
1.30
0.77
0.60
1.20
0.84
0.68
0.8
1.2
0.9
0.8
0.9
1.0
1.3
1.1
1.4
1.9
1.7
1.4
1.7
1.7
1.7
1.7
1.0
0.8
1.7
1.2
0.8
1.2
1.7
1.3
1.2
1.3
1.4
2.0
1.5
1.9
2.5
2.2
1.9
2.2
2.3
2.1
2.1
1.4
1.3
2.4
1.6
1.3
2.4
2.7
2.7
1.9
2.1
2.1
3.1
2.5
2.5
3.2
2.9
2.5
2.6
2.9
2.6
2.6
2.0
2.8
3.2
2.4
2.0
4.9
4.9
5.5
4.6
4.6
3.8
5.3
4.7
3.1
4.8
3.8
3.4
3.0
3.9
3.2
3.2
3.2
5.7
4.0
3.6
3.3
8.0
7.2
8.0
8.0
7.0
6.3
7.6
7.2
3.6
8.0
5.0
4.0
3.4
5.3
3.7
3.85
4.3
7.6
5.1
4.8
4.7
(1) Runs prior to run 80 the sample number refers to the order fn which the Balston filters were
taken with the old Balston filter sampling system.
(2) Excluding 7
(3) Excluding 7, 8, 9, 13
-------
APPENDIX M-6 (CONT'D). PARTICLE SIZE DISTRIBUTION AND GRAIN
LOADING - FLUE GAS PARTICULATES AFTER TERTIARY CYCLONE
Particle Size (ym)
/,x Grain Loading 5% 10% 25% 50% 75% 90% 95%
Run/Sample No. ' gr/SCF Less Than Less Than Less Than Less Than Less Than Less Than Less Than
ro
00
en
112 BA-1
112 BA-4
112 BA-7
113 BA-3
113 BA-6
113 BA-8
114 BA-1
114 BA-3
0.067
0.132
0.203
0.109
0.072
0.118
0.133
0.172
0.60
0.72
0.56
0.62
0.68
0.79
0.68
0.78
0.68
0.90
0.64
0.73
0.78
0.98
0.82
0.98
0.9
1.3
0.8
1.0
1.1
1.3
1.2
1.4
1,4
1.7
1.3
1.4
1.4
1.7
1.7
1.9
2.4
2.3
2.4
2.6
2.0
2.2
2.6
2.9
4.9
3.4
4.2
5.2
3.3
3.9
4.6
4.9
6.4
4.3
5.4
6.6
4.8
5.2
6.5
6.4
115 BB-4 0.075 0.74 0.92 1.3 1.8 2.6 4.3 5.6
(1) Runs prior to run 80 the sample number refers to the order In which the Balston filters were
taken with the old Balston filter sampling system.
-------
APPENDIX N. MINIPLANT SOLIDS ANALYSIS
Run No,
61
Source
CO
at
62
Initial Bed
Final Bed
Second Cyclone No,
No,
GBP Filter 2
Filter 3
Filter 2
Filter 3
1
2
Dump No. 3
Dump No. 3
Dump No. 6
Dump No. 6
GBF Flange Fines
Participates - GBF Outlet
Bed Probe Sample No. 1
No. 2
No. 3
Dip!eg - 1st Cyclone
Initial Bed
Final Bed
Second Cyclone No,
No,
No,
No.
GBF Filter 2
Filter
Filter 2
Filter 3
Filter 2
1
2
6
7
Dump No. 1
Dump No. 1
Dump No. 2
Dump No. 2
Dump No. 7
Filter 3 Dump No. 7
GBF Flange Fines
Particulates - GBF Outlets
Bed Probe Sample No. 2
No. 5
No. 6
01 pi eg - 1st Cyclone
32.0
38.0
16,4
13.7
8.42
8.26
10.4
24.4
12.8
14.5
7.70
6.82
12.1
40.2
37.3
37.6
12.7
Weight Percent
S
7.21
10.50
4.24
3.34
4.44
3.83
4.65
4.51
4.26
2.41
7.04
7.23
8.68
5.45
10.0
10.51
3.24
3.06
2.83
3.09
5.67
5.20
4.70
4.20
4.51
5.33
5.60
5.98
9,10
11.0
11,0
5.67
S04
22.52
33.28
13.48
10.84
13.20
10.39
13.56
12.47
10.42
15.85
21.83
24.93
28.70
20.38
33.37
32.15
9.03
9,20
8.08
7.17
15.57
14.24
12.49
12.38
12.57
11.54
17.02
18.20
27.96
34.66
32.39
16.62
C03
24.06
0.62
11.27
4.05
3.73
3.94
1.42
0.84
4.14
1.17
27.60
0.54
0.71
2.52
19.91
0.53
4.78
2.17
0.32
0.28
2.89
4.20
4.33
3.93
0.68
0.50
1.53
0.01
1.71
0.65
1.78
1.44
Total C
5.01
0.28
5.62
3.05
1.59
2.14
1.44
1.70
1.63
1.19
5.96
0.35
0.37
1.51
5.46
0.42
3.50
1.87
2,21
3.16
0.98
1.49
1.72
1.86
1.52
0.85
1.75
2.78
0.37
0.58
0.32
0.61
1
4
7,
5,
5,
5,
5,
4.
4,
,83
,92
19
91
44
53
39
73
28
3.27
5.
2,
.27
.36
4.10
7.37
0,43
0.88
4.33
4.18
2.26
1.93
4.62
11.6
5.06
5.05
1.62
1.32
4.54
1.56
1.08
1.02
4,91
-------
Run No.
63
00
64
APPENDIX N (CONT'D),
Source
Initial Bed
Final Bed
Second Cyclone No. 3
No. 6
No. 9
GBF Filter 2 Dump No. 3
Filter 3 Dump No. 3
Filter 2 Dump No. 7
Filter 3 Dump No. 7
Filter 2 Dump No. 10
Filter 3 Dump No. 10
GBF Flange Fines Filter 2
Particulates
No.
No.
No.
Filter 3
1
2
3
Bed Probe Sample No. 2
No. 3
No. 4
No. 5
No. 6
Dip!eg - 1st Cyclone
Initial Bed
Final Bed
Second Cyclone No. 2
No. 3
No. 6
Dump No,
3
Dump No. 3
Dump No. 6
Dump No. 6
Filter 2 Dump No. 8
Filter 3 Dump No. 8
GBF Filter 2
Filter 3
Filter 2
Filter 3
MINIPLANT SOLIDS ANALYSIS
Weight Percent
Ca
37.7
36.4
7,
7,
4,
3,
2,
2,
2
2
1
62
15
93
34
81
15
,01
,32
,62
4.47
4.55
4.74
4.14
5.02
78.0
35.9
39.8
36.1
35.6
9.42
32.6
30.0
4.93
6.13
3.46
0.72
2.16
0.89
0.92
1.96
1.80
s
8.55
10.32
2.31
2.64
1.49
1.55
2.14
1.15
1.21
1.08
4.93
3.97
-
8.00
9.83
10.10
9.39
10.41
3.13
9.36
11.6
1.55
0.80
1.00
0.34
1.52
0.59
0.45
0.48
1.17
S04
27.36
33.96
6.13
4.25
4.18
6.47
4.14
3.98
2.82
3.30
2.87
11.12
7.66
-
29.07
30.97
32.54
33.63
34.23
8.02
28.90
34.16
5.27
4.77
3.24
0.61
4.28
1.00
5.03
4.71
1.39
C03
11.88
0.50
0.51
0.63
0.16
0.36
0.30
0.06
0.05
0.03
0.24
0.29
0.53
-
6.37
0.43
0.42
0.46
0.79
1.10
3.35
0.78
0.18
0.25
0.11
0.05
0.12
0.27
0.02
0.22
0.81
Total C
2.52
0.23
4.51
4.38
5.00
0.81
1.15
0.80
0.49
0.24
0.59
1.76
1.22
0.12
1.34
0.12
0.30
0.50
0.33
0.57
2.14
0.35
7.30
15.51
4.92
<0.05
1.14
0.18
<0.05
0.17
0.42
0.73
1.56
0.60
0.54
0.14
0.23
0.17
0.06
0.34
0.25
0.16
0.34
0.47
-------
APPENDIX N (CONT'D). MINIPLANT SOLIDS ANALYSIS
Weight Percent
ro
oo
CO
Run No.
64
(Cont'd,
65
67
67.1
Source
GBF Flange Fines No. 2
No. 3
Participates No. 1
No. 2
No. 3
Bed Probe Sample No. 4
No. 7
No. 9
Dipleg - 1st Cyclone
Initial Bed
Final Bed
Second Cyclone No. 2
No. 6
GBF Filter 2 Dump No. 3
Filter 3 Dump No. 3
Filter 2 Dump No. 7
Filter 3 Dump No. 7
Bed Probe Sample No. 3
No. 7
Dipleg . 1st Cyclone
Initial Bed
Final Bed
Dipleg - 1st Cyclone
Second Cyclone Nos. 5-9
Nos. 10-14
Bed Overflow Nos. 5-9
Nos. 10-14
Third Cyclone Nos. 5-9
Nos. 10-14
Bed Probe Sample
Bed Probe Sample
Ca
4.56
5.52
5.60
4.92
4.50
43.4
41.1
32.8
5.53
37.8
37.4
5.10
2.80
1.52
8.25
1.06
1.78
39.1
30.5
6.40
22.3
18.9
15.1
7.3
7.1
20.8
21.6
6.7
6.8
19.6
22.1
S
2.17
1.85
3.82
4.52
5.81
8.85
8.78
10.09
1.59
10.6
12.72
2.62
1.31
0.74
0.44
0.78
1.55
11.37
12.78
2.99
6.87
8.17
7.07
5,39
5.03
7.16
6.62
6.05
6.90
10.53
7.01
S04
3.23
8.19
9.28
15.32
24.88
28.56
31.24
33.56
4.83
28.32
35.95
6.26
3.27
1.64
1.15
1.78
3.01
33.42
35.51
6.38
14.54
24.82
17.95
14.01
20.84
21.08
14.13
19.85
20.16
15.72
24.71
CDs
0.03
0.02
_
—
—
0.45
0.27
0.52
0.08
7.95
0.11
0.52
0.10
0.22
0.10
0.01
0.07
0.53
0.62
0.01
.
12.36
9.38
0.45
0.51
17.19
15.77
0.03
0.01
17.80
20.27
Total C
0.46
1 .09
-
1.61
1.63
0.44
0.34
O.05
0.23
2.21
0.29
6.21
5.05
0.12
0.21
0.17
0.62
0.25
0.12
0.29
4.43
2.15
1.94
4.69
4.20
2.63
2.67
1.26
1.23
3.18
4.50
Mg
0.40
0.28
0.23
0.23
0.21
0.56
0.65
0.68
0.36
0.75
0.59
0.24
0.16
0.19
0.35
0.05
0.12
0.62
0.70
0.27
11.7
10.7
8.0
3.1
3.4
11.2
12.0
3.9
4.1
10.9
12.2
-------
APPENDIX N (CONT'D). MINIPLANT SOLIDS ANALYSIS
Run No.
67.1
(Cont'd.)
67.2
Source
ro
oo
67.3
67.4
Turbine Section No. 1
(Pressure Reduc. Section)
Turbine Section No. 4
(Turbine Housing After
Distribution Plate)
Second Cyclone Nos. 18-22
Nos. 23-27
Bed Overflow Nos. 18-22
Nos. 23-27
Third Cyclone Nos,
Nos
Bed Probe Sample
Bed Probe Sample
18-22
23-26
Second Cyclone Nos. 28-32
Bed Overflow Nos. 28-32
Third Cyclone Nos. 28-32
Bed Probe Sample
Second Cyclone Nos. 42-46
Nos. 47-51
Bed Overflow Nos. 42-46
Nos. 47-51
Third Cyclone Nos. 41-45
Nos. 46-50
Bed Probe Sample
Bed Probe Sample
Bed Probe Sample
Bed Probe Sample
8.4
8.4
20.6
20.1
6.9
6.8
16.3
20.9
7.3
23.1
6.1
23.3
4.3
6.9
22.1
22.1
6.7
6.9
23.5
20.4
22.1
23.1
s
4.04
9.18
5.26
5.44
10.19
10.19
8.02
6.19
8.07
8.73
4.99
8.16
8.02
8.74
5.17
5.78
7.45
6.63
9.30
8.78
6.75
11.84
7.99
6.38
S04
11.46
29.55
16.37
17.68
34.36
34.61
16.20
18.64
29.55
23.20
15.34
28.83
21.53
29.96
15.66
16.19
19.20
23.81
28.78
28.29
24.80
21.58
22.59
23.22
C03
0.58
0.39
0.84
0.49
12.80
10.00
0.13
0.08
15.74
20.62
0.51
17.86
0.21
16.44
0.49
0.48
24.12
21.95
0.09
0.04
21.53
22.43
21.38
18.87
Total C
0.33
0.18
3.43
3.66
1.78
2.47
2.26
1.16
4.39
3.73
4.01
4.44
0.40
3.50
4.78
4.53
4.53
5.07
1.06
1.15
4.91
2.46
4.31
5.26
4.3
4.2
12.1
10.9
3.5
3.4
11.9
12.1
3.6
13.7
3.6
13.5
2.1
3.4
12.8
12.4
5.0
4.9
13.4
11
12,
,9
.7
13.4
-------
APPENDIX N (CONT'D). MINIPLANT SOLIDS ANALYSIS
Weight Percent
Run No.
68
68.1
68.2
ro
69
Source
Initial Bed
Final Bed
Second Cyclone No. 6
No. 7
Third Cyclone No. 6
No. 7
Bed Probe Sample No. 6
No. 7
No. 8
Second Cyclone No. 8.1
No. 8.2
No. 9
No. 10
Third Cyclone No. 8.1
No. 8.2
No. 9
No. 10
Bed Probe Sample No. 9
No. 10
No. 11
No. 12
Initial Bed
Final Bed
Ca
^•^•^•••^H
22.6
21.6
15.8
14.9
11
11
.0
.1
23.8
24.2
23.7
17.4
18.7
18.5
17.6
10.8
11
12
13.1
25
22
21,
23.7
27.2
24.1
S04
30.50
37.12
23.19
21.70
24.01
25.90
38.80
33.26
32.98
25.68
24.96
24.92
25.54
24.42
24.95
25.00
21.20
35.36
33.84
36.05
32.73
29.53
38.16
C03
17.87
0.60
4.49
5.32
0.72
1.10
0.93
0.71
1.40
4.93
5.21
5.04
5.13
1.28
1.57
0.97
1.32
0.88
0.69
0.58
0.72
6.42
0.71
Total C
3.44
0.18
2.29
2.78
4.14
4.63
0.12
0.31
1.04
2.01
2.40
1.90
2.53
4.62
5.78
3.40
3.16
0.21
0.32
0.62
0.44
1.34
0.13
Mg
13.8
14.0
10.0
9.3
5.9
6.0
14.6
15.2
15.6
11.4
11.7
12.1
11.1
5.79
5.85
5.68
7.3
16.6
15,2
15.1
16.1
12.9
11.9
-------
APPENDIX N (CONT'D). MINIPLANT SOLIDS ANALYSIS
Run No,
69.1
69.2
70
70.1
Source
Ca
Weight Percent
Me
Second Cyclone No. 5
No. 6
No. 7
Bed Overflow No. 3
Bed Probe Sample No. 1
No. 5
No. 6
No. 7
No. 8
Balston Filter No. 1
Second Cyclone No. 9
No. 10.1
No. 10.2
No. 11
Bed Overflow No. 5
No. 6.1
No. 6.2
No. 7
Bed Probe Sample No. 10
No. 11
No. 12
Balston Filter No. 2
Initial Bed
Final Bed
Second Cyclone No. 6
No. 7
Bed Overflow No. 1
No. 2
Balston Filter No. 1
Bed Probe Sample No. 1
No. 8
No. 9
No. 10
15.6
16.5
15.0
18.0
22.9
23.5
22.4
15.1
19.3
14.1
13.4
14.0
13.7
10.5
23.2
23.0
22.1
20.0
18.7
21 .0
23.4
1 .99
24.7
24.6
16.2
17.7
21.9
26.5
13.8
25.5
32.8
24.4
26.2
27.02
26.40
28.78
35.76
26.29
40.38
40.44
38.36
31.77
44.74
27.61
9.43 27.94
9.70 26.79
24.92
34.29
12.19 36.88
10.94 33.05
32.42
40.20
12.92 40.28
39.02
24.48
31.12
33.68
27.72
29.78
36.19
29.04
30.56
37.45
35.25
37.93
31.39
-------
APPENDIX N tCONT'D). MINIPLANT SOLIDS ANALYSIS
Run No,
ro
<&
ro
71
Source
72
Initial Bed
Final Bed
Second Cyclone No.
Bed Overflow No.
Third Cyclone No.
Bed Probe Sample No.
Initial Bed
Final Bed
Ca
70.2 Second Cyclone No. 8
No. 9
No. 10.1
No. 10.2
No. 11
Bed Overflow No. 3
No. 4
No. 5.1
No. 5.2
No. 6
Balston Filter No. 2
Bed Probe Sample No. 11
No. 12
No. 13
No. 14
No. 15
^•^^•••M
15.3
11.7
15.1
16.2
16.8
16.1
19.9
18.0
21.7
24.6
11.9
24.4
22.7
24.0
23.9
20.5
). 4
9
, 4
No.
No.
No.
No.
1
5
6
18
25.3
22.4
14.9
22.0
10.6
24.5
25.1
23.0
22.3
21.1
18.9
Weight Percent
S04
28.28
28.10
28.22
28.63
27.57
32.06
31.71
35.75
35.84
33.43
27.52
33.44
34.69
29.39
35.30
39.57
27.70
31.95
27.60
32.94
23.58
33.74
31.85
33.44
34.51
39.55
33.96
C03
2.99
2.59
2.40
2.63
2.70
6.11
5.74
4.82
6.12
4.86
_
1.05
1.12
1.30
1.41
0.56
2.21
3.22
0.91
9.35
1.32
5.93
3.86
2.22
3.16
3.56
0.46
Total C
1.13
1.14
1 .00
0.98
1.04
1.40
1.25
0.98
1.28
0.97
0.42
0.31
0.27
0.30
0.33
0.13
0.53
0.69
0.96
2.00
1.09
1.36
0.88
0.57
0.81
0.95
0.15
Mg
9.92
7.85
10.5
9.59
9.40
9.09
10.8
12.3
12.7
13.5
6.43
13.7
13.4
13.3
13.2
12.2
14.5
13.0
7.74
12.1
4.71
13.0
13.6
12.1
12.9
13.8
12.8
-------
APPENDIX N (CONT'D). MINIPLANT SOLIDS ANALYSIS
Run No.
72.1
ro
us
co
72.2
73
73.1
Source
Second Cyclone No. 8.1
No. 8.2
No. 9
Bed Overflow No. 3
Third Cyclone No. 8.1
No. 8.2
No
Bed Probe Sample
. 8
. 8
, 9
No.
No.
No.
No.
1
8
9
10
Second Cyclone No. 12
Third Cyclone No. 12
Bed Probe Sample No. 12
No. 13
No. 14
Initial Bed
Final Bed
Second Cyclone No. 6
No. 7
Bed Overflow No. 3
Third Cyclone No. 6
No. 7
Bed Probe Sample No. 1
No. 5
No. 6
No. 7
No. 8
Ca
^^•M^MM^
15.8
16.2
19.1
20
11
11
11
Weight Percent
25.7
21.0
20.4
19.2
17.4
14.5
22.0
18.8
23.4
27.3
21.8
16.7
16.5
21.9
16.2
16.3
28.1
25.7
23.7
26.4
24.7
S04
27.07
27.60
27.17
27.83
22.64
23.30
20.93
26.90
34.34
37.38
29,65
23.23
15.98
36.65
38.65
32.94
30.29
34.75
26.55
26.33
35.41
24.97
24.78
34.70
24.81
30.73
30.23
35.20
COs
1.34
1.33
0.83
11.49
0.15
0.27
0.99
6.43
0.85
0.88
0.74
1.33
1.71
1.08
0.72
0.53
5.52
1.03
3.33
3.04
3.98
1.60
1.42
3.41
2.18
2.27
1.93
1.01
Total C
1.84
1.58
1.45
0.99
1.18
1.14
1.07
1.37
0.33
0.34
0.40
8.56
4.91
0.93
0.23
0.17
1.21
0.24
2.32
2.28
0.87
1.13
1.09
0.76
0.59
0.61
0.55
0.33
10.3
5.93
13.6
13.0
14.1
16.2
12.9
9.75
8.80
11.6
7,
7,
44
02
15.7
14.3
13.4
15.3
14.1
-------
APPENDIX N (CONT'D). MINI PLANT SOLIDS ANALYSIS
Run No.
73.2
Source
Weight Percent
74
74.1
ro
vo
74.2
75
Second Cyclone No. 11
Third Cyclone No. 11
Bed Probe Sample No. 11
No. 12
No. 13
Initial Bed
Final Bed
Second Cyclone No. 3
No. 4
Third Cyclone No. 3
No. 4
Bed Probe Sample No. 1
No. 4
No. 5
Second Cyclone No. 9
No. 10
Third Cyclone No. 9
No. 10
Bed Probe Sample No. 10
No. 11
No. 12
Initial Bed
Second Cyclone No. 5
No. 6
24.5
38.6
47.7
24.0
27.2
19.3
13.7
44.8
47.5
47.8
22.8
20.3
19.4
15.6
52.9
52.8
51.9
52.9
13.73
14.2
S04
16.88
19.86
33.63
33.04
31.87
19.62
19.7
14.7
14
17,
16.6
15.7
18.0
15.8
13.4
13.6
19.8
18.9
12.8
14.6
17.85
17.43
13.11
14.12
COs Total C
2.48
1.63
0.86
1.04
0.48
28.60
1.54
2.62
2.46
3.54
4.56
5.99
2.59
2.08
1.93
1.51
3.93
3.96
2.32
1.84
2.11
1.61
0.67
0.78
2.44
1.06
0.47
0.46
0.20
6.34
0.34
5.44
5.21
3.99
3.04
1.37
0.65
0.47
5.85
6.03
2.84
3.06
0.43
0.53
0.43
0.36
6.87
5.48
10.2
6.50
13.9
14.5
14.3
-------
APPENDIX N (CONT'D). MINIPLANT SOLIDS ANALYSIS
Run No.
75
(Cont'd.)
Source
76
ro
vo
01
77
78
80
Bed Overflow No. 5
Third Cyclone No. 5
No. 6
Bed Probe Sample No. 1
No. 6
No. 7
No. 8
Final Bed
Initial Bed
Final Bed
Second Cyclone No. 6
No. 7
Third Cyclone No. 6
No. 7
Bed Probe Sample No. 1
No. 7
No. 8
No. 13
Second Cyclone No. 6
Bed Overflow No. 4
Third Cyclone No. 6
Second Cyclone No. 72
Bed Overflow No. 72
Third Cyclone No. 72
Bed Probe Sample No. 8
No. 9
Second Cyclone No. 100
Bed Overflow No. 100
Third Cyclone No. 100
50.8
50.7
45.6
.3
.7
25,
29,
15.0
16.4
12.6
12.0
27.0
27.7
29.2
24.0
13.0
24.8
9.16
12.9
23.5
8.41
24.8
22.13
12.30
13.67
24.69
6.48
11.07
5.06
Weight Perceat
$04
24.76
19.19
18.69
16.98
19.23
20.99
19.04
24.72
39.63
40.83
16.53
16.28
14.70
14.07
38.44
40.13
39.14
38.16
20.87
36.00
17.87
9.14
39.10
15.16
39.41
29.53
20.19
31.93
15.12
C03
1.41
-
8.72
2.21
1.89
1.70
0.61
1.36
0.61
0.33
0.30
0.32
1.57
1.34
3.61
0.99
0.93
0.40
0.52
0.49
0.90
0.27
0.68
0.82
1.27
2.65
0.90
3.03
0.57
Total C
0.37
3.05
3.26
0.53
0.43
0.44
0.45
0.29
0.22
0.11
4.38
4.12
2.01
2.39
0.69
0.22
0.25
0.13
3.54
0.15
1.86
0.56
0.27
0.50
0.43
0.62
0.62
0.65
0.39
13,
11,
5(
5(
2
2
6
9
33
84
02
36
13.1
10.8
11.6
11.7
6.83
12.3
2.83
7.43
12.9
3.82
13.5
14.35
7.48
7.67
14.44
-------
APPENDIX N (CONT'D). MINIPLANT SOLIDS ANALYSIS
Run No.
81
Source
Ca
S04
Weight Percent
ro
to
99
99.2
99.3
Initial Bed
Final Bed
Second Cyclone No. 15
No. 51
No. 62
No. 74
No. 86
Third Cyclone No. 15
No. 51
No. 62
No. 74
No. 86
Bed Overflow No. 15
No. 51
No. 62
No. 74
No. 86
25.34
24.47
_
17.18
16.74
-
-
-
10.51
10,15
-
_
_
25.38
23.11
-
-
11.43
13.15
-
7.65
7.92
-
_
-
8.40
8.60
-
_
_
12.65
13.93
_
—
— '
37.24
40.87
32.10
24.16
_
^
—
26.01
26.44
_
_
—
37.78
42.50
—
«.
Dolomite Blend "A" 21.67
Second Cyclone No. 13 14.76 7.17 20.20
Third Cyclone No. 13 6.63 4.93 18.81
Bed Probe Sample No. 1 25.24 14.95 42.25
No. 2 24.73 14.33 42.43
Second Cyclone No. 14 12.94 6.30 19.54
No. 15 9.87 5.52 16.63
Third Cyclone No. 15 5.98 5.93 14.98
Bed Probe Sample No. 3 24.40 15.04 43.76
No. 5 25.28 13.50 44.34
C03
1.22
0.10
4.48
4.93
5,43
5.07
3.89
0.44
0.49
0,22
0.11
0.08
0.95
0.40
0.39
0.28
0.33
64.70
4.91
0.24
0.31
0.37
3.73
1.77
0.03
0.37
0.25
Total C
0.31
0.07
2.22
1.64
2.93
1.55
1.77
1.14
0.54
1.57
0.44
0.56
0.26
0.08
0.16
0.11
0.25
-
4.29
2.35
0.09
0.13
2.39
1.57
3.11
0.09
0.12
15.06
14.23
10.52
11.01
6.52
7.25
15.18
14.41
13.19
8.59
4.19
12.91
13.55
7.96
6.05
3.59
12.94
13.76
-------
APPENDIX N (CONT'D). MINIPLANT SOLIDS ANALYSIS
ro
UD
Run No.
99.4
99.5
Source
Ca
Weight Percent
99.7
100
100.1
100.2
100.4
Second Cyclone No. 2
Third Cyclone No. 21
Bed Probe Sample No. 11
Second Cyclone No. 22
No. 23
Third Cyclone No. 23
Bed Probe Sample No. 12
No. 14
Second Cyclone No. 36
Third Cyclone No. 36
Bed Probe Sample No. 32
Dolomite Blend "A"
Second Cyclone No, 4
No. 5
Third Cyclone No. 4
Bed Probe Sample No. 6
No. 7
Second Cyclone No. 6
Third Cyclone No. 6
Bed Probe Sample No. 8
Second Cyclone No. 12
No. 13
No. 14
Third Cyclone No. 12
No. 14
Bed Probe Sample No. 13
No.
No.
15
17
13.46
7.06
24.47
17.21
19.49
6.91
24.49
23.53
18.40
7.90
22.32
21.85
14.16
14.47
7.83
25.16
25.03
10.36
6.36
24.19
10.00
9.31
12.91
6.43
5.03
24.51
24.87
24.74
6.02
7.21
14.09
7.96
7.11
7.78
13.35
14.95
6.81
5.58
14.16
0.05
4.81
4.62
4.65
14.58
13.60
3.59
4.04
13.73
3.45
2.92
4.79
3.37
3.16
13.89
14.12
13.76
18.94
23.14
44.46
21.84
23.07
23.28
44.5
42.38
20.46
18.15
45.03
-
15.10
14.69
15.45
46.44
44.01
11 .93
11.67
44.60
9.20
9.85
13.55
10.75
10.61
44.71
45.33
44.04
C03
1.54
0.01
0.54
2.62
3.69
0.10
0.70
0.32
4.71
0.91
0.39
64.97
2.19
2.16
1.21
0.15
0.28
1.40
0.25
0.14
1.01
1 .08
2.19
0.35
0.00
0.32
0.17
0.13
Total C
1.46
0.75
0.21
3.76
5.59
1.62
0.14
0.08
2.60
1.46
0.09
-
5.48
4.38
1.95
0.04
0.06
4.09
1.66
0.05
3.97
5.56
4.39
1.65
1.86
0.08
0.06
0.07
6.36
5.14
13.23
8.11
9.21
5.43
12.95
13.14
8.71
4.06
13.02
13.07
6.19
.21
.95
11.99
12.00
4.52
2.95
11.60
3.78
3.89
5.61
2.54
2.41
11.96
11.71
11.96
-------
APPENDIX N (CONT'D). MINIPLANT SOLIDS ANALYSIS
Run No.
100.6
Source
Ca
Weight Percent
102
103
Second Cyclone No. 29
No. 31
Third Cyclone No. 31
Bed Probe Sample No. 32
No. 33
15.61
14.25
6.59
22.99
24.33
5.52
5.33
4.41
13.49
13.32
17.00
15.31
13.16
42.43
42.09
3.26
2.42
0.12
0.06
0.10
3.43
4.09
1.26
0.03
0.02
u
7.24
6.33
3.11
10.86
11.60
COMBUSTOR
Second Cyclone No. 30
Third Cylcone No. 30
Bed Probe Sample No. 4
Final Bed
REGENERATOR
First Cyclone No. 29
Final Bed
COMBUSTOR
Second Cyclone No. 33
No. 40
No. 52
No. 60
No. 75
No. 87
No. 99
Third Cyclone No. 33
No. 40
No. 52
No. 60
No. 75
No. 87
No. 99
15.61
14.25
6.59
22.99
24.33
29.2
18.6
55.3
47.8
40.0
49.9
23.3
15.8
12.9
18.0
19.0
14.1
6.04
16.3
10.5
8.93
13.5
13.9
12.1
7.46
5.52
5.33
4.41
13.49
13.32
2.63
2.87
2.05
3.43
3.74
1.65
4.32
4.29
3.00
3.58
4.00
2.95
1.58
3.15
3.70
3.14
2.98
2.96
2.34
2.14
17.00
15.31
13.16
42.43
42.09
7.01
8.02
6.87
11.67
9.67
4.24
12.5
11.0
9.19
10.89
11.97
8.83
4.61
9.34
11.08
9.07
9.48
8.99
7.49
6.67
C03
3.26
2.42
0.12
0.06
0.10
1.70
3.13
0.83
0.69
3.91
0.18
1.23
0.52
0.29
0.71
0.93
0.65
0.05
1.24
1.25
0.20
0.55
0.49
0.26
0.05
Total C
3.43
4.09
1.26
0.03
0.02
4.63
3.11
0.25
0.22
0.13
2.60
6.43
8.80
4.69
3.28
5.64
9.60
1.49
3.23
3.52
1.19
0.93
1.57
2.45
Me
-------
APPENDIX N (CONT'D). MINIPLANT SOLIDS ANALYSIS
ro
10
Run No. Source Ca
103 Bed Probe Sample No. 28 47.8
(Cont'd.) No. 40 48.3
No. 52 46.1
No. 60 45.1
No. 15 44.1
No. 87 45.2
No. 99 38.8
Final Bed 39.4
REGENERATOR
Sand Filters No. 40 35.4
No. 50 32.8
No. 97 30.1
Reg. to Comb. Pulse Pot 39.1
Final Bed 40.1
104 Second Cyclone 12.6
Third Cyclone 8.36
Bed Overflow 20.8
Bed Probe 21 .7
Final Bed 20.7
105 COMBUSTOR
Second Cyclone No. 28 14.7
No. 44 13.4
Third Cyclone No. 28 15.9
No. 44 13.2
Bed Probe Sample 44.2
43.1
Initial Bed 37is
Final Bed 37.3
8.76
6.10
8.60
10.15
7.67
8.05
11.64
13.60
5.16
4.95
5.61
11.07
9.26
5.92
3.24
11 .50
11.93
10.82
2.70
3.56
2.08
2.21
7.60
7.16
9.42
14.77
24.21
17.88
22.72
29.21
23.48
22.90
33.51
40.5
13.51
13.73
14.29
31.87
27.58
19.43
10.69
33.34
39.64
36.09
19.56
9.84
7.54
7.50
25.31
23.80
16.84
31.66
Weight
COg
0.53
4.33
0.54
1.83
0.91
1.34
1.45
0.88
1.03
0.76
3.78
1.53
0.69
0.04
0.00
0.64
0.39
1.07
0.42
0.24
0.18
0.31
0.33
0.33
9.09
0.70
Percent
Total C
0.14
0.85
0.17
0.44
0.20
0.32
0.33
0.21
1.46
1.34
1.67
0.42
0.21
1.06
0.43
0.19
0.15
0.33
6.10
6.14
1.69
1.95
0.26
0.22
7.99
0.18
7,
3,
12,
14,
96
22
4
0
12.5
-------
APPENDIX N (CONT'D). MINIPLANT SOLIDS ANALYSIS
Run No.
Source
Ca
S04
Weight Percent
o
o
105
(Cont'd.)
107
108
REGENERATOR
Cyclone No. 25
No. 41
Bed Sample
Initial Bed
Final Bed
Second Cyclone No. 27
No. 50
Third Cyclone No. 27
No. 50
Bed Overflow No. 24
No. 44
Second Cyclone - Final
Third Cyclone - Final
Bed Overflow - Final
32.8
35.0
37.7
35.7
44.7
8.35
9.16
5.70
6.10
20.4
22.2
11.5
7.62
21.9
2.46
1.78
3.84
7.01
3.11
4.07
3.42
2.97
3.21
11.07
12.84
.
-
-
^^•••^^^B
7.26
7.44
10.28
24.59
10.47
12.14
10.42
9.57
10.06
34.42
39.72
16.93
10.68
41.90
3.66
3.57
1.58
8.19
0.59
0.45
0.22
0.20
0
7.27
4.81
0.36
0.03
3.39
2.45
1.91
3.80
1.27
0.27
4.62
5.78
1.66
1.33
1.54
1.06
3.33
1.24
0.93
•IMi^MMrt
-
-
-
-
-
4.41
4.88
2.41
2.48
12.5
13.6
6.49
3.32
13.8
-------
APPENDIX 0. MINI PLANT SAMPLE SHIPMENTS
CO
o
Requestor
Acurex-Aerotherm
Mountain View, CA
Atlantic Richfield Corp.
Alexandria, VA
Battelle
Columbus, OH
Colorado State University
Fort Collins, CO
Cornell University
Ithaca, NY
EFB, Inc.
Woburn, MA
ERE
Baton Rouge, LA
Florham Park, NJ
Linden, NJ
EPA
Research Triangle Park, NC
Sample Description
3rd Cyclone Flyash
2nd Cyclone Flyash
2nd Cyclone Flyash
3rd Cyclone Flyash
3rd Cyclone Flyash
Scrubber Water
2nd Cyclone Flyash
3rd Cyclone Flyash
2nd Cyclone Flyash
Bed Overflow
2nd Cyclone Flyash
Bed Overflow
3rd Cyclone Flyash
Bed Overflow
3rd Cyclone Flyash
Fresh Dolomite
3rd Cyclone
3rd Cyclone
2nd Cyclone
2nd Cyclone
2nd Cyclone
3rd Cyclone
Flyash
Flyash
Flyash
Flyash
Flyash
Flyash
2nd Cyclone Flyash
3rd Cyclone Flyash
2nd Cyclone Flyash
3rd Cyclone Flyash
Sorbent
Dolomite
Dolomite
Dolomite
Dolomi te
Dolomite
Dolomite
Limestone
Limestone
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Run No.
67
67
Combination
of Many Runs
67
80
81
81
81
74
74
78
78
78
78
78
--
67
78
80
93
67
67
50. IB
80
81
81
Amount
15 Gallons
2 Drums
18 Drums
2 Grams
20 Grams
1 Drum
1 Quart
1 Quart
10 Gallons
10 Gallons
10 Gallons
10 Gallons
5 Gallons
4 Drums
1/2 Drum
1-1/2 Tons
10-1/2 Ibs.
1 Drum
1 Gallon
1 Drum
1 Ib.
1 Ib.
1 Drum
20 Grams
5 Gallons
5 Gallons
Date
1/27/78
4/4/78
4/6/78
4/4/78
12/4/78
11/7/78
11/7/78
11/7/78
8/15/78
8/15/78
8/15/78
8/15/78
9/28/78
11/14/78
11/28/78
12/21/78
6/16/78
7/19/78
1/8/79
3/9/79
9/28/78
9/28/78
8/30/77
12/4/78
3/28/79
3/28/79
-------
APPENDIX 0 (CONT'D). MINIPLANT SAMPLE SHIPMENTS
Requestor
Sample Descrlption
Sorbent
Run No.
Amount
Date
CO
EPA (Cont'd)
Rivesville, WV
GCA Corporation
Bedford, MA
General Electric
King of Prussia, PA
Schenectady, NY
Valley Forge, PA
MIT
Cambridge, MA
N.J. Institute of Technology
Newark, NJ
N.Y. University
Westbury, NY
Pratt & Whitney Aircraft Co.
Middletown, CT
Radian Corporation
Austin, TX
2nd Cyclone Flyash
3rd Cyclone Flyash
Bed Overflow
Bed Overflow
Bed Overflow
2nd Cyclone Flyash
Final Bed (Combustor)
Final Bed (Regenerator)
3rd Cyclone Flyash
3rd Cyclone Flyash
3rd Cyclone Flyash
Illinois Coal No. 6
Pfizer Dolomite
2nd Cyclone Flyash
3rd Cyclone Flyash
Bed Overflow
Bed Overflow
2nd Cyclone Flyash
Illinois Coal No. 6
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Limestone
Limestone
Dolomite
Dol omi te
Dolomite
—
—
Dol omi te
Dolomite
Dolomite
Dolomite
Dolomite
--
108
108
108
66
67
67
105
105
68
68
81
—
—
50.4
80
78
78
78
-.
1 Quart
1 Quart
1 Quart
1 Drum
12 Drums
5 Drums
10 Gallons
10 Gallons
5 Gallons
5 Gallons
1 Small Jar
1 Quart
1 Quart
1 Drum
5 Gallons
5 Gallons
25 Gallons
25 Gallons
25 Ibs.
8/13/79
8/13/79
8/13/79
2/2/78
2/2/78
2/2/78
5/14/79
5/14/79
4/17/79
1/18/79
11/13/78
3/14/78
3/14/78
9/30/77
9/22/78
8/18/78
10/17/78
10/17/78
10/20/78
Champion Coal
2nd Cyclone Flyash Dolomite
3rd Cyclone Flyash Dolomite
Bed Overflow Dolomite
2nd Cyclone Flyash Dolomite
Bed Overflow Dolomite
2nd Cyclone Flyash Dolomite
3rd Cyclone Flyash Dolomite
1 Drum
68
67
79
79
80
80
80
10
10
2
3
21
15
1
Ibs.
Ibs.
Drums
Drums
Drums
Drums
Drum
3/20/79
5/2/78
5/2/78
1/24/79
1/24/79
1/24/79
1/24/79
1/24/79
-------
APPENDIX 0 (CONT'D). MINIPLANT SAMPLE SHIPMENTS
Requestor
Radian Corporation (Cont'd)
Sandia Laboratory
Llvermore, CA
TVA
Muscle Shoals, AL
University of Cincinnati
Cincinnati, OH
University of Denver
Denver, CO
University of S. California
Los Angeles, CA
Sample Description Sorbent
3rd Cyclone Flyash Dolomite
Bed Overflow Dolomite
2nd Cyclone Flyash Dolomite
Bed Overflow Dolomite
2nd Cyclone Flyash Dolomite
3rd Cyclone Flyash Dolomite
3rd Cyclone Flyash Dolomite
2nd Cyclone Flyash Dolomite
3rd Cyclone Flyash Dolomite
Bed Overflow Dolomite
3rd Cyclone Flyash Dolomite
3rd Cyclone Flyash Dolomite
2nd Cyclone Flyash Dolomite
3rd Cyclone Flyash Dolomite
2nd Cyclone Flyash Dolomite
Granular Bed Dolomite
Filter Particulates
Illinois Coal No. 6
Pfizer Dolomite
2nd Cyclone Flyash Dolomite
3rd Cyclone Flyash Dolomite
Bed Overflow Dolomite
Champion Coal
Ohio Coal
Pfizer Dolomite
Grove Limestone
2nd Cyclone Flyash Dolomite
Bed Overflow Dolomite
2nd Cyclone Flyash Limestone
3rd Cyclone Flyash Limestone
Run No.
81
79
79
80
80
80
81
78
78
78
80
81
78
106
59
59
•• •-
--
78
78
78
—
--
_.
—
50
50
75
75
Amount
1/2 Drum
1 Small Bottle
1 Small Bottle
1 Small Bottle
1 Small Bottle
1 Small Bottle
1 Small Bottle
10 Ibs.
10 Ibs.
6 Drums
5 Gallons
5 Gallons
5 Gallons
10 Gallons
1 Ib.
1 Ib.
1 Ib.
1 Ib.
1 Ib.
1 Ib.
1 Ib.
1 Ib.
1 Ib.
1 Ib.
1 Ib.
1 Ib.
1 Ib.
1 Ib.
1 Ib.
Date
1/24/79
1/19/79
1/19/79
1/19/79
1/19/79
1/19/79
1/19/79
2/6/79
2/6/79
5/8/79
2/6/79
2/27/79
2/27/79
8/9/79
9/7/77
9/7/77
8/15/78
8/1 5/78
8/15/78
8/15/78
8/15/78
10/2/78
10/2/78
10/2/78
10/2/78
10/2/78
10/2/78
10/2/78
10/2/78
-------
APPENDIX 0 (CONT'D). MINIPLANT SAMPLE SHIPMENTS
CO
o
Requestor
University of S. California
(Cont'd)
Valley Forge Laboratories
Devon, PA
Vanderbilt University
Nashville, TN
Westinghouse Research Labs
Pittsburgh, PA
Sample Description
Bed Overflow
2nd Cyclone Flyash
3rd Cyclone Flyash
Bed Overflow
Bed Overflow
2nd Cyclone Flyash
Bed Overflow
Bed Overflow
Bed Overflow
3rd Cyclone Flyash
3rd Cyclone Flyash
3rd Cyclone Flyash
2nd Cyclone Flyash
2nd Cyclone Flyash
3rd Cyclone Flyash
2nd Cyclone Flyash
Bed Overflow
2nd Cyclone Flyash
Bed Overflow
2nd Cyclone Flyash
3rd Cyclone Flyash
Bed Overflow
2nd Cyclone Flyash
3rd Cyclone Flyash
Bed Overflow
2nd Cyclone Flyash
3rd Cyclone Flyash
Bed Overflow
2nd Cyclone Flyash
3rd Cyclone Flyash
Bed Overflow
3rd Cyclone Flyash
Sorbent
Limestone
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Limestone
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Limestone
Limestone
Limestone
Dolomite
Run No.
75
77
77
77
67
67
26
45
52
67
79
59
79
68
68
69
69
70
70
71
71
71
72
72
72
73
73
73
74
74
74
81
Amount
1 Ib,
1 Ib.
1 Ib.
1 Ib.
25 Ibs.
25 Ibs.
5 Ibs.
5 Ibs.
5 Ibs.
500 Grams
1 Drum
10 Ibs.
10 Ibs.
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Gallon
Date
10/2/78
10/2/78
10/2/78
10/2/78
2/2/78
2/2/78
4/26/78
4/26/78
4/26/78
2/16/78
10/27/78
10/27/78
10/27/78
1/8/79
1/8/79
1/3/79
1/3/79
1/3/79
1/3/79
1/3/79
1/3/79
1/3/79
1/3/79
1/3/79
1/3/79
1/3/79
T/3/79
1/3/79
1/3/79
1/3/79
1/3/79
2/26/79
-------
APPENDIX 0 (CONT'D). MINIPLANT SAMPLE SHIPMENTS
Requestor
Westinghouse Research Labs
(Cont'd)
Sample Description Sorbent
1st Cyclone (Dipleg) Dolomite
Material
3rd Cyclone Flyash Dolomite
2nd Cyclone Flyash Limestone
3rd Cyclone Flyash Limestone
Bed Probe Limestone
Regenerator Cyclone Ash Limestone
Regenerator Final Bed Limestone
Run No.
79
81
105
105
105
105
105
Amount
2-1/2 Gallons
2 Gallons
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
1 Small Jar
Date
4/28/79
5/11/79
8/17/79
8/17/79
8/17/79
8/17/79
8/17/79
CO
o
en
-------
APPENDIX P-l. BENCH COMBUSTOR RUN SUMMARY
OJ
o
en
Program
Run No. 1.1
Operating Conditions
Pressure (kPaa)
Bed Temperature (°C)
Air Flow Rate (m3/m1n)
Excess A1r (%)
Coal Feed Rate (kg/hr)
Expanded Bed Height (m)
Superficial Velocity (m/s)
Gas Residence Time (s)
Ca/S Molar Ratio
Run Length at Steady State (hr) 3.7
Flue Gas Emissions
S02 (ppm)
NT (ppm)
COX (ppm)
C02 (X)
/\ » I at \
Initial Checkout
(1)
Results
S02 Retention
Lbs SOp/MBTU
Lbs NOVMBTU
rt
1.2
(1)
2.1
(1)
3.1
(1)
3.2
(1)
4.1
(1,2)
4.2
(1,2)
770
880
2.9
146
10.3
0.5
1.9
0.28
2.43
•) 3.7
550
200
175
12.0
13
62
2.0
0.50
770
890
2.9
153
10.5
0.5
1.9
0.27
2.71
0.6
450
180
125
12.0
13
69
1.6
0.46
760
910
3.0
111
10.1
0.5
1.9
0.26
3.54
1.9
570
270
500
11.4
11
60
2.2
0.74
760
880
1.6
37
8.0
0.9
1.1
0.83
2.78
2.2
790
190
325
13.6
5.8
57
2.0
0.35
760
780
1.7
41
7.5
0.9
1.1
0.88
2.87
1.5
900
200
350
11.4
6.2
47
2.6
0.42
770
865
2.2
16
12.1
1.1
1.4
0.77
2.87
1.4
1000
120
200
16.8
2.9
54
2.4
0.20
770
870
2.2
13
12.9
1.2
1.4
0.84
2.68
0.5
800
140
225
17.8
2.5
65
1.8
0.22
(1) Grove Limestone; ArkwHght Coal
(2) 1st Cyclone Sol Ids Recycled
-------
APPENDIX P-l (CONT'D). BENCH COMBUSTOR RUN SUMMARY
Program Initial Checkout
Run NO. s.i'1.2' 6.1"' 6.2'1' r.i'1' e.i'1'2' 9..
Operating Conditions
Pressure (kPaa) 760 760 750 750 750 770
Bed Temperature (°C) 880 915 900 950 930 960
Air Flow Rate (m3/min) 1.8 1.9 2.0 1.3 1.6 1.4
Excess Air (%) 16 41 13 44 37 11
Coal Feed Rate (kg/hr) 10.5 10.2 11.3 4.7 7.1 11.0
Expanded Bed Height (m) 0.9 0.8 1.1 0.5 0.6 0.9
Superficial Velocity (m/s) 1.2 1.3 1.7 1.0 1.0 1.3
Gas Residence Time (s) 0.76 0.61 0.84 0.54 0.61 0.66
Ca/S Molar Ratio 0.96 1.09 1.58 3.77 1.30 1.97
Run Length at Steady State (hr) 2.2 0.9 0.4 1.3 1.2 0.3
Flue Gas Emissions
SO, (ppm) 630 430 100 100 160 218
NCT (ppm) 95 150 120 180 160 160
CO (ppm) 175 125 150 75 250 125
CO, (%} 15.4 14.2 15.0 12.5 14.8 16.4
Og (%) 3.0 6.2 2.5 6.5 5.8 2.1
Results
S02 Retention (%) 72 81 95 92 91 89
Lbs SOo/MBTU 1.4 1.0 0.23 0.36 0.47 0.35
Lbs NO /MBTU 0.15 0.26 0.20 0.46 0.34 0.18
/\
(1) Pfizer Dolomite; Arkwright Coal
(2) 1st Cyclone Solids Recycled
-------
APPENDIX P-2. BENCH COMBUSTOR RUN SUMMARY
Program Two Stage Combust1on-NOX Control
Run No. 10.r1^ 11.1^ 11.2^ 12.1^ 12.2^ 13.
Operating Conditions
Pressure (kPaa) 770 765 765 770 770 770
Bed Temperature (°C) 920 870 870 960 950 950
Air Flow Rate (m3/min) 0.9 1.9 1.8 1.5 1,4 1.7
Excess A1r (%} 39 23 18 25 8 17
Coal Feed Rate (kg/hr) 8.3 18.2 18.2 10.9 10,9 13.5
Expanded Bed Height (m) 0.9 1.0 1.2 1.1 1,3 1.1
Superficial Velocity (m/s) 0.6 1.2 1.1 1.3 1.3 1.5
Gas Residence Time (s) 1.4 0.8 1.1 0.79 1.03 0.70
Ca/S Molar Ratio 6.2 2.9 2.9 3.5 3.5 3.1
Run Length at Steady State (hr) 0.8 0.3 0.2 0.7 0.4 0.6
Primary A1r (% Stoic.) 70 66 61 105 75 97
Secondary/Primary A1r 0.37 0.32 0.34 0 0.34 0
o Secondary A1r Injection (cm Above Grid) 30.7 30.7 30.7 — 30.7
Flue Gas Emissions
SO, (ppm) 900 1455 1170 588 960 400
NO (ppm) 120 60 140 250 120 240
CO (ppm) 200 >5000 650 125 150 140
CO-, (%) 13.0 15.4 18.6 15.4 19.0 17.6
6.0 4.0 3.2 4.3 1.6 3.1
Results
SOo Retention (%) 42 17 37 65 53 79
Lbs SOo/MBTU 1.27 1.86 1.47 0.98 1.54 0.62
Lbs NO^/MBTU 0.12 0.06 0.12 0.30 0.14 0.27
rt
(1) Pfizer Dolomite; Champion Coal
(2) Grove Limestone; Champion Coal
-------
APPENDIX P-2 (CONT'D). BENCH COMBUSTOR RUN SUMMARY
Program Two Stage Combustion-N0x Control
Run No. 13.2'^ 14.1^' 14.2^ 17.1^ 17.2^^ 17A.l'2^
Operating Conditions
Pressure (kPaa) 770 760 760 770 770 480
Bed Temperature (°C) 950 880 930 850 830 870
Air Flow Rate (m3/min) 1.7 1.8 1.5 1.3 1.3 1.5
Excess Air (%) 16 19 20 21 31 41
Coal Feed Rate (kg/hr) 11.4 12.5 11.0 10.5 9.1 10.0
Expanded Bed Height (m) 1.0 1.6 1.2 1.2 1.2 1.0
Superficial Velocity (m/s) 1.5 1.5 1.4 1.0 1.1 2.0
Gas Residence Time (s) 0.70 1.1 0.87 1.2 1.1 0.5
Ca/S Molar Ratio 3.1 2.8 2.9 2.7 3.2 2.9
Run Length at Steady State (hr) 1.0 1.0 2.0 0.5 0.5 0.6
w Primary Air (% Stolch.) 91 112 79 92 86 116
S Secondary/Primary A1r 0.24 0 0,38 0 0.33 0
Secondary Air Injector (cm Above Grid) 5.9 — 5.9 — 5.9
Flue Gas Emissions
S02 (ppm) 600 840 453 769 829 795
NO^ (ppm) 145 180 150 140 160 190
COX (ppm) 125 150 150 275 250 400
C0? (%} 16.4 15.2 15.4 14.6 14.0 14.4
02 (%) 2.9 3.4 3.5 3.8 5.0 6.2
Results
S02 Retention (%) 68 60 75 57 50 48
Lbs S02/MBTU 1.01 1.48 0.79 1.15 1.52 1.48
Lbs NOX/MBTU 0.19 0.23 0.19 0.15 0.21 0.26
(1) Grove Limestone; Champion Coal
(2) Grove Limestone; Arkwright Coal
-------
APPENDIX P-2 (CONT'D). BENCH COMBUSTOR RUN SUMMARY
Program Two Stage Combustion-N0x Control
Run No. 17A.2 17A.3^ 20.1^ 20. 2^ 20.3^ 21
(1) Grove Limestone; Arkwrlght Coal
(2) Grove Limestone; Champion Coal
Operating Conditions
Pressure (kPaa) 480 480 490 490 490 480
Bed Temperature jf°C) 810 820 900 890 900 840
Air Flow Rate (m3/m1n) 1.3 1.4 1.6 1.5 1.5 1.2
Excess Air (%) 18 18 15 18 19 18
Coal Feed Rate (kg/hr) 9.4 9.5 10.8 11.2 11.3 10.0
Expanded Bed Height (m) 1.4 1.2 1.1 1.2 1.2 1.2
Superficial Velocity (m/s) 1.7 1.8 2.2 2.0 2.1 1.6
Gas Residence Time (s) 0.76 0.67 0.52 0.59 0.57 0.77
Ca/S Molar Ratio 3.1 3.1 3.0 2.9 2.8 2.9
Run Length at Steady State (hr) 1.9 2.0 1.3 1.5 2.5 1.3
Primary Air (% Stolen.) 89 72 115 85 74 93
Secondary/Primary A1r 0.24 0.57 0 0.22 0.42 0
Secondary A1r Injector (cm Above Grid) 15 15 — 30 30
Flue Gas Emissions
S0? (ppm)
N(T (ppm)
COX (ppm)
COo
°2
Results
SO? Retention (%) 33 25 43 42 42 19
Lbs SOo/MBTU 2.16 2.53 2.02 1.79 1.78 2.25
Lbs NO /MBTU 0.23 0.25 0.32 0.23 0.20 0.20
1230
185
500
15.0
3.3
1390
190
550
15.0
3.2
1080
240
300
16.4
2.8
1080
190
300
15.8
3.2
1060
190
300
15.8
3.4
1490
170
540
14.4
3.3
-------
APPENDIX P-2 (CONT'D). BENCH COMBUSTOR RUN SUMMARY
Program Two Stage Combustion-N0x Control
Run No. 21.2^) 21.3^ 22.1^ 22.2^ 22.3^
Operating Conditions
Pressure (kPaa) 480 505 515 505 505
Bed Temperature (°C) 860 840 950 880 890
A1r Flow Rate (m3/m1n) 1.3 1.2 2.0 1.7 1.6
Excess A1r (%) 16 15 35 37 31
Coal Feed Rate (kg/hr) 10.6 9.6 14.2 11.5 11.4
Expanded Bed Height (m) 1.3 1.4 1.3 1.1 1.2
Superficial Velocity (m/s) 1.7 1.5 2.7 2.2 2.1
Gas Residence Time (s) 0.74 0.91 0.48 0.52 0.56
Ca/S Molar Ratio 2.7 2.93 3.20 3.13 3.15
Run Length at Steady State (hr) 1.3 1.4 1.5 2.0 1.7
Primary A1r (% Stolch.) 80 73 111 91.2 80
Secondary/Primary A1r 0.23 0.33 0 0.24 0.39
Secondary A1r Injector (cm Above Grid) 30 30
Flue Gas Emissions
S02 (ppm) 1410 1020 500 900 880
N(T (ppm) 190 140 245 230 210
CO (ppm) 425 500 225 350 350
CO, (%) 15.0 15.6 14.2 15.0 15.0
02 (%) 3.0 2.8 5.5 5.7 5.0
Results
SOo Retention (%) 25 46 67 43 47
Lbs SO-/MBTU 2.23 1.57 0.89 1.63 1.56
Lbs NO^/MBTU 0.22 0,15 0.31 0.30 0.27
3\
(1) Grove Limestone; Champion Coal
(2) Pfizer Dolomite; Champion Coal
-------
APPENDIX P-3. BENCH COMBUSTOR RUN SUMMARY
co
H->
ro
Program
Run No.
Operating Conditions
Pressure (kPaa)
Bed Temperature (°C)
Air Flow Rate (m^/min)
Excess Air (%)
Coal Feed Rate (kg/hr)
Expanded Bed Height (m)
Superficial Velocity (m/s)
Gas Residence Time (s)
Ca/S Molar Ratio
Run Length at Steady State (hr)
NH3/NOX Molar Ratio
H2/NH3 Molar Ratio
NH3 Injection Location
Flue Gas Emissions
SOp (ppm)
NO* (ppm)
COX (ppm)
C02 (X)
022 (X)
Results
SO? Retention (%)
Lbs SOo/MBTU
Lbs NO /MBTU
NH3 Injection-N0x Control
1.1
61
1.25
0.18
1.2
1.4
3.1
3.2
4.1
63
1.24
0.16
65
1.17
0.23
68
1.08
0.17
75
0.81
0.13
70
0.98
0.13
4.2
760
885
1.5
19
8.4
1.3
1.3
0.98
3.05
1.5
--
—
--
760
890
1.5
16
8.4
1.3
1.3
0.95
3.07
2.2
0.88
0.0
LI
760
885
1.5
16
7.9
1.2
1.3
0.93
3.22
0.9
0.88
2.00
LI
760
890
1.5
15
9.4
1.5
1.3
1.14
2.94
2.8
--
--
--
760
890
1.5
13
10.7
1.4
1.3
1.09
2.57
1.3
2.27
1.48
LI
760
895
1.4
16
11.5
1.4
1.2
1.12
2.08
1.8
--
—
—
760
900
1.4
14
11.9
1.4
1.2
1.14
2.04
1.4
2.00
1.00
LI
700
140
200
16.0
3.8
675
117
550
16.0
3.1
600
160
450
16.0
3.0
675
145
200
16.0
2.9
588
133
500
16.0
2.7
825
155
200
15.2
3.3
800
143
200
15.4
2.9
72
0.90
0.12
Grove Limestone; Champion Coal
-------
APPENDIX P-3 (CONT'D). BENCH COMBUSTOR RUN SUMMARY
CO
t—•
CO
Program
Run No.
Operati ng Cond i t1 ons
Pressure (kPaa)
Bed Temperature (°C)
Air Flow Rate (m^/min)
Excess Air (%)
Coal Feed Rate (kg/hr)
Expanded Bed Height (m)
Superficial Velocity (m/s)
Gas Residence Time (s)
Ca/S Molar Ratio
Run Length at Steady State (hr)
NH3/NOX Molar Ratio
Ho/NH3 Molar Ratio
N&3 Injection Location
Flue Gas Emissions
SOo (ppm)
NO^ (ppm)
CO* (ppm)
C0
Results
S02 Retention (%)
Lbs S02/MBTU
Lbs N
NH, Injection-NO,, Control
«3 X
4.3
5.1
5.2
5.3
5.4
5.5
5.6
760
900
1.4
17
11.9
1.4
1.2
1.13
1 .99
1.0
__
_—
__
760
895
1.5
19
7.9
1.3
1.3
1.01
3.07
1.5
_-
-_
__
760
905
1.5
19
10.0
1.1
1.3
0.89
2.46
0.8
2.18
2.94
L2
760
915
1.5
18
10.0
1.4
1.3
1.06
2.46
1.7
1.27
2.50
L2
760
905
1.5
18
9.0
1.4
1.3
1.03
2.67
1.3
--
_-
--
760
900
1.5
18
8.8
1.3
1.3
0.99
2.78
1
3.41
0.0
12
760
890
1.5
17
8.8
1.4
1 .3
1.06
2.70
1.1
8.05
0.0
12
725
170
450
15.4
3.4
725
145
350
14.8
3.3
625
210
475
14.8
3.4
625
200
375
14.8
3.3
600
160
325
15.2
3.3
575
210
300
15.2
3.2
550
250
375
15.2
3.1
75
0.79
0.13
58
1.49
0.22
71
0.99
0.24
71
0.90
0.23
69
1.04
0.20
70
1.03
0.27
71
0.99
0.32
Grove Limestone; Champion Coal
-------
APPENDIX P-3 (CONT'D), BENCH COMBUSTOR RUN SUMMARY
Program NH3 Injection-N0x Control
Run No- 5.7 6.1 6.2 6.3 7.1 7.2 7.3
Operating Conditions
Pressure (kPaa) 760 660 660 660 660 660 660
Bed Temperature (PC) 890 890 880 890 890 890 890
Air Flow Rate (m3/min) 1.5 1.4 1.4 1.4 1.4 1.4 1.4
Excess Air (%) 17 16 18 17 19 18 17
Coal Feed Rate (kg/hr) 8.8 9.2 8.4 8.5 9.2 9.2 9.2
Expanded Bed Height (m) 1.3 1.3 1.3 1.4 1.3 1.3 1.3
Superficial Velocity (m/s) 1.3 1.4 1.4 1.4 1.5 1.4 1.4
Gas Residence Time (s) 1.0 0.93 0.92 0.96 0.90 0.92 0.90
Ca/S Molar Ratio 2.70 2.77 3.03 2.97 2.80 2.80 2.80
Run Length at Steady State (hr) 0.9 1.2 1.0 1.0 1.2 1.1 1.3
NH3/NOX Molar Ratio — — 0.82 -- -- 1.36 2.25
H2/NH3 Molar Ratio — -- 0.0 -- — 0.0 0.0
NH3 Injection Location — — L3 -- — L3 L3
Flue Gas Emissions
S02 (ppm)
NOX (ppm)
COX (ppm)
C02 (*)
02 (X)
Results
S0£ Retention (%) 75 70 71 71 68 75 82
Lbs S02/MBTU 0.85 0.99 0.89 0.88 1.11 0.84 0.58
Lbs NO^/MBTU 0.22 0.16 0.12 0.19 0.18 0.12 0.09
475
170
400
15.4
3.1
650
150
475
16.0
3.0
563
103
650
16.8
3.5
575
170
675
16.8
3.5
675
155
525
14.8
3.3
525
100
625
15.2
3.3
375
83
825
16.0
3.3
Grove Limestone; Champion Coal
-------
APPENDIX P-3 (CONT'D). BENCH COMBUSTOR RUN SUMMARY
Program NH3 Injection-NO Control
Run No. 7.4 8.1 8.2 8.3 8.4 8.5
Operating Conditions
Pressure (kPaa) 660 660 660 660 660 660
Bed Temperature (°C) 890 890 890 890 885 880
Air Flow Rate (m^/min) 1.4 1.3 1.3 1.3 1.3 1.3
Excess Air (%) 16 19 18 19 17 15
Coal Feed Rate (kg/hr) 9.2 7.9 7.9 7.9 10.1 10.1
Expanded Bed Height (m) 1.2 1.3 1.3 1.3 1.3 1.3
Superficial Velocity (m/s) 1.4 1.3 1.3 1.3 1.3 1.3
Gas Residence Time (s) 0.84 0.95 0.98 0.97 0.96 0.93
Ca/S Molar Ratio 2.8 3.12 3.12 3.12 2.44 2.48
Run Length at Steady State (hr) 0.8 0.9 1.1 0.7 0.8 1.1
NH3/NOX Molar Ratio — — 1.24 2.29 1.39
H2/NH3 Molar Ratio — — 0.0 0.0 1.46
NH3 Injection Location — -- LI LI LI
Flue Gas Emissions
SO- (ppm) 550 625 600 563 513 450
NO, (ppm) 150 133 115 100 115 132
CO (ppm) 425 225 363 550 550 500
CO, (%) 15.8 15.8 15.4 14.8 15.2 15.8
3.0 3.6 3.3 3.4 3.2 3.0
Results
S0? Retention (%) 74 68 70 71 79 82
Lbs S02/MBTU 0.87 1.04 1.02 1.00 0.68 0.58
Lbs NOx/MBTU 0.17 0.16 0.14 0.13 0.11 0.12
Grove Limestone; Champion Coal
-------
APPENDIX P-4. BENCH COMBUSTOR RUN SUMMARY
Program Simulated Flue Gas Reclrculatlon-NO Control
Run No. 1A.1 1A.2 1A.3 2.1 2.2
Operating Conditions
Pressure (kPaa) 810 810 810 810 810
Bed Temperature (°C) 875 890 880 875 890
Air Flow Rate (m3/min) 1.4 1.4 1.4 1.4 1.5
Excess A1r (%) 19 13 17 18 7
Coal Feed Rate (kg/hr) 9.0 9.0 9.0 7.9 10.4
Expanded Bed Height (m) 1.4 1.3 1.3 1.3 1.4
Superficial Velocity (m/s) 1.2 1.3 1.2 1.2 1.5
Gas Residence Time (s) 1.20 1.02 1.15 1.13 0.92
Ca/S Molar Ratio 2.97 3.10 3.10 3.36 2.85
Run Length at Steady State (hr) 1.5 2 0.7 0.5 1.0
N2 Flow Rate (m3/m1n) -- 0.16 -- — 0.35
Redrculatlon Ratio (%) -- 11.2 -- -- 23.4
Flue Gas Emissions
SO, (ppm) 550 600 450 419 525
NO^ (ppm) 160 117 170 155 70
COX (ppm) 200 400 475 200 300
CO, (%) 15.4 15.4 16.0 14.8 14.5
02Z (%) 3.5 2.1 3.3 3.2 1.1
Results
S02 Retention (%) 73 67 78 77 72
Lbs S09/MBTU 0.89 1.17 0.72 0.81 1.09
Lbs NO^/MBTU 0.19 0.16 0.19 0.22 0.07
Grove Limestone; Champion Coal
-------
APPENDIX P-4 (CONT'D). BENCH COMBUSTOR RUN SUMMARY
Program Simulated Flue Gas Reclrculation-NO Control
/V
Run No. 3.1 3.2 4.1 4.2
Operating Conditions
Pressure (kPaa) 810 810 810 810
Bed Temperature (°C) 910 900 820 820
Air Flow Rate (m3/min) 1.8 1.8 1.3 1.3
Excess Air (%) 27 25 17 13
Coal Feed Rate (kg/hr) 9.6 10.9 8.4 8.7
Expanded Bed Height (m) 1.3 1.2 1.3 1 .4
Superficial Velocity (m/s) >.5 1.7 1.0 1.1
Gas Residence Time (s) 0.86 0.73 1.39 1.29
Ca/S Molar Ratio 3.05 2.87 2.65 2.87
Run Length at Steady State (hr) 1.5 2.0 1.0 2.0
N2 Flow Rate (m3/min) — 0.20 — 0.14
Redrculation Ratio (%) — 11.2 -- 5.0
Flue Gas Emissions
S0? (ppm) 575 575 750 938
NO* (ppm) 160 130 260 150
COX (ppm) 175 550 425 700
CO- («) 16.0 15.4 14.8 14.2
02 (%} 5.0 4.0 3.0 2.0
Results
S02 Retention (%) 67 67 66 54
Lbs S02/MBTU 1.01 1.07 1.18 1.72
Lbs NOV/MBTU 0.20 0.17 0.29 0.13
f\
Grove Limestone; Champion Coal
-------
APPENDIX P-5. BENCH COMBUSTOR RUN SUMMARY
Program Combined NOX Control Methods
Run No.^ 1.1 1.2 1.3 1.4 1.5
Operating Conditions
Pressure (kPaa) 500 500 500 500 500
Bed Temperature (°C) 890 900 890 900 900
A1r Flow Rate (m3/m1n) 1.5 1.5 1.5 1.5 1.5
Excess A1r (%) 25 23 25 23 22
Coal Feed Rate (kg/hr) 9.0 8.7 9.0 10.4 9.0
Expanded Bed Height (m) 1.2 1.3 1.2 1.2 1.3
Superficial Velocity (m/s) 1.9 1.9 2.0 2.0 2.0
Gas Residence Time (s) 0.63 0.65 0.63 0.60 0.64
Ca/S Molar Ratio 3.08 3.17 3.08 2.64 3.08
Run Length at Steady State (hr) 0.8 0.6 0.4 0.6 0.7
2 Primary A1r (% Stolch.) 105 109 106 91 72
00 Secondary/Primary Air (2) 0 0 0 0 0.53
NH3/NOX Molar Ratio (3) — 0.89 -- 1.43
Flue Gas Emissions
S02 (ppm)
MT (ppm) 180 120 125 95 82
CO (ppm) 325 550 650 740 525
CO, (%} 14.8 14.8 14.8 14.8 15.2
Og (%) 4.5 4.0 4.5 4.3 4.0
Results
S0? Retention (%)
Lbs SO-/MBTU
Lbs MT/MBTU 0.21 0.15 0.15 0.09 0.10
(1) Grove Limestone; Champion Coal
(2) Supplementary air Injected 33 cm above grid.
(3) NH3 Injected 290 cm above grid.
-------
APPENDIX P-5 (CONT'D). BENCH COMBUSTOR RUN SUMMARY
CO
>->
10
Program
Run No.
Combined N0 Control Methods
(1)
Operating Conditions
Pressure (kPaa)
Bed Temperature (°C)
Air Flow Rate (m3/min)
Excess Air (%)
Coal Feed Rate (kg/hr)
Expanded Bed Height (m)
Superficial Velocity (m/s)
Gas Residence Time (s)
Ca/S Molar Ratio
Run Length at Steady State (hr)
Primary Air (% Stoich.)
Secondary/Primary Air (2)
NH./NOV Molar Ratio (3)
»5 X
Flue Gas Emissions
S02 (ppm)
NO (ppm)
CO (ppm)
C02 (%)
f\ *•* f a/ \
Results
S0? Retention (%)
Lbs SOp/MBTU
Lbs NCT/MBTU
2.1
500
880
1.5
31
9.7
1.1
2.0
0.56
3.02
0.4
101
0
575
153
400
14.2
5.5
73
0.87
0.17
2.2
500
880
1.6
25
9.7
1.2
2.1
0.58
3.02
0.4
72
0.50
650
120
375
15.2
4.5
68
1.03
0.14
2.3
675
105
425
14.8
5,0
65
1.15
0.13
2.4
510
890
1.6
25
9.3
1.3
2.1
0.61
3.15
0.5
75
0.50
1.61
510
880
1.6
24
9.0
1.2
2.1
0.60
3.28
0.3
78
0.50
—
675
113
440
14.8
4.3
63
1 .22
0.15
(1) Grove Limestone; Champion Coal
(2) Supplementary air Injected 33 cm above grid.
(3) NH3 injected 290 cm above grid.
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO.
EPA-600/7-80-013
RECIPIENT'S ACCESSION NO.
4. TITLE ANDSUBTITLE
Miniplant and Bench Studies of Pressurized Fluidized-
bed Coal Combustion: Final Report
REPORT DATE
January 1980
PERFORMING ORGANIZATION CODE
EXXON/GRU.18GFGS.79
7.
t c. Hoke ,E.S. Matulevicius ,M. Ernst,J. L.
Goodwin,A.R.'Garabrant,I.B.Radovsky,A. S. Lescar-
ret.R.R.Bertrand.L.A.Ruth.V.J.Siminski (see blk 15)
. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING OFIOANIZATION NAME AND ADDRESS
Exxon Research and Engineering Co.
P.O. Box 8
Linden, New Jersey 07036
0. PROGRAM ELEMENT NO.
INE825
11. CONTRACT/GRANT NO.
68-02-1312
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; 8/77 - 8/79
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES IERL-RTP project officer is D. Bruce Henschel, Mail Drop 61,
919/541-2825. (From blk 7: M.S.Nutkis,M.D. Loughnane,H.R.Silakowski, M.W.Gre-
gory, and A.Ichel.) EPA-600/7-78-069, -77-107, and -76-011 are related reports
16. ABSTRACT
The report gives further results of studies on the environmental aspects
of the pressurized fluidized-bed coal combustion process, using the 218 kg coal/hr
continuous combustion/sorbent regeneration Miniplant (0.63 MW equivalent), and a
13 kg coal/hr bench-scale system. Tests on the Miniplant combustor confirmed its
ability to achieve over 90% SO2 removal with either limestone or dolomite sorbent.
Studies of dynamic response indicated that the combustor responds much more
quickly to changes in coal sulfur content than to changes in sorbent feed rate. High
temperature/pressure particle control device testing on the Miniplant addressed:
conventional and alternative-design cyclones, a ceramic fiber filter, and a granular
bed filter. Three stages of cyclones may reduce dust loading sufficiently to protect
a gas turbine. A conventional low-pressure electrostatic precipitator and fabric
filter were also tested. Further tests on the Miniplant regenerator confirmed that
regeneration can reduce fresh sorbent feed requirements by a factor of 3 to 4. Addi-
tional sampling was completed on the Miniplant combustor and regenerator for com-
prehensive analysis of emissions. NOx control studies in the bench combustor sug-
gested that NOx emissions might be reduced by 20 to 50% through two-stage combus-
tion or ammonia injection; flue gas recirculation had little effect.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
Pollution
Coal
Combustion
Fluidized Bed
Processors
Flue Gases
Calcium Carbonates
Dolomite
Sulfur Dioxide
Dust
Cyclone Separators
Ceramics
Fabrics
Release to Public
b.lDENTIFIERS/OPEN ENDED TERMS
Pollution Control
Stationary Sources
Particulate
Fabric Filters
19. SECURITY CLASS (THil Report)
Unclassified
20. SECURITY CLASS (This page)
Unclassified
c. COSATi Field/Gioup
13B 07B
2 ID 08G
2 IB
11G
131, 07A
11B
11E
21. NO. OF PAGES
332
22. PRICE
EPA Form 2220-1 (9-73)
320
------- |