United States Industrial Environmental Research EPA-600/7-80-015a
Environmental Protection Laboratory January 1980
Agency Research Triangle Park NC 27711
Experimental/
Engineering Support
for EPA's FBC Program:
Final Report
Volume I. Sulfur
Oxide Control
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-600/7-80-015a
January 1980
Experimental/Engineering Support
for EPA's FBC Program:
Final Report
Volume I. Sulfur Oxide Control
by
N.H. Ulerich, W.G. Vaux, R.A. Newby,
and D.L Keairns
Westinghouse Research and Development Center
1310 Beulah Road
Pittsburgh, Pennsylvania 15235
Contract No. 68-02-2132
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
-------
PREFACE
The Westinghouse R&D Center is carrying out a program to provide
experimental and engineering support for the development of fluidized-bed
combustion systems under contract to the Industrial Environmental
Research Laboratory (IERL), U. S. Environmental Protection Agency (EPA),
at Research Triangle Park, NC. The contract scope includes atmospheric
and pressurized fluidized-bed combustion processes as they may be
applied for steam generation, electric power generation, or process
heat. Specific tasks include work on calcium-based sulfur removal
systems (e.g., sorption kinetics, regeneration, attrition, modeling),
alternative sulfur sorbents, nitrogen oxide emissions, particulate
emissions and control, trace element emissions and control, spent
sorbent and ash disposal, and systems evaluation (e.g., impact of new
source performance standards on fluidized-bed combustion system design
and cost).
This report contains the results of work defined and completed
under the sulfur oxide control task of the contract. Work performed on
this task was performed from January 1976 to January 1979 and is docu-
mented in the:
• Present report which presents results on desulfurization per-
formance of limestones and dolomites, prediction of desulfuriza-
tion performance for FBC plants, and sorbent attrition behavior
• Report on the "Effect of SO- Emission Requirements on Fluidized-
Bed Combustion Systems: Preliminary Technical/Economic
Assessment," issued in August 1978 (EPA-600/7-78-163)
• Report on "Regeneration of Calcium-Based S02 Sorbents for
Fluidized-Bed Combustion: Engineering Evaluation," issued in
March 1978 (EPA-600/7-78-039)
• Report on "Alternatives to Calcium-Based SO- Sorbents for
Fluidized-Bed Combustion: Conceptual Evaluation," issued in
January 1978 (EPA-600/7-78-005).
iii
-------
ABSTRACT
The desulfurization performance and attrition behavior of limestones
and dolomites are investigated for atmospheric and pressurized fluidized-
bed combustion (FBC) systems. Results from these experimental and
analytical studies are important in providing information for the design
of FBC processes to achieve energy cost objectives and environmental
requirements. Results are presented on the impact of selected FBC
operating conditions on desulfurization performance, further comparisons
of the ability of the Westinghouse desulfurization model to predict
desulfurization performance, and the development of an understanding of
sorbent attrition. The studies show that:
• PFBC systems can be operated at high temperatures (e.g.,
1000°C) or high excess air (e.g., 300%) and achieve sulfur
control without increasing sorbent requirements; this
allows for higher operating efficiency and greater flexi-
bility in turndown without sacrificing sulfur removal
efficiency.
• The agreement between fluidized-bed data and the Westinghouse
kinetic model utilizing thermogravimetric (TG) data has been
further demonstrated using data collected at atmospheric and
pressurized operation; this permits a practical, economical
method for determining sulfur sorbent requirement, given a
plant design and a specific sorbent, or for selecting the
optimal sorbent to achieve the desired sulfur removal
performance.
• Sorbent attrition screening tests indicate that sorbent
type and operating parameters will effect particle attrition.
An attrition model is presented for sorbent attrition in
-------
the bubbling bed region of the FBC, and the model is supported
by experimental data. The understanding of attrition permits
a basis for selecting sulfur sorbents and design and operating
conditions to minimize attrition and for improving fine
particle carry-over and sulfur control predictions.
These investigations provide further information for the development of
an integrated fluidized-bed combustion model incorporating sulfur
control and particulate profile models
vi
-------
NOMENCLATURE
AFBC = atmospheric-pressure fluidized-bed combustion
ANL = Argonne National Laboratories
EGAS = Energy Conversion Alternatives Study
EPA = Environmental Protection Agency
FBC = fluidized-bed combustion
PFBC = pressurized fluidized-bed combustion
STP = standard temperature and pressure
TG = thermogravimetric
TGA = thermogravimetric apparatus
TVA = Tennessee Valley Authority
DESULFURIZATION
2
A « cross-sectional area of bed, cm
a * stoichiometric reaction coefficient for the solid - 1
b * stoichiometric reaction coefficient for the gas - 1
C a S02 concentration, mole/cc
C » S02 concentration fed to batch fluidized bed, mole/cc
D = diffusion coefficient
D » pore diffusion coefficient
Ea • activation energy, kcal/mole
F =» total superficial volumetric gas flow rate, cc/s
f » mole fraction of SO,, in effluent gas
FQ • total gas flow rate, mole/s
h • static bed height, cm
K - rate constant for sulfur sorption by limestone
K - surface rate constant, cm/s
M * moles of calcium in batch fluidized bed, mole
ca
m * molecular weight
N = moles of Ca reacted per particle of limestone
n = order of reaction
P = pressure, kPa
PC02 = Partlal pressure of C02» kPa
vii
-------
PCO .1 = equilibrium partial pressure of CC^
R = gas constant
Ra = (da/dt) /(da/dt)_ ,
a r r— J.
r = particle radius
T = average pore radius
2
S = surface area of particles in bed, cm
T = temperature, K
t = time, s
w = bed weight, g
X = particle porosity
a = mole fraction of sorbent calcium sulfated
6 = volume fraction of bubble phase
e = bed voidage in emulsion fraction
p = particle density, mole Ca/cc
2
ft = aPr
6bDpC
6 = absolute temperature
T = tortuosity (2-6)
PARTICLE ATTRITION
A = extent of particle attrition
c = constant in Gonzales and Otero equation
C = extent of calcination
D = particle diameter defined by sieve analysis
D = volume-surface particle diameter
vs
F = mass fraction in cumulative size distribution
F(t) = decreasing function
f = fraction of mass smaller than 710 vim
g = gravity acceleration
g = Newton's Law conversion factor
K = unspecified constant
k = regrouping of K
viii
-------
L = rate of loss of coarses, mass loss per unit time
M = mass of large particles in a fluidized-bed
M = mass of large particles before an interval of attrition
M = mass of large particles after an interval of attrition
P = pressure; AP = pressure drop
R = attrition rate, % loss per unit time
R. = rate of heating
o
t = time
t = reference time interval
o
S = percent of possible sulfation
T = temperature
U = superficial gas velocity
U , = minimum fluidization velocity
X = fraction of calcium in bed solids
Y = mass fraction loss of CO on ignition
Z = depth into the bed, measured from the surface
p = gas velocity
p = particle density
S
a = particle strength
V = gas viscosity
Subscripts
1 • stone after attrition
o = original stone
f = filter
ix
-------
ACKNOWLEDGEMENT
We want to express our high regard for and acknowledge the con-
tribution of Mr. D. B. Henschel who served as the EPA project officer.
Mr. P. P. Turner and Mr. R. P. Hangebrauck, Industrial Environmental
Research Laboratory, EPA, are acknowledged for their continuing con-
tributions through discussions and support of the program.
We thank Mr. R. E. Brinza, Mr. J. Capozzi, Ms. L. J. Cwynar,
Ms. C. A. Hill, and Mr. W. F. Kittle for performing sample analyses and
carrying out the thermogravimetric and batch fluidized-bed experiments
on desulfurization. We thank Mr. A. W. Fellers for his work on the sorbent
attrition test program.
-------
TABLE OF CONTENTS
Page
1 INTRODUCTION 1
2 SUMMARY AND CONCLUSIONS 3
3 RECOMMENDATIONS 9
4 SULFUR OXIDE CONTROL - CALCIUM-BASED SORBENTS 11
Desulfurization 11
Particle Attrition 56
Sorbent Regeneration 115
5 SULFUR OXIDE CONTROL - ALTERNATIVE SORBENTS 118
6 REFERENCES 120
APPENDICES
A SULFUR OXIDE REMOVAL DATA BASE AND MODEL 125
B SORBENT INFORMATION AND TG RATE DATA 145
C FLUIDIZED-BED DATA 179
D A MODEL FOR PARTICLE ATTRITION BY ABRASION IN THE UPPER
ZONE OF A FLUIDIZED BED 187
E 3.5-in FLUIDIZED-BED TEST SYSTEM 225
xi
-------
LIST OF TABLES
Page
1 TG Experimental Program Outline: Operating Range 13
Impact on Desulfurization Performance
2 The Effect of Temperature on Sorbent Utilization at 20
1013 kPa (10 atm)
3 Relative Effectiveness of Large-Grained Dolomite in 41
Sulfation
4 Batch Fluidized-Bed Experiments 45
5 Summary of Models Used to Analyze Fluidized-Bed Data 47
6 Range of Values of Test Variables in Several Test 62
Systems
7 Percent of Solids Attrited in Four Hours 65
8 Summary of Attrition Test Data Statistics 65
9 Description of Test Conditions 68
10 Attrition Test Data 69
11 Percentages of Fines Formed during Attrition Testing . 73
and Percentages Attributable to Fluidization Only
12 Relation between Extent of Attrition and Degree of 88
Roundness as Judged by Six Observers
13 Comparison of Mean Sizes of Grove Limestone Particles 91
Calcined and Untreated
14 Effect of Resieving a Single Size Fraction of Tymochtee 93
Dolomite
15 High and Low Levels of the Independent Variables 100
16 The Results of High-Temperature Attrition Testing of 106
Grove 1359
17 Factorial Model Coefficients Describing Attrition of 107
Grove 1351
18 Dependence of Solids Specific Surface on Time of 115
Fluidization
19 Distribution of Particle Surface Area for Various 116
Particle Sizes after 15 Minutes of Fluidization of Grove
Limestone at 25°C
xii
-------
LIST OF FIGURES
Page
1 The Pressurized TG Apparatus 16
2 Schematic Diagram of the Pressurized TG System 17
3 The Effect of Temperature on the Pressurized Sulfation 21
of Limestone and Dolomite
4 The Effect of High Temperature on the Pressurized 21
Sulfation of Small and Large Limestone Particles
5 The Effect of Temperature on the Pressurized Sulfation 22
of Greer Limestone (74-149 ym particles)
6 The Effect of Temperature on the Pressurized Sulfation 22
of Greer Limestone (420-500 pm particles)
7 The Effect of Temperature on the Pressurized Sulfation 22
of Greer Limestone (2380-3360 um particles)
8 The Effect of C02 Pressure during Calcination on S02 25
Emissions from a Fluidized Bed
9 The Influence of Oxygen Partial Pressure on the Rate 29
of Dolomite Sulfation (101.3 kPa/1 atm)
10 The Influence of Oxygen Partial Pressure on the Rate 29
of Dolomite Sulfation (1013 kPa/10 atm)
11 The Influence of Excess Air Level on the Pressurized 30
Sulfation of Uncalcined Limestone
12 The Influence of Sorbent Residence Time on the Sulfation 30
of Greer Limestone (1013 kPa/10 atm)
13 The Influence of Sorbent Residence Time on the Sulfation 33
of Limestone 1359 (1013 kPa/10 atm)
14 The Influence of Sorbent Residence Time on the Sulfation 34
of Greer Limestone (101.3 kPa/1 atm)
15 The Influence of Sorbent Residence Time on the Sulfation 34
of Limestone 1359 (101.3 kPa/1 atm)
16 Comparison of Sulfation Rates at 101.3 and 1013 kPa 36
(1 and 10 atm) Pressure (Greer Limestone)
xiii
-------
LIST OF FIGURES (Continued)
Page
17 Comparison of Sulfation Rates at 101.3 and 1013 kPa 36
(1 and 10 atm) Pressure (Tymochtee Dolomite)
18 The Effect of Pressure on the Sulfation Rate of 37
Tymochtee Dolomite
19 TG Sulfation of Canaan Dolomite (420-500 ym particles) 39
20 Sulfur Penetration at Periphery and Grain Boundaries of 40
500 ym Particle of Canaan Dolomite (^300 ym grains)
21 TG Sulfation of Canaan Dolomite (74-149 ym particles) 41
22 Schematic of Batch Fluidized-Bed Reactor 43
23 Comparison of Fluidized-Bed Models (Bellefonte limestone) 49
24 Comparison of Fluidized-Bed Models (Carbon limestone) 49
25 Comparison of Rate Constants Derived from Fludizided-Bed 51
Data (Model 1) and TG Data: Limestone 1359
26 Comparison of Rate Constants Derived from Fluidized-Bed 51
Data (Model 1) and TG Data: Carbon Limestone
27 Comparison of Rate Constants Derived from Fluidized-Bed 52
Data (Model 1) and TG Data: Brownwood Limestone
28 Comparison of Rate Constants Derived from Fluidized-Bed 52
Data (Model 1) and TG Data: Ames Limestone
29 Comparison of Rate Constants Derived from Fluidized-Bed 53
Data (Model 1) and TG Data: Bellefonte Limestone
30 Comparison of Rate Constants Derived from Fluidized-Bed 53
Data (Model 1) and TG Data: Mississippi Limestone
31 Comparison of Rate Constants Derived from Cambridge 54
Fluidized-Bed Data (Model 1) and Westinghouse TG Data
32 Attrition Rate Dependence upon Stone Type and Atmosphere 67
33 Gas Velocity and Temperature Patterns in the Attrition 71
versus Time Tests
34 Effect of Duration of Fluidization on Extent of Attrition 72
in Fluidization of Grove Limestone at 815°C
35 Temperature History When Cold Grove Limestone Is Added to 75
815°C 3.5 cm Attrition Cell (Runs Al, Al Repeated, A2)
36 Measurement of Perpendicular Dimensions for Measurement 78
of Particle Shape
37 Grove Limestone Particles before and after Fluidization 80
at U - U = 30 cm/s for 329 Hours
mf
xiv
-------
LIST OF FIGURES (Continued)
Page
38 Sorbent Micrographs - Grove Limestone 82
39 Particle Size Distributions before and after Hot 37
Fluidization of Grove Limestone, Run A-9
40 Mean Rankings of Particle Angularity 89
41 Mean Rankings of Particle Angularity for the Effect-of- 89
Duration Tests
42 Apparent Swelling of Sausage-Shaped Particles 90
43 Micrographs of Grove Limestone before and after 92
Calcination
44 Size Distributions of Tymochtee Dolomite after Hot 96
Fluidization at 100 and 1000 kPa
45 Attrition Test Cell 98
46 Test Procedure for Attrition Testing of Grove Limestone 99
for Effects of Grain Size, Temperature, and Sulfation
47 Determination of Umf from the AP-U Curve 101
48 Flow Diagram of Sorbent Attrition Test System 102
49 Attrition Test Cell 103
50 Leak in 10.3-cm Fluidized Bed 108
51 Size Frequency Plots for Grove Limestone 110
52 Cumulative Distributions for Grove Limestone 110
53 Flow Diagram for Room-Temperature Fluidized Bed 113
54 Photomicrograph of Elutriated Grove Limestone Recovered 113
from the Balston Filter
xv
-------
1. INTRODUCTION
The design of fluidized-bed combustion (FBC) systems for electric
power generation, industrial steam or process heat, or cogeneration
applications depends on an understanding of desulfurization in order to
achieve energy cost and environmental objectives and requirements.
Fluidized-bed combustion systems can offer energy cost and environmental
advantages when compared with alternative process choices if the system
design incorporates an understanding of the component/subsystem per-
formance, limitations, and available trade-offs as a function of
operating and design parameters. Limestone and dolomite desulfurization
performance and attrition behavior represent two areas that must be
understood when selecting design and operating parameters to achieve the
process objectives and environmental requirements at the lowest energy
cost. The results reported in this document extend our previous under-
standing of these phenomena to provide a basis for FBC design and per-
formance evaluation.
A data base of over 700 atmospheric-pressure and 100 pressurized
thermogravimetric (TG) tests is now available, based on experiments
performed at Westinghouse. These results and the results from other
investigators have been used to develop an understanding of desulfuri-
zation phenomena. The quantitative prediction of sorbent desulfurization
performance was shown in our earlier work under contract to EPA utilizing
reaction rate constants derived from TG data and the Westinghouse FBC
desulfurization model. The selection of the work reported here is based
on an analysis of the available laboratory and plant data and systems
-------
evaluation studies which identified research and development needs.
The results of work in five areas are presented: The impact of selected
FBC operating conditions on desulfurization in atmospheric (AFBC) and
pressurized fluidized-bed combustion (PFBC) systems, the comparison of
reaction rate constants derived from TG data and batch fluid-bed data,
further testing of the Westinghouse desulfurization model against
available bench-scale and pilot plant data, screening tests to assess
the effect of sorbent type and operating conditions on sorbent attrition
in the fluidized bed, and the development and experimental confirmation
of an attrition model for attrition in the "bubbling bed."
-------
2. SUMMARY AND CONCLUSIONS
Experimental and analytical studies were carried out to investigate
desulfurizatlon performance of limestones and dolomites for AFBC and
PFBC systems to extend our capability for predicting desulfurization
performance and to initiate a program to understand particle attrition
in fluidized-bed combustion systems.
DESULFURIZATION PERFORMANCE
Westinghouse has conducted an extensive program, under sponsorship
of EPA and other organizations, to study limestone and sulfur sorption
utilizing thermogravimetric analysis test facilities. Previous results
have indicated areas where additional data and analyses were needed to
evaluate performance. The areas selected for this study included the
impact of selected operating conditions on desulfurization:
• The effect of high temperature (e.g., 1000°C) PFBC opera-
tion, which is of interest for achieving high plant
efficiency through increased turbine inlet temperatures
and for providing turndown flexibility.
• The effect of oxygen concentration, which is of interest
for assessing the desulfurization performance for adiabatic
FBC process concepts utilizing high excess air (e.g., 300%).
• The performance of uncalcined, impure limestones for PFBC
operation, which is of interest for increasing the potential
sorbents available for commercial applications.
• The effect of sorbent residence time, which is of interest
for assessing the impact of long (e.g., 10-20 hours)
exposures at operating temperature in commercial plants
and for interpreting TG or batch fluid-bed data taken at
short residence times.
-------
• The use of large-grained dolomites, which is of interest
for increasing the potential sorbents available for
application.
Conclusions from the experimental test program are that:
• Desulfurization performance for PFBC operation at 1013 kPa
(10 atm) pressure will be maintained or improved at high
temperature (900-1000°C) operation. Sulfur removal of
85 and 90 percent can be achieved at temperatures as high
as 900 to 1000°C without increased sorbent feed requirements
(over 800-850°C operation) for most sorbents.
• Desulfurization performance will be maintained at high
excess air operation. Dolomite sulfation is zero order
in oxygen concentration throughout the range of FBC opera-
tion (0.75-16% oxygen) at pressures of 101.3 and 1013 kPa.
Sulfur removal of 85 and 90 percent can be achieved at high
excess air levels without increasing sorbent feed require-
ments over low excess air operation.
• Impure limestones, such as Greer, can be effective sorbents
in the uncalcined form, provided that the carbon dioxide
(CO-) partial pressure is not much greater than the equili-
brium for calcination. This result increases the sorbents
available for FBC systems where operating temperature and
pressure ranges result in operation with calcium carbonate
(CaC03).
• The residence time of a sorbent at the operating tempera-
ture may change the sulfation kinetics for some sorbents
from the first-order relationship observed at initial
reaction periods. The effect of sorbent residence time
on desulfurization performance should be tested when TG
data are used to project calcium feed requirements.
-------
• Large-grained dolomites tested at elevated pressure show
that low calcium-to-sulfur ratios can be achieved only if
fine particles (e.g., 74 to 149 ym) are utilized. Results
are consistent with plant test data.
DESULFURIZATION PERFORMANCE PREDICTIONS
The desulfurization performance of fluidized-bed combustion systems
can be predicted by utilizing a kinetic model for sulfur dioxide (SO-)
1 L
capture previously developed by Westinghouse. This model - using rate
constants derived from TG data - is capable of projecting sorbent require-
ments as a function of FBC operating and design parameters. The
importance of selecting proper fluid-bed combustor operating conditions
(e.g., gas velocity, bed depth, sorbent particle size) was shown in a
2
previous Westinghouse study. Further confirmation of the accuracy of
model projections was an objective under this contract. Comparison of
the models' predictions with available bench-scale and pilot plant FBC
2
data were previously reported, with additional comparisons presented in
this report, with specific focus on the 90 percent desulfurization
obtained in the pressurized Exxon miniplant. The use of batch fluidized-
bed tests as an alternative to TG tests to obtain rate constant, data
was also investigated.
The conclusions from the experimental and analytical work on pre-
dicting desulfurization performance are that:
• The agreement between fluidized-bed data and the kinetic
model utilizing TG data has been further demonstrated with
data collected at 1013 kPa (10 atm) pressure as well
as at atmospheric pressure, and at sulfur removal efficien-
cies of up to 90 percent.
• The ability to compare predictions with plant performance
is limited by the availability of complete pilot plant
data (e.g., particle size distribution, fraction of inert
particles in the bed, bed expansion, etc.) and the accuracy
of pilot plant data
-------
• Batch fluidized-bed data can also be utilized to obtain
reaction rate constants for predicting performance. The
calculation of the rate constant requires the use of a
fluid-bed model that represents the test unit. Rate con-
stants from batch fluidized-bed tests performed by
Westinghouse with six sorbents and by Cambridge University
with one sorbent compare favorably with rate constants from
TG data.
SORBENT ATTRITION
The attrition of sorbent particles will effect the selection of
sulfur sorbents, the desulfurization performance given a sulfur sorbent,
and the design of the particulate control system for fine particles
to achieve process (e.g., turbine tolerance) and environmental require-
ments. An understanding of particle attrition in FBC systems is important
for the design and operation using limestones or dolomites and is critical
3
for regenerative sulfur control processes utilizing alternative sorbents.
An approach to understanding particle attrition in FBC systems was
developed as part of the present effort. The objective was to develop
a unified attrition model that would incorporate the separate attrition
mechanisms occurring in a FBC system. This model will be integrated
into the sulfur removal system and particle profile models being developed
under this contract and under contract 68-02-3110.
The initial effort was to review the available literature, to perform
screening tests to gain perspective on the effect of sorbent type and
operating conditions on attrition in the fluidized bed, and to develop
and confirm an attrition model for the attrition phenomena due to the
"bubbling bed" behavior.
The conclusions from the review of available information, our
experimental test programs, and our modeling work are that:
• The understanding of particle attrition in fluidized-bed
processes is fragmented and incomplete - comprehensive
understanding or models for the attrition mechanisms are
not available.
-------
Screening tests on the effect of sorbent type and operating
conditions on attrition show that sorbent type, particle
atmosphere and temperature, particle composition (degree
of calcination and degree of sulfation), duration of
fluidization, residence time at temperature, rate of heating,
and gas jets will effect the extent of particle attrition.
A model was developed to relate the rate of attrition in
the bubbling zone of the FBC above the influence of any
grid jets. The relationship is
RZ
U-U
mf
- [P(t)
m
where
g = gravity acceleration
p - particle density
5
s
U
U
mf
particle strength
superficial gas velocity
minimum fluidization
velocity.
g = Newton's Law conversion factor
c
R = attrition rate
Z = depth into the bed,
measured from the surface
F(t) = a transient function describing
the variation of attrition rate
with time. F(t)-K) as t -*• »,
dF(t)/dt <_ for t > 0
• Experimental results from our test program and available
literature data support the model. Attrition by the
bubbling bed results in the formation of fines but does
not alter the basic particle shape.
• The attrition rate in the bubbling zone of a fluidized-
bed combustor can be controlled by choosing a weak or
strong sorbent and by specifying bed depth, gas velocity,
and particle diameter as it affects U ^.
• Attrition from jets, thermal shock, cyclones, and impact
devices that may be incorporated in the system design are
expected to result in the greatest extent of attrition;
these and other sources, such as the bubbling bed, may all
-------
be important in the production of fines (<3 urn), which are
important for process considerations and environmental
impact.
-------
3. RECOMMENDATIONS
The following recommendations are made for further work on desul-
furization and sorbent attrition:
• Continue the work to compare model predictions with
available bench-scale, pilot plant, and commercial plant
performance for confirmation and identification of suggested
areas for improvement. The projection of high sulfur
removal efficiencies (>90 percent) should be emphasized.
• Extend the desulfurization model to particle size distri-
bution and particle size residence times, incorporating
particle attrition and carry-over models.
• Comprehensive reporting of pilot plant and commercial
plant data is required in order to extend the capability
for desulfurization performance of the present predictive
model,to develop comparisons of attrition performance with
model predictions,and to permit accurate interpretation
of plant results.
• Extend the development of fundamental gas-solid modeling.
Development of a reaction model of the sulfation of lime-
stones and dolomites could result in innovative methods
for improving sorbent utilization through an understanding
of the reaction mechanisms; improved tests to determine
reaction performance; or, as an ultimate objective, the
prediction of sorbent utilization based on physical and
chemical properties of the sorbent.
• Investigate techniques to improve sorbent performance based
on experimental and modeling work.
-------
• Review and extend the test procedures developed for sorbent
selection in FBC applications. Principal considerations
would be the sorbent reactivity for high sulfur removal
requirements and attrition characterization.
• Develop an understanding of attrition phenomena in FBC
systems. Priority areas for study are sorbent and fuel
attrition resulting from jets (e.g., grid), thermal shock,
solids transport, and cyclones.
• Develop an integrated attrition model that incorporates
the important attrition mechanisms in FBC systems. The
model will permit extension of the desulfurization per-
formance model and the particle profile model for estimat-
ing particle size distributions and loadings through the
FBC system.
Recommendations for work on sorbent regeneration using calcium-based
and alternative sorbents and for further systems analyses are presented
2-4
in companion reports previously issued.
10
-------
4. SULFUR OXIDE CONTROL - CALCIUM-BASED SORBENTS
DESULFURIZATION
The impact of sulfur removal on the operation of a fluidized-bed
power generation unit depends on the sorbent calcium utilization in
desulfurization. The reaction kinetics of limestone with S0? must be
understood in order to develop a rational basis for sorbent selection
and to maximize calcium (Ca) utilization in fluidized-bed combustion.
The evaluation of sorption kinetics was achieved by mathematically
modeling the fluid-bed combustor and performing TG laboratory experi-
ments simulating fluidized-bed operating conditions to define the
kinetic rate constants for the model, investigate data gaps, and study
the effect of sorbent type on desulfurization.
Westinghouse has conducted a substantial TG program studying lime-
stone and dolomite sulfur sorption under the sponsorship of EPA and
other organizations. ' ' Previous results have indicated areas in
which additional data and analysis were needed to evaluate the sulfur
removal performance of sorbents. These specific areas, addressing the
impact of FBC operating conditions on desulfurization in AFBC and PFBC^
were investigated.
The desulfurization performance expected in fluidized-bed units can
be projected by using rate constants derived from TG data. The fluidized*
bed model used and the method of making the projections are summarized
in Appendix A. The models' projections were compared to data obtained
from bench-scale and pilot-plant fluidized-bed units. In particular, we
modeled the achievement of more than 90 percent desulfurization in high-
pressure work at the Exxon miniplant.
11
-------
Previous Work Perspective and Approach
FBC Operating Range Impact on Desulfurization Performance
Previous TG work has identified the following subjects, where little
information is available, as important in understanding the impact of
various operating conditions on desulfurization. The specific data sets
collected, and their scope and operating conditions, are outlined in
Table 1.
High Temperature Operation. The range of fluidized-bed combustion
conditions has been generally considered to lie in the temperature range
of 730 to 950°C. The Energy Conversion Alternatives Study (EGAS),
however, showed the desirability of extending the operating range to
1010°C for pressurized operation. Almost no data are available to
12
project sulfur removal efficiency at these temperatures. Work at Exxon
showed that desulfurization adequate for achieving EPA sulfur oxide (SO )
A
emission limits could be achieved at operating temperatures above 980°C
in PFBC.
Previous pressurized TG studies at Westinghouse, using Tymochtee
dolomite, showed only a slight decrease in the reaction rate for sulfa-
tion in the range of 843 to 954°C. Pressurized studies with Limestone
1359 gave ambiguous results, but the data showed only a slight decline
in calcium utilization, from 40 to 35 percent in the range of 900 to
950°C.
Limestone usage in PFBC may be practical at temperatures above 950°C,
since calcination would occur. The stable form of dolomite at high tem-
peratures would also be in the fully calcined (CaO'MgO) rather than in
the half-calcined (CaCO-'MgO) form. The effect of temperature on the
i
pressurized sulfation of limestone and dolomite, therefore, was studied
in data sets 1 and 2. Since there is also little data available on
large (>2000 ym) and small (<100 pm) sorbent particles, the temperature
effect was examined using various particle sizes of limestones.
12
-------
Table 1
TG EXPERIMENTAL PROGRAM OUTLINE:
OPERATING RANGE IMPACT ON DESULFURIZATION PERFORMANCE
Data Set
1
2
3 .
4
5
6
7
Scope
Effect of Temperature on
Pressurized Limestone
Sulfation
Effect of Temperature on
Pressurized Dolomite
Sulfation .
Effect of Q£ Concentration
on Atmospheric Pressure
Desulfurization of
Dolomite
Effect of 02 Concentration
on Pressurized Desulfur-
ization of Dolomite
Effect of Excess Air on
Uncalcined Limestone
Sulfation
Effect of Sorbent
Residence Time on
Desulfurization
Large-Grained Dolomite
Performance
Sorbent (s)
Greer
limestone
Dolomite 1337
Tymochtee
dolomite
Tymochtee
dolomite
Greer
• limestone
Pressure,
kPa
1013
1013
101
1013
1013
Greer limestone 101
Limestone 1359
1013
Kaiser dolomite 101
Canaan dolomite 1013
Particle Size,
Vim
420-500
74-149
149-420
2380-3360
420-500
1000-1190
1000-1190
1000-1190
1000-1410
420-500
74-149
% Excess Air
300
300
15.2 kPa C02
(2-16% 02)
15.2 kPa C02
(0.7-16% 02)
15-200
20
20
Temperature »
°C
840-1010
840-1010
815
815
815
815
815
-------
High and Low-Oxygen Concentrations. Previous tests showed no
increase in the sulfation rate of dolomites at high partial pressures of
oxygen (0.) (4-11%). Since stoichiometric utilization of the calcium in
dolomites has been achieved, however, the influence of oxygen concentra-
tion on dolomite sulfation through high levels of calcium conversion is
of interest. The effect of oxygen concentration over a wide range
(0.75-16%) was studied in data sets 3 and 4.
The Effect of Excess Air on Uncalcined Limestone Sulfation. Results
from one limestone in the uncalcined form showed that the sorbent
absorbed very little SCL. The limestone tested, however, was a very pure
stone with small, interlocking grains. Since the sulfation of half-
calcined dolomite proceeds readily, and since most pressurized operating
conditions do not permit complete calcination of the sorbents, the use
of an impure limestone in the uncalcined form was investigated (data
set 5).
Sorbejit Residence Time. The influence of prolonged exposure of
sorbents to FBC temperatures has not been studied on the TGA. TG sulfa-
tions have been carried out in 0.5% SCL. At this SO- concentration the
sorbent sulfation time is convenient for laboratory study, about two
hours. In order to simulate the longer residence times of sorbents in
fluidized beds, TG sulfations were carried out in lower SO- concentra-
tions (data set 6). . .
Large-Grained Dolomites. The worst sorbents tested on the TGA have
been large-grained dolomites. The possible use of these sorbents in
powderized, fine-particle form was tested in data set 7.
Prediction of Desulfurization Performance
The qualitative information gained from TG studies has been greatly
enhanced by the development of models that permit the quantitative pre-
diction of sorbent performance in fluidized-bed units. We had previously
developed a model to project fluidized-bed desulfurization performance
using reaction rate constants derived from TG data. The model was tested
14
-------
here against bench-scale and pilot-plant data. In addition, rate
constants derived from TG data were compared to those derived from batch
fluidized-bed tests we had carried out, using several models for gas/
solid dynamics in the batch unit.
FBC Operating Range Impact on Desulfurization Performance
TG Experimental Facility
The fractional utilization of a sorbent (fraction of the calcium
oxide [CaO] sulfated) was determined as a function of time by sulfating
sorbents suspended from a Du Pont 951 thermogravimetric balance. This
5 8
system has been described previously. ' The gas flow path and the
reaction tube geometry have been modified to improve the temperature
measurement. Linear mass flow controllers have replaced rotameters for
flow-rate measurement and control. Schematics of the apparatus are given
in Figures 1 and 2.
TG experiments were run according to the following procedure:
• Size limestone by sieving.
• Suspend a 20 tng sample in a platinum mesh basket from the
balance arm. Place the thermocouple into the basket, about
1 mm from the sample.
• Pressurize the system.
• Heat the sample at a rate of 10°C/min to reaction tempera-
ture in the reactant gas, minus SCL, flowing at a rate of
2 1/min at standard temperature and pressure (STP).
• After complete calcination or half-calcination as indicated
by a stable sample weight, introduce S02 into the reactant
gas mixture.
• Monitor the sample weight gain as a function of reaction time.
TG Experimental Program
Twelve sorbents were used throughout the TG experimental program.
Sorbent quarry and supplier information and a summary of sorbent analyses
are given in Appendix B. TG rate data referenced in the report are also
appended.
15
-------
ff1
*t
B16
l_
B17
B18
Bl
B2
B3
B4
B5
B6
B7
B8
B9
BIO
— •-'
i_
Iff
*t
^_I ^
\
\
\
\
\
x
\
\
'-n'
\
iy o u o o'
/^ B14
/ _n
~\
po-
bo
B-
V
\
/
^" ^...- •!
^SJ B4
B2
rtft oon^oc
' ^ ^r/ 7"
- o o o o y
5D6w7B5////
A l.™//^W/
1 5A«-HiU
IT" f / H * f / / /
-1 ///Biiy//
.. r> o o
E
(OftOpOOoJ
i O O
\\w
B9
A\\\\\V
000-
////////S
— — — ' — DO oououy uuuuuwuuw I
10"
y
c
~?r^
/
o
A.
S"
0-
/
._
/
— •>
^ —
Figure 1 - The Pressurized TG Apparatus
KEY
Stainless Steel Pressure Vessel Bll Sample Thermocoi
External Cooling Coil B12 Baffle Assembly
Internal Cooling Coil B13 Reaction Tube R
TG Balance Housing B14 TG Bell Jar
Reaction Zone Furnace B15 Flexible Metal 1
Reaction Zone Thermocouple B16 Atmospheric-Pre
Preheat Zone Furnace B17 TG Balance Elec
Preheat Zone Thermocouple
Quartz Reaction Tube B18 Inert Pur8e Gas
Sample Basket B19 ExhauSt Gas ^
B20 Reaction Gas In:
s
\
\
"\
\
\
\
\
\
__-
^
1"
jpl
&ta
Exh
ssu
tri
In
Let
Let
jl
3=wB19
J
C^~B2°
a-
43
e
ining Ring
aust Hose
re Vent
cal Feed-
Let
16
-------
Dwg.
Relief Valve
Check Valve -W-
On-Off Valve -&*~
Control Valve I 1
Transducer (T
Filter
Pressure
Regulator
Pressure
Gauge
©
Atm Vent
0>
"5.J*
E
CO
CO QC
(!)
Vent
1 T T T
^u ouuiiuy wui ~-
¥
1 H 0 Pool inn In — —
TG Pressure
Choll
jucll
-1 L—
Purge Gas
Flow
Controller
i
J
G
— s
~Q
T 2
(u
N2( Purge)
o
-/
1
/!
V
1
F D
TG Console
rtm
ent
Data
Acquisition
System
Figure 2 - Schematic Diagram of the Pressurized TG System
-------
Temperature. Previous work with Limestone 1359 calcines has shown
that the initial sulfation rates at 101.3 kPa (1 atm) pressure at tempera-
tures from 750 to 950°C are nearly identical. After 20 percent utiliza-
tion of the calcium, however, the reaction rate varied. The utilization
of the stone after one hour of reaction time improved with increasing
temperature, up to 860°C. Further increasing the temperature caused
13 14
sorbent utilization to decrease. Others ' have also found this phe-
nomenon of an optimum temperature for limestone utilization at atmospheric
pressure in laboratory- and bench-scale studies. Westinghouse TG results
at 1013 kPa (10 atm) pressure, however, showed that the extent of lime-
stone sulfation increased with increased temperatures up to 900°C;
further increasing the temperature to 950°C gave inconclusive results.
Operating fluidized-bed combustors at high temperatures is desirable,
since at high pressures higher temperatures are required to calcine
CaCO«. (Increased turbine inlet temperature will also improve turbine
efficiency.) The use of limestone in high-pressure (1013 kPa/10 atm)
FBC, therefore, is dependent on its performance at higher temperatures
(>900°C). We, therefore, assessed the effect of temperature on the
pressurized sulfation of limestone and dolomite.
Thermogravimetrie experiments studying the influence of temperature
on the pressurized sulfation of limestone and dolomite were carried out
at 1013 kPa (10 atm) pressure. Greer limestone and Dolomite 133? were
the sorbents tested. Particles of 420 to 500 um diameter were precalcined
by heating them at a rate of 10°C/min up to the sulfation temperature in
4.3% C02 and 15.8% 02 in nitrogen (N2). The oxygen and C02 concentra-
tions are representative of a 300 percent excess air level in combustion.
This condition was chosen so that sorbents would be in the fully calcined
state throughout the range of temperatures tested. The sorbents were
then sulfated in the same atmosphere, plus 0.38% S0_.
Improved sorbent utilization was observed for both limestone and
dolomite with increased temperature, up to 1000°C. Only a few runs were
18
-------
done at temperatures greater than 1000°C because of furnace limitations.
These runs, however, indicate that maximum sorbent utilization occurs at
around 1000°C at 1013 kPa (10 atm) pressure. Table 2 summarizes the
results. Figure 3 illustrates the temperature effect. The extent of
sulfation was compared at the point at which the reaction rate was 0.1%
calcium sulfating per minute. Although the sorbent is not saturated at
this point, its rate of reaction is too slow for additional effective
sulfur capture.
We suspect that the scatter in the data from Dolomite 1337 runs is
the result of gas leakages in the TG system. Inlet and outlet seals were
readjusted after this TG run series was completed.
The increased utilization of limestone with temperature (up to
1000°C) in pressurized TG sulfation was tested with particles of Greer
limestone in three additional size ranges, 74 to 149 ym, 149 to 420 urn,
and 2380 to 3360 ym. In agreement with the results obtained with 500 ym
particles, the sorbent did not lose reactivity at high temperatures
(see Figure 4). Tests on the large particles of Greer limestone, however,
showed some scatter in the results. The larger particles of Greer lime-
stone reached 25 to 50 percent of the utilization obtained by the 500 urn
particles. No improvement in extent of sulfation occurred when the par-
ticle size was reduced to 149 to 420 ym. Little improvement in sorbent
utilization would be expected, however, if the actual size of the sample
was near 420 ym. The 74-to-l49-ym particles were utilized 25 to 35 per-
cent more than the 500 ym particles.
The TG rate data used to generate the plots are shown in Figures 5
through 7. The initial rates of reaction are fairly insensitive to tem-
perature variation. Only after about 30 percent sulfation does the
variation in sulfation rates with temperature become evident. At this
extent of reaction, diffusion of S0? through the sorbent's pores would
have an effect on the reaction rates.
19
-------
Table 2
THE EFFECT OF TEMPERATURE ON SORBENT UTILIZATION AT 1013 kPa (10 atm)
Particle Size - 420-500 ym
Sulfation Atmosphere - 0.38% S02> 4.3% C02, 15.8% 02, balance N2
Precalcination - Heated at 10°C/min. up to sulfation temperature in 4.3% C02, 15.
Pressure - 1013 kPa (10 atm)
balance N-
Run No.
76-122
76-123
76-124
76-125
76-126
76-127
76-128
P-23
Greer
Limestone
Temperature » °C
842 +
896 +
866 +
955 +
928 +
978 +
980.3
8
4
8
2
2
3
+ 0.4
1011 + 1
% Sulfation*
43
55
55
67
60
68
65
55.5
Dolomite 1337
Run No.
P-7
P-10
P-16
P-8
P-17
P-15
P-ll
P-13
P-14
P-18
Temperature, °C
840
868.8 + 0.4
901 + 1
924 + 2
929 + 3
951 + 3
972 + 1
976 + 1
979
1005 + 2
% Sulfation*
60
65
58
> 67
67
77
> 75
88
83
85
*% conversion of CaO to CaSO, when the sulfation rate falls below 0.1% Ca/min.
-------
Curve 690958-A
100
90
80
_c
.E 70
^
f—I
0- 60
Al
i 50
oe
_o>
i 40
5 30
20
10
"^ ' ' I '
o Dolomite 1337
a Greer Limestone
T
TG Sulfation at 1013 kPa (10 atm)
Sorbents 420-300 \an
Sulfation Atmosphere -0.38%S02, 4.3% C02,
15.8%02, balance N2
Precalcination- heated at 10°C/ /nin up to
Sulfation Temperature in 4.3% C02, 15.8%
0^ balance
800
' 'LI.1 I. J. .1, .L
20 40 60 80 20 40 60 80
900
Temperature, °C
1000
Figure 3 - The Effect of Temperature on
the Pressurized Sulfation of
Limestone and Dolomite
90
80
70
60
3
* 50
Al
^
I *
£
5 *
3
20
10 -
Particle Size um:
• 74-149
A 149-420
o 420-500
02380-3360
Curve 691Z65-B
Greer limestone
1013k Pa (10 atm)
Sulfation Atmosphere - a 38* SO,, 4.3* CO,.
15.8* Oz. Balance ^ i i
Precalcination - Heated at 10°C/min up to
Sulfation Temperature in 4.3* CO., IS. 8* 0.,
Balance N •
800
80 w 2040 60 80 1MO 20
Temperature, "C
Figure 4 - The Effect of High Temperature on
the Pressurized Sulfation of Small
and Large Limestone Particles
-------
Greer Limestone, 74-149 pm particles. 1013 IcPa
Calcined at T in 4.3Z CO,, 15.8% 0
Sulfated at T in 0.38% SO.,
• Run P49, T - 948°C
* Run P52, T - 997'C
2
15.8* 0,
Ti Tz Ti T«
Ts
rr
FRACTION SULFATED
Figure 5 - The Effect of Temperature on the Pressurized Sulfation
of Greer Limestone (74-149 |im particles)
Ift
Greer Limestone, 420-500 un, 1013 kPa
Calcined Nonlsothermally up to T In
4.3Z C02, 15.81 02
Sulfated at T In 0.381 SO., 4.3Z CO,,
15.8Z 02
• Run 122, T - 842'C
* Run 123, T • 896'C
X Run 125, T - 955'C
+ Run P23. T • 1010'C
V
.1
.7
Fraction Sulfated
Figure 6 - The Effect of Temperature on the Pressurized Sulfation
of Greer Limestone (420-500 ym particles)
Creer Limestone, 2380-3360 urn particles, 1013 kP«
Calcined at T In 4.3Z O>2, 15.8Z Oj
Sulfated at T In 0.38J SOj, 15.8X Oj
• Run P20, T - 894'C
* Run P21, T • 954'C
x Run P22, T • 1010'C
+ Run P27, T • 1010'C
-2
Tl T2
FRACTION SULFATED
.3
Figure 7 - The Effect of Temperature on the Pressurized Sulfation
of Greer Limestone (2380-3360 ym particles)
22
-------
Several mechanisms have been postulated to explain the occurrence
of an optimum sulfation temperature at atmospheric pressure and the
absence of such a temperature (up to 980°C) under pressure.
1) The Sulfite Mechanism
13
Moss has proposed a mechanism of sulfation based on the formation
of sulfite (S03).
1) S02 + 1/2 02 ;—* S03
2) S03 + CaO -»• CaSO^.
Using the Arrhenius equation and the observation that the rate is propor-
tional to the gas concentration, we have for reaction 2
rate - K[S03]n e~|| .
The exponential term has the rate increasing with temperature, but the
equilibrium concentration of S0_ from reaction 1 decreases with tempera-
ture. Thus, an optimum temperature for sulfation should be observed, as
it has been in the atmospheric case. The mechanism predicts a shift to
higher optimum temperatures in pressurized sulfation because of increas-
ing partial pressures of SO. as pressure is increased. Thus, one would
expect that an optimum temperature would be observed in pressurized
testing as well, when the testing is extended to higher temperatures.
More recent work by Burdett, however, argues that the SO., concen-
tration in fluid beds is far from the equilibrium level. It is likely
that raising the temperature in the bed increases the oxidation rate and
thus raises the local SO. concentration. As the temperature rises,
therefore, the sulfation tends to occur at the outer edge of the stones,
halting diffusion, lowering sorbent utilization, and producing a maximum
in the sulfur retention/temperature curve.
2) The Sulfide/Sulfate Mechanism
An alternative explanation is that unstable calcium sulfite (CaSOj)
is formed first, and disporportionation to sulfate and sulfide occurs:
CaO + S02 ^± CaS03 ^^ 1/4 CaS + 3/4 CaS04 .
23
-------
This reaction is readily reversible, and the equilibrium SO- increases
with temperature. Opposing the tendency to reject S02, however, is the
reaction in which the sulfide is air oxidized to the sulfate:
CaS + 2 02 •+ CaSO, .
This oxidation reaction is extremely rapid, particularly at low sulfur
contents (but, in the case of limestone, it is greatly impeded by forma-
tion of an impervious sulfate shell). A competition between S0~ rejection
and sulfide oxidation then controls the rate of reaction. At increased
pressures the rate of S02 rejection would decrease, but the rate of S0«
reaction with CaO and the sulfide oxidation would both increase. The
balance point between the competing reactions would shift to higher tem-
peratures, and the temperature maximum would shift to a higher temperature,
3) The Pore Structure Mechanism
The pore structure mechanism postulates that the rate of sulfation
becomes too slow to be useful when the pores in CaO are filled with prod-
uct calcium sulfate (CaSO^). The pore structure is formed during calci-
nation: when calcination is slow, i.e., retarded by a high local partial
pressure of C02, fine pores consolidate into wider pores, thus increasing
the capacity of the sorbent. When calcination is rapid, CO^ is expelled,
reducing the local C02 partial pressure and forming fine pores. The
reaction front moves rapidly away from the fine pores, freezing' in place
the fine pore structure initially formed.
When limestone is calcined in a fluidized-bed combustor at tempera-
tures of around 820°C or lower, the partial pressure of CO. present is
high relative to the equilibrium partial pressure over CaCO_ (because of
coal combustion), and the calcination reaction is slow, permitting con-
solidation of the pore volume among pores with larger radii and eventu-
ally greater sulfate capacity. If the temperature is raised to 900°C,
the retarding action of the local CO- pressure is relatively trivial,
since it is now only a small fraction of the equilibrium partial pressure.
Calcination is rapid and mainly fine pores are created in the solid.
This effect is illustrated in Figure 8 by plotting sulfur removal
24
-------
Curve 718607-A
CO
o>
E
o
«*—
i
E
CNJ
1000
900
800
700
600
500
400
300
200
100
0
' I ' I T I T I I I M ' I ' I ' _
Welbeck Coal
U.K. Limestone, (890-1114) um-
Median Diameter
Ca/ S Mole Ratio, 2. 8
Coal Feed Rate, 137 kg/hr
Velocity, 2. 5-2. 3 m/ s
Bed Height, 0. 67m
799-888°Cll470-1630°R
Atmospheric Pressure
I i I i I i I
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Fraction of Equilibrium CC^ in Effluent, PCO ' PCO equil
Figure 8 - The Effect of CC>2 Pressure during Calcination on S02
Emmlssions from a Fluidized Bed (calculated by
Westinghouse using data from Reference 12,
page A1.95)
25
-------
results obtained by the National Coal Board (NCB), not as a function
of temperature (the controlled variable) but as a function of partial
pressure of C02 (P ) relative to the equilibrium value (PCQ ).
P.,_ /P/^n ..-I is a measure of the therraodynatnic potential for slowing
C0« CO- equil
down calcinationo
At atmospheric pressure calcination will not proceed to completion
at low temperatures. As the temperature is increased, calcination occurs
slowly to produce an optimum calcine. Further increasing the tempera-
ture increases the rate of calcination to the point where an inactive
pore structure is formed and sorbent activity drops. If the system is
operated at pressure, the partial pressure of CCL generated from combus-
tion is greater than that generated at atmospheric pressure. Calcination,
therefore, is slow, producing porous, active calcines at temperatures
greater than those at atmospheric pressure. Active calcines are produced
at those partial pressures of CCL that correspond to a Pc09/pco e il
ratio of 0.6. A combustor operating at 1013 kPa (10 atm) with 10 per-
cent excess air would be operating at a PCQ /PCQ uil of 0.6 at 980°C.
Two other mechanisms that have been postulated to explain the tem-
perature effects, the silica deactivating mechanism1 and the oxidizing/
reducing cycle mechanism, do not account for the fact that the phenomena
are observed in the TGA where these mechanisms are not operative-.
All of the proposed explanations for the temperature effects are
incomplete in that they oversimplify a complex gas/solid reaction. The
S0» and the sulfate/sulfide mechanisms are based on a chemically con-
trolled reaction rate. They do not consider the influence of pore diffu-
sion on the reaction rate. The pore structure mechanism does not explain
why an optimum temperature for sulfation is observed when identical cal-
cines, prepared at the same temperature, are sulfated at varied tempera-
10
tures.
Temperature influences the performance of the sulfur sorbent in two
ways. It directly affects sulfation kinetics and indirectly affects the
sorbent's performance by dictating the speed at which calcination occurs
26
-------
(if at all) and, therefore, determines the pore structure of the calcine.
Westinghouse TG, porosity, and electron microprobe data have shown that
the pore-size distribution developed during calcination has a profound
8 10
influence on the temperature response of different sorbents. ' The
model for sulfation of porous limestone, developed by Hartman and
Coughlin, which considers the chemical rate of sulfation at the individual
grains of CaO, as well as the rate of diffusion of SC>2 through the sor-
bent's pores and through the product sulfate shell, was used to explain
the TG results obtained.
In conclusion, high-temperature operation will not hinder the reactiv-
ity of limestones and dolomites at 1013 kPa (10 atm) pressure. Sorbent
utilization, in fact, improves with increased temperatures up to 1000°C.
The optimum temperature for desulfurization will depend on the sorbent
type and its calcination process or pore structure. Thermogravimetric
data can, experimentally, be used to determine the specific optimum temper-
ature for any specified sorbent and operating conditions. To determine
the optimum temperature theoretically, however, requires a much better
understanding of the mechanism of sulfation than is currently possible.
Excess Air (02 and C02 Concentration) Effects. Fluidized-bed com-
bustors are being designed to operate at excess air levels of 10 to 300
percent. Increasing the amount of excess air used to burn the coal
changes the gas composition in which the sulfur sorbent reacts. The
partial pressure of oxygen is increased, and the partial pressure of CO-
is reduced. Typical concentrations of oxygen and CO^ found in the bed
are tabulated below as a function of excess air level.
% excess air
10
20
100
200
300
%o2
1.9
3.5
10.5
14.0
15.8
% co2
15.7
14.4
8.7
5.8
4.3
27
-------
The effects of oxygen and CO- concentrations on calcium-based sorbent
sulfation were studied independently by varying the gas composition in
TG experiments.
Previous work at Westinghouse has shown no effect of oxygen on the
sulfation of dolomite with gas concentrations of 2 to 11 percent oxygen.
In this work the range of oxygen concentrations tested was expanded to
include all FBC operating conditions.
The effect of oxygen concentration on Tymochtee dolomite sulfation
was studied at 101.3 and 1013 kPa (1 and 10 atm) pressure (Figures 9 and
10). The dolomite was sieved to 16 to 18 mesh and fully calcined on the
TGA at 815°C in 15.2 kPa (0.15 atm) of CO. before the sulfations. The
entire reaction was zero order in oxygen concentration at levels from
0.75 to 16% 0 (0.075-1.6 atm 0. for pressurized sulfation). The vari-
ations observed in the sulfation rates were independent of oxygen con-
centration and, therefore, attributable to sorbent inhomogeneity.
The sulfation rate curves for Tymochtee dolomite indicate greater
than 100 percent sulfation of the sorbent. Since the fraction sulfated
is calculated from the original calcium content of the sorbent, an incon-
sistency between the calcium in the analyzed sorbent and the sample
sulfated might account for the excess sulfation. Sulfation of the mag-
nesium fraction, however, or formation of the double salt found by Hubble
18
et al., - Mg-Ca(SO,),, may be occurring.
The CO. level in the combustor directly determines the form of the
sorbent, oxide or carbonate that will be absorbing SO-. The partial
pressure of C02 in the combustor also has important secondary effects
on sulfur sorption kinetics. In AFBC the calcined sorbent*s pore
structure and, hence, its reactivity depends on the partial pressure
of CO- under which the sorbent is calcined. This phenomenon is a topic
^ 7
of another report and, therefore, was not investigated here. During
PFBC, the C0_ partial pressures generated are often greater than the
equilibrium CO- partial pressure for CaCO_ decomposition. The CaCO-
28
-------
§
r-l
X
4)
U
4
M
00
o
-1
-2
8 l
u
I
•H
e
g
iH
M
V
* -
-2
Tymochtee Dolomite,1000-1900 urn, 101 kPa
815'C, 0.5Z S02, 600 ml/min
Calcined nonisothemially up to 815*C
in 15.2 kPa C0 and N
* Run 335, 2Z 02
- Run 348, 7Z DZ
o Run 334, 10.5Z 0
• Run 336, 14Z 02
Run 333, 16Z 02
T3 \k .5
Fraction sulfated
Figure 9 - The Influence of Oxygen Partial Pressure on the
Rate of Dolomite .Sulfation (101.3 kPa/ 1 atm")
Tymochtee Dolomite, 1000-1190 urn. 1013 kPa
8158C, 0.5Z S02, 1000 ml/min
in 15.2 kPa C02 and
Calcined nonlaothermally up to 815*C
• Run P76, 16Z 02
* Run P79, 10.5Z Oj
x Run P46, 7Z 0,
+ Run P71, 2Z 02
# Run P72, 0.75Z 02
.2
.6 .7 .8 .9
Fraction sulfated
Figure 10 - The Influence of Oxygen Partial Pressure on the
Rate of Dolomite Sulfation (1013 kPa/10 atm)
29
-------
w
H
P
M
SC
SB
O
o
W
H
06
Greer Limestone, 1000-1190 pm (1013 kPa)
Heated at 10°C/min to 815°C in C02> 02 and NZ
Sulfated at 815°C in 0.5% SO,, CO,, 0 and N
Z. £• £• *•
x Run P29, 15% C02 (15% excess air)
* Run P28, 8.7% C02 (100% excess air)
+ Run P30, 5.8% C02 (200% excess air)
• • • Run 76-122, calcined sorbent
-2
\
*
1 1
.1 .2
**
la !4 Is
FRACTION SULFATED
Figure 11 - The Influence of Excess Air Level on the Pressurized
Sulfation of Uncalcined Limestone
Greer limestone, 1000-1410 urn
Fluid bed calcined at 815°C In 151 C(>2 (101 kPa)
Sulfated at 815°C in « 0^ SO^ and N2(1013 kPa)
Rate curve predicted for 0.1Z S02 sulfation from:
* Run F57 (0.1Z S02)
• Run F54 (0.5Z S02)
+ Run P60 (0.3Z SO2)
x Run P58 (0.05Z S02)
§
H
M
2
00
-1
-2
\\
!l !2 !3 J4 15 16 17 IB 19
Fraction Sulfated
Figure 12 - The Influence of Sorbent Residence Time on the
Sulfation of Greer Limestone (1013 kPa/10 atm)
30
-------
fraction of limestone and dolomite, therefore, is often not calcined.
2
Previous work on carbonated sorbents has indicated that
• Most half-calcined dolomites (CaCO.-MgO) are very active
sulfur sorbents.
• Uncalcined limestone is a very poor sulfur sorbent.
Since it may be desirable to use limestone in high-pressure operation,
the possibility of using uncalcined limestones was investigated further.
The sulfation of limestone in the presence of C0_ at pressures
greater than the equilibrium for the reaction CaCO ->• CaO + CO^ is
kinetically limited but thermodynamically favorable. The equilibrium
concentration of SO. is less than 1 ppb at 871°C in 10% C02> 4% oxygen,
at 1013 kPa; yet, TG experiments using 500 ym particles of Limestone 1359
have shown only 3.4 percent sulfation at 850°C in 0.18% S02, 4% oxygen,
and 60% CO. in nitrogen (101 kPa). Limestone 1359, however, is a very
pure calcitic stone, with small, interlocking grains. Diffusion of S02
through its carbonated structure is slow. Since the sulfation of half-
calcined dolomite proceeds readily, an impure limestone, Greer, was
tested for reactivity with SO- in the carbonated state. Although the
magnesium (Mg) content of Greer limestone is low, 0.67 percent, the
stone loses weight when heated in CO. because of the reactions of the
impurities, aluminum (Al) (2.6%), silicon (Si) (7.2%), and iron (Fe)
(1.3%). The structure of the Greer limestone was expected to be open
to SO. diffusion.
Greer limestone particles of 1000 ym sulfated up to 36 percent in
the carbonated state of 1013 kPa. The effect of excess air on the reac-
tion was studied by using varied amounts of oxygen and C0» in the reac-
tant gas. The sorbent was preheated at 10°C/min at 815°C in CO., oxygen,
and nitrogen. The sulfation atmosphere was 0.5% S02 and 15% (2.7% 02
and 15% C02), 100% (10.5% 02 and 8.7% C02), and 200% (14% 02 and 5.8% C02)
excess air. As the excess air level increased (% 0. increased and % CO.
decreased), the utilization of the sorbent improved (see Figure 11).
1 q
Argonne National Laboratories (ANL) have observed increased rates of
31
-------
sulfation of half-calcined Dolomite 1337 when the CO- concentration was
decreased from 100 to 40 percent at 101 kPa, 640 to 800°C. In high concen-
trations of CO , it is likely that retarded diffusion of CO- away from the
£. &
product CaSO, decreases the rate of sulfation in the carbonated material.
Impure limestones, such as Greer, should be useful sulfur sorbents
in low-temperature combustion where the carbonated form of the sorbent
is stable. TG results indicate that the reactivity of the carbonated
sorbent will improve as the amount of excess air used in combustion is
increased.
5 20 21
Sorbent Residence Time. Various investigators '' have reported
that the sulfation of limestone is first order with respect to SO- con-
centration. The application of a first-order reaction model at high
sulfate loadings, however, must be questioned.
The Tennessee Valley Authority (TVA)2* used the rate of sulfation
after one minute of reaction to justify a first-order reaction in SO-.
20
Borgwardt's data were taken at a 10.5 percent conversion level of
Dolomite 1337 at 870°C. This is also early in the reaction, since Dolo-
mite 1337 has been shown to sulfate 100 percent at 815°C. Data from
22
Battelle's dispersed phase reactor have indicated that the apparent
order of sulfation increases with sulfate loading.
Thermogravimetric sulfation has typically been carried out.in
0.5% S0_. Using 0.5 percent SO- provides sulfations that occur in about
2 hours and, therefore, are convenient to study in the laboratory. Sor-
bent residence times in fluidized beds, however, may be 12 hours. In
burning a coal that contains 4 percent sulfur (S), the resulting SO-
level is in the range of 0.09 to 0.34 percent, depending on the amount
of excess air used.
The longer residence times and lower SO concentrations in the fluid
bed could cause sintering of the sorbent, changing its pore structure and,
thus, its reactivity toward S02« The effect of residence time at temper-
ature on the sulfation kinetics of limestones at 101.3 and 1013 kPa (1
and 10 atm) was studied by sulfating sorbents on the TGA in gases of
varied SO- concentrations.
32
-------
The rate of sulfation at 1013 kPa (10 atm) in 0.1 percent SO was
predicted for Greer limestone using TG data from runs with 0.05 to
0.5 percent SO^ in the gas and assuming first-order kinetics (Figure 12).
The predicted curve is fairly consistent, indicating a first-order reac-
tion is followed and the sorbent is unaffected by the time held at tem-
perature (the time required for 37% utilization varied from 15 to 89 min-
utes). The same analysis using Grove limestone, however, indicates
sintering of the stone decreases its reaction rate with temperature
exposure (Figure 13). (The time of exposure when sulfated 8% varied
from 4 to 66 minutes.) Similar results were obtained for Grove and
Greer limestone at 101.3 kPa (atmospheric) pressure (Figures 14 and 15).
I
o
o
4J
1
M
-1
-2
Limestone 1359, 1000-1190 go (1013 kPa)
Calcined Nonisothermally up to 815°C
in 1.5* C02
Sulfated at 815°C in 4Z 02> S02, and N2
Rate curve predicted for 0.1% SO^
sulfation from:
* Run P64 (0.52 SO2>
* Run P65 (0.3Z S02)
X Run P68 (0.1Z S02)
+ Run P69'(0.05X SO.)
Fraction Sulfated
Figure 13 - The Influence of Sorbent Residence Time on the
Sulfation of Limestone 1359 (1013 kPa/10 atm)
33
-------
M
W
H
D
95
o
o
M
H
-
u
o
.J
-2
Greer Limestone, 1000 - 1190 ym
Fluid bed calcined at 815'C in 15% CO- (101 kPa)
Sulfated at 815°C in 4% 02, S02, and NZ (1013 kPa)
Rate curve predicted for 0.1% S02 sulfation from:
* Run 486 (0.1% S02)
+ Run 339 (0.5% S02) -
11 minutes for 30% sulfation
• Run 536 (0.3% S02)
Run 487 (0.05% S02) -
82 minutes for 30% sulfation
'.2 ;3 .4 .5
FRACTION SULFATED
'.6
.7
Figure 14 - The Influence of Sorbent Residence Time on the
Sulfation of Greer Limestone (101.3 kPa/1 atm)
ta
H
JE
Z
§ -1
U
H
at
-2
u
o
-3
Limestone 1359, 1000 - 1190 urn (101 kPa)
Calcined nonisothermally up to 815°C in 15% CO
Sulfated at 815°C in 4% 02> S02, and NZ
Rate curve predicted for 0.1% SOj sulfation from:
• Run 524 (0.5% S02) - 3 minutes
for 10% sulfation
* Run 549 (0.1% S02> - 28 minutes
sulfation
.1
FRACTION SULFATED
.2
Figure 15 - The Influence of Sorbent Residence Time on the
Sulfation of Limestone 1359 (101.3 kPa/1 atm)
34
-------
We conclude that the residence time of a sorbent at temperature may
change the sulfation kinetics by varying the sorbents1 pore structure,
from the first-order relationship observed at initial reaction periods.
Depending on the type of sorbent, this effect could lead to errors in
rate constants projected from TG data. Unfortunately, the types of sor-
bent that will show varied kinetics with residence time have not yet
been identified. It is possible that the higher sodium content in the
Greer Limestone, 0.2 percent sodium (Na), prevented shrinkage of the
calcine's structure during the high-temperature exposure.
Pressure. Early Westinghouse studies at 1013 kPa (10 atm) pres-
sure found that the sulfation rates at such high pressures were not
significantly greater than those at atmospheric pressure. The variation
in the sulfation rates of Greer Limestone and Tymochtee dolomite with
pressure, when calcined under corresponding conditions, that was found
during this test series is illustrated in Figures 16 and 17. The rate
of sulfation increases with pressure; the increase, however, is not of
the magnitude predicted by a first-order reaction in S02.
The ratio of the rate of sulfation at 1013 kPa (10 atm) to the rate
at 101.3 kPa (1 atm) is shown in Figure 18 for Tymochtee dolomite sulfa-
tion. The dolomite was sieved to 16 to 18 mesh and calcined at 815°C in
0.15 atmosphere of C0_ before sulfation. Increasing the pressure from
101.3 to 1013 kPa (1 to 10 atm) has increased the reaction rate .by a fac-
tor of 2 to 3 over most of the sulfation. This relation can be rational-
ized by considering the rate of pore diffusion in a shrinking core model,
*
da m 1
dt 29[(1 - a)'1/3 - 1]
where
G a r2
6b Dp C
The gas concentration, C, is proportional to P, and the pore diffusion
coefficient, D , is the sum of two terms, one representing Knudsen
*See Nomenclature.
35
-------
C
•rl
X
X
01
•u
to
00
o
-1
-2
Greer Limestone, 1000 - 1190 urn
Fluid Bed Calcined at 815°C in 15%
(101.3 kPa)
Sulfated at 815°C in 0.5%
Run P59, 1013 kPa
* Run 339, 101.3 kPa
-4-
•4-
•4-
•4-
•4-
•4-
•4-
.1 .2 .3 .4 .5 .6 .7 .8 .9
Fraction Sulfated
4%
Figure 16 - Comparison of Sulfation Rates at 101.3 and 1013 kPa
(1 and 10 atm) Pressure (Greer Limestone)
3
I
C
O
o
X
-------
Curve 693690-A
Ji
"no
C 50
S
s 40 r-
Oi
30 -
Tymochtee Dolomite, 1000-1190 pm
815° C. 0.5%S02,02, andN2
Calcined at 815°C in
15.2kPaC02
% 02 in Sulfating Atmosphere:
2%
_E
(O
§ 20
£
1 10
"no
re
O£
)l
/
>^'
****
»j*i^f+^^^~^^
1 1 1 1 I 1
1 i
"\\
V
\v
1 1 T1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 LO
Fraction Sulfated
Figure 18 - The Effect of Pressure on the Sulfation Rate
of Tymochtee Dolomite
diffusion in small pores where collisions between gas molecules and the
pore walls are more frequent than collisons between gas molecules:
D -
P T
3 /_jm_\ + I
"4? 2R0A D
\ A/
-1
Now the ratio of the rate at pressures to that at atmospheric, R , is
a fda\
L) "P °so2 <" p>
"9
-------
If no small pores are present, Knudsen diffusion is unimportant and
R
a D(P = 1) A '
Since D a ^, the rate is not affected by pressurization. If only Knudsen
diffusion were important, however, the rate would be proportional to the
pressure. In reality, limestone particles have a broad distribution of
pore sizes, so D is a function of the fractional conversion, and this
function could vary with pressure. The ratio of the diffusion controlled
rates, therefore, could exceed 10. As the stones reach 100 percent con-
version, the rates as well as the ratio approach zero.
The reaction, however, does not proceed by a simple shrinking core
mechanism, and the combined effects of pore diffusion with chemical reac-
tion must be considered to accurately project the effect of pressure on
the sulfation reaction.
Sulfation of Large-Grained Dolomites. Two dolomites, Canaan from
Connecticut and Kaiser from California, were chosen to study the sulfa-
tion of large-grained sorbents. We observed poor TG sulfation of 420 to
500 pm particles of Canaan dolomite in run P5 at 1013 kPa (10 atm) pres-
sure, 843°C (see Figure 19). Electron microprobe scans of the sulfated
product from Run P5 to show sulfur is concentrated almost solely at the
grain boundaries of the sorbent (Figure 20). The low utilization is
consistent with the finding by Combustion Power Company that sulfur pen-
etration into large particles of massive-grained dolomites is not very
deep and that fine particles are needed to achieve low calcium-to-sulfur
(Ca/S) feed ratios and good desulfurization.
To determine if the large-grained dolomite could be utilized in a
fine particle form, 74 to 149 urn particles of the sorbent were tested at
815°C at atmospheric pressure. The fine particles were precalcined to
activate the sorbent. Greater than 50 percent utulization was achieved
(see Figure 21). On a weight basis, however, the fine, precalcined
sorbents were still less useful than 1000 urn particles of other
38
-------
* *
Q
U
fe
»J
£3
V)
o
M
H
,* *
Run # P5 (1013 kPa)
Canaan Dolomite, 420 - 500 pro
Calcined Noniso thermally up to
84 3° C in 4.3% C02 and 15.8% 0,
Sulfated at 843°C in 0.5% SO
4.3%
15.
"to !o fc to toJoto to to loo
TIME/MINUTES
Figure 19 - TG Sulfation of Canaan Dolomite
(420 - 500 ym particles)
smaller-grained, precalcined sorbents (Table 3). Although this extent
of sulfation would require a Ca/S molar feed ratio of only 1.6 for
80 percent sulfur removal with Canaan dolomite, other limestones, or
Dolomite 1337, would require a smaller feed on a weight basis, and the
additional pulverizing would be unnecessary.
Prediction of Desulfurization Performance
Analysis of Pilot Plant Fluidized-Bed Data
We derived the projections of Ca/S molar feed ratio required to
achieve any selected degree of desulfurization in a fluidized-bed com-
bustor by using a simplified model for fluidized-bed desulfurization,
with kinetic rate constants developed using laboratory TG data. For
confirmation, where possible, we have compared the TG-supported model
with available data from fluidized-bed combustors. The TG data base,
39
-------
Photomicrograph of Scanned Area
Sulfur Scan
Figure 20 -• Sulfur Penetration at Periphery and Grain Boundaries
of 500 um Particle of Canaan Dolomite (^300 ym grains)
40
RM-72659
-------
.6 *
5 |
* * *
o .4
H
O
M
H
Run // 317
Canaan Dolomite, 74 - 149 pm
Calcined Nonisothermally up to
900°C in 602 C(>2
Sulfated at 815°C in 0.5Z SO ,
4% 0, Z
to"
1o
TIME/MINUTES
Figure 21 - TG Sulfation of Canaan Dolomite
(74-149 pm particles)
Table 3
RELATIVE EFFECTIVENESS OF LARGE-GRAINED
DOLOMITES IN SULFATION
Sorbenta
Dolomite 1337
Greer Limestone
Lowellville Limestone
Limestone 1359
Canaan Dolomite
Kaiser Dolomite
Grain Size,
ym
27
6
40
6
300
>400
S03 Pickup ,b
mg S03/mg Sorbent
0.36
0.31
0.29
0.28
0.22
0.25
Particle Size,
V>m
1000-1190
1000-1190
1000-1190
1000-1190
74-149
74-149
a
Sorbents were precalcined at 900°C in 60% C0,/N0.
b £• £
SOa pickup during 815°C sulfation (0.5% SO,, 4% 0?) while rate is
^0.« Ca/min.
41
-------
the model, and its agreement with fluidized-bed data are discussed in
Appendix A. Areas for improvement of the projection technique are also
identified.
Westinghouse Batch Fluidized-Bed Data
Batch fluidized-bed experiments were performed under an Electric
Power Research Institute (EPRI) contract to compare the attrition
rates of six sorbents. During these runs the limestones were sulfated,
and the concentration of SO- in the effluent line was monitored as a
function of time. The data collected (Appendix C) were analyzed here to
determine how well rate constants derived from TG data compare to those
found in the fluidized-bed unit. TG data from a seventh sorbent, Penrith
limestone, were used to compare the rate constants from TG data to batch
fluidized-bed data obtained by Cambridge University.23
Fluid Bed Experimental Facility. Batch fluidized-bed experiments
were performed in a 3.5-cm-diameter Inconel 600 batch reactor (Figure 22).
The reactor sits in a shell that may be pressurized to 1013 kPa (10 atm).
For these experiments the shell was open to the atmosphere. The four
electric heating coils that surround the furnace were controlled by a
Trendtrak programmer. Bed temperature was recorded by a thermocouple
extending through the flanged reactor lid and down into the solids
region.
The reactant gas, metered through rotameters, was preheated as it
passed between the reactor and the furnace before entering the base of
the bed. The distributor plate was made of an Inconel 600 disk through
which 37 holes 0.012 cm in diameter were drilled. Pressure gauges were
located on the manifold and effluent lines. The effluent gas passed
through a continuously monitoring Dynasciences S0~ meter.
The minimum fluidization velocity for the raw stone, 1000 to 1410 ym,
was about 67 cm/s. Two run procedures were used. Some sorbents were
sulfated immediately after calcination (Procedure 1). Other sorbents
were cooled to room temperature after calcination (Procedure 2). A
42
-------
0.25 in.Inconel Tube
Weld
Owo. f386A07
Inconel 600
0.25 in.Plate
Inconel 600
Reactor Shell
Distribution
"late
Lid Assembly
To Filter
Thermocouple Well
Weld
Figure 22 - Schematic of Batch Fluidized-Bed Reactor
-------
sample of the fluidized-bed calcine was sulfated on the TG. The sorbent
was later reheated and sulfated. The two procedures are outlined as
follows :
• Procedure 1: Raw limestone, lOOg, was heated at 10°C/min
up to 815 °C. The fluidizing gas was switched
on (15% CCL in nitrogen) . After four hours
in the calcining atmosphere, the sulfating
atmosphere was introduced (0.5 percent S0_,
4 percent oxygen in nitrogen). When the moni-
tor indicated a 20 or 80 percent breakthrough
of S02 (0.1 or 0.4 volume percent S02 in the
effluent) , the fluidizing gas was turned off
and the sample was cooled.
• Procedure 2; After four hours in the calcining atmosphere,
the sorbent was cooled to room temperature
and stored in a desiccator. A sample of cal-
cine was TG sulfated. Before being sulfated
in the fluidized bed, as outlined in Proce-
dure 1, the sorbent was reheated at 108C/min
up to 815 °C.
The operating conditions for the batch fluidized-bed runs are summarized
in Table 4.
Estimation of Reaction Kinetics. The reaction rate of the sorbents
can be defined in terms of an effective surface rate constant, K , based
on the external surface area of the particles used,
dt
where
N ™ moles of solid reactant/particle
C = S0« concentration in emulsion gas, mole/cc
r = particle radius, cm.
44
-------
Table 4
BATCH FLUIDIZED-BED EXPERIMENTS
Run
Limestone
Fluidizing
Velocity,
cm/s
Bed Height, cm3
Static
Fluidized
Procedure
Number
Surface Mean Particle Radius
of Bed Material, cm
After Calcination
After Sulfation
Oi
W )3.5 cm Reactor
NU5 1359 95
NU6 Carbon 95
NU7 Brownwood 95
NU8 Ames 95
L6 Bellefonte 100
Lll Mississippi 100
Cambridge 7.8 cm
Reactor
Cambridge:
Penrith 43.6
7.6
4.8 6.7
5.4
2
2
2
1
2
2
0.056
0.053
0.047
0.051
0.043
0.055
0.052
0.051
0.046
0.056
9.5
As measured in open bed, prior to experiment.
-------
The batch fluidized-bed data were analyzed with four models, as out-
lined in Table 5. All of the models assume perfect mixing of the solid
phase. The particle surface area was calculated from the surface mean
particle size of the bed material (Table 4); the particles are assumed
to be spheres. The fraction of the bed calcium sulfated, a, was calcu-
lated from the effluent S0_ concentrations, f = fraction of S0_ in
effluent line, by
t
F f (f - f)dt
o/ o
0
. ,
MCA
where M . is the number of moles of calcium in the bed and F_ is the
(jA
total gas molar flow rate, mole/min.
The model that corresponded most closely to data from TG experiments
(model 1) assumed plug flow of the gas phase, with sulfur generation
occurring at the bed's base.
Ks=fln(T '
where
F « total gas flow rate, cc/s
S ™ surface area of particles in the bed.
If we assume spherical particles in the bed,
pr
(p = density of the limestone, g/cc; w, = bed weight, g) .
The effective surface rate constant was calculated from TG data as
follows:
46
-------
Table 5
SUMMARY OF MODELS USED TO ANALYZE FLUIDIZED-BED DATA
Model
KD from Fluid Bed Data
Comments
• Plug flow with S generation
at base of bed
• Quasi-steady-state
• Plug flow with uniform
S generation
• Quasi-steady-state
• Perfect gas mixing
• Quasi-steady—state
• Perfect gas mixing
• Transient
Nomenclature:
K,
K = F/S In C /C
s o
F ~SKs
c/co ' si" (1 - e -T
,
s " St 1+6 ln C
Best agreement with TG data
Ks in fluid-bed initially much
higher than from TG data
Ks in fluid-bed initially much
higher than from TG data
Valid only where t is on the
order of gas residence time
.s - surface rate constant, cm/s
F = superficial volumetric gas flow rate, cc/s
S = surface area of particles in bed, cm2
= S02 concentration in effluent line
= S02 concentration fed to bed
static bed height, cm
volume fraction of bubble phase
void fraction of emulsion phase
= time, s
= cross-sectional area of bed,
-------
or
_ da pr
Ks ~ dt 3C '
Modeling Results. Figures 23 and 24 compare the rate constant
derived from the four fluidized-bed models. The surface rate constants
calculated from the transient model are much smaller than the others.
The transient model would be expected to apply at reaction times on the
order of the gas residence time, which was only 0.05 to 0.08 s for the
fluidized-bed runs. Sulfur dioxide evolution data were not measurable
until one or two minutes of reaction. Steady-state models, therefore,
would apply. At low values of the ratio of the SO- concentration leaving
the bed to that fed, C/C , the second and third models gave rate constants
that were much greater than those predicted from model 1 and the rate
constant measured in TG data. At higher levels of S0_ emission, C/CO,
however, the rate constants predicted using models 1 through 3 were very
similar. The similarity is illustrated by comparing the ratio of the
rate constant predicted from model 1 (plug flow of gas with below-bed
sulfur generation) to the rate constant predicted from model 3 (perfect
gas mixing) as a function of C/CQ:
C/C K (Model 3)/K (Model 1)
o s s
0.1 3.9
0.2 2.5
0.5 1.4
0.8 1.3
At low C/CO, where a large fraction of the sulfur fed is absorbed by the
bed, perfect gas mixing is not likely. Since sulfur is fed at the bed's
base, uniform generation of sulfur is an unlikely model when most of the
sulfur is being absorbed by the sorbent (i.e., at low C/CO). As might
be expected, rate constants derived from model 1 (plug flow of gas with
48
-------
at
u:
X +
+
x +
Beliefonte Limestone, ^ 1000 pm
Calcined at 815°C in 15% C02
Sulfated at 815°C in 0.5% S02> 4%
• Model 1
+ Model 2
x Model 3
* Model 4
C/C maximum =0.40
o
+.
****** ********
.1
FRACTION SULFATED
Figure 23 - Comparison of Fluidized-Bed Models (Beliefonte limestone)
09
B
u
en
10
9
8
7
6
5
4
3
2
1
0
Carbon Limestone, ^ 1000 pm
* Calcined at 815CC in 15% C02
\ Sulfated at 815°C in 0.5% S02> 4%
*
\ • Model 1
* x Model 2
x * Model 3 .
A C/C maximum - 0.10
\
• TL
»* «• K
' '"•V'to-'sw
1 2
FRACTION SULFATED
Figure 24 - Comparison of Fluidized-Bed Models (Carbon limestone)
49
-------
sulfur generation at the bed's base) corresponded most closely to the
rate constants measured on the TGA. At high C/C , however, there was
little difference in the quasi-steady-state models.
The comparison between rate constants derived from fluidized-bed
data using model 1 and rate constants measured on the TGA is illustrated
in Figures 25 through 30. The agreement is excellent for Carbon, Grove,
and Brownwood limestones. The worst agreement was found for Ames lime-
stone. During previous work, ^ however, this stone was found to be
inhomogeneous. In addition, the calcine used for the TG sulfation was
prepared in the TGA. If any large deviation in calcination conditions
(i.e., temperature, C0_ partial pressure) existed between the fluidized-
bed and the TGA, the sorbent's pore structure and, therefore, its sulfa-
tion rate would be varied. The rate constants of Mississippi and
Beliefonte limestone, derived from fluidized-bed data, differed from
those measured on the TGA by, at most, a factor of three. Sorbent
utilization at any particular rate constant does not vary greatly since
the sorbents were not highly utilized.
Some shortcomings of the fluidized-bed apparatus should be noted.
The bed heights were very low, and, therefore, the fluidization produced
may not be representative of larger units. The drilled distributor
plate sometimes became partially plugged during operation. Dead regions
in the bed near the plate, therefore, may have restricted the active
section of the bed.
The comparison of TG rate constants for Penrith limestone to those
obtained from a 28-cm batch fluidized-bed unit at Cambridge University
using a bubbling bed model is shown in Figure 31. Rate constants
obtained in the batch fluidized-bed were greater at low sulfate conver-
sions than at those derived from TG data.
In conclusion, rate constants derived from batch fluidized-bed data
are fairly insensitive to the bed model used to interpret the data, pro-
vided C/C is >0.2. Quasi-steady-state models tested include plug flow
50
-------
4%
e
o
Limestone 1359, ^ 1000 pm
Calcined at 815°C in 15% CO,
A
Sulfated at 815°C in 0.5% SC^
• Fluid-bed data, NU5
* TG data, Run 366 (Using
Calcine from NU5)
FRACTION SULFATED
Figure 25 - Comparison of Rate Constants Derived from Fluidlzed-Bed
Data (Model 1) and TG Data: Limestone 1359
6 4
5 3
Carbon Limestone, ^ 1000 um
Calcined at 815°C in 15Z C0£
Sulfated at 815°C in 0.5% SO
*
\-
.
<
2, ™ ~2
• Fluid bed data, NU6
* TG data, Run 368 (using
calcine from NU6)
.2
.3
FRACTION SULFATED
Figure 26 - Comparison of Rate Constants Derived from Fluidized-Bed
Data (Model 1) and TG Data: Carbon Limestone
51
-------
3.5 *
2.5
^ 2
B
u
» 1.5
.5
* *
**\
Brownwood Limestone, ^ 1000 ym
Calcined at 815°C in 15% C02
Sulfated at 815°C in 0.5% S02> 4% 02
• Fluid bed data, NU7
* TG data, Run 358
*
* *
**
***
* *
*****
.2
.3
FRACTION SULFATED
Figure 27 - Comparison of Rate Constants Derived from Fluidized-Bed
Data (Model 1) and TG Data: Brownwood Limestone
B
o
0
*
*
Ames Limestone, ^ 1000 urn
Calcined at 851°C in 15% C02
Sulfated at 815°C in 0.5% S02, 4% 0
• Fluid-bed data, NU8
* TG data, Run 354
.1
.2
.3
.4
FRACTION SULFATED
Figure 28 - Comparison of Rate Constants Derived from Fluidized-Bed
Data (Model 1) and TG Data: Ames Limestone
52
-------
«
B
**
Beliefante Limestone, ^1000 ym
Calcined at 815°C in 15% C02
Sulfated at 815°C in 0.52 S02l 4% 02
• Fluid-bed data, L6
* TG data, Run 420 (using
calcine from L6)
.1 .2
FRACTION SULFATED
.3
Figure 29 - Comparison of Rate Constants Derived from Fluidized-Bed
Data (Model 1) and TG Data: Beliefonte Limestone
-J 3
0
U
Mississippi Limestone, ^ 1000 urn
Calcined at 815°C in 15% C0
Sulfated at 815°C in 0.5%
4%
• Fluid-bed data, Lll
TG data, Run 435 (using
calcine from Lll)
FRACTION SULFATED
Figure 30 - Comparison of Rate Constants Derived from Fluidized-Bed
Data (Model 1) and TG Data: Mississippi Limestone
53
-------
Curve 718700-A
12
Cambridge Fluid-Bed Data
Westinghouse TG Data, Run 713
Penrith Limestone, 710-840 Mm
Sulfated at 840°C in 0. 25% SO,, 2% (X,
0.2
Fraction Sulfated
0.3
0.4
Figure 31 - Comparison of Rate Constants Derived from
Cambridge Fluidized-Bed Data (Bubbling
Bed Model) and Westinghouse TG Data
(with uniform sulfur generation and with below-bed sulfur generation)
and perfect gas mixing. The same conclusion was reached at Cambridge
where a bubbling bed model was also tested. Since modeling assumptions
must be made to analyze batch fluidized-bed data, the rate constants
derived are subject to more uncertainty than those measured directly on
the TGA. Care must be taken, however, to assure good gas/solid con-
tacting on the TGA. A mass transfer correlation for the initial rates
measured on the TGA would also be useful. Uncertainties in measurements
and modeling techniques during initial reaction in the fluidized-bed
make this technique unsuitable for deriving initial reaction rates.
Conclusions
FBC Operating Range Impact on Desulfurization Performance
High-temperature (900-1000°C) operation is recommended for pres-
surized fluidized-bed combustion. At higher temperatures limestones may
54
-------
be used in the calcined form. Thermogravimetric tests indicate that sor-
bent (limestone or dolomite) utilization in PFBC will not decrease at
temperatures of from 900 to 1000°C.
Dolomite sulfation is zero order in oxygen concentration throughout
the range of typical FBC operation (0.75-16% 0») at pressures of
101.3 kPa (1 atm) and 1013 kPa (10 atm).
Impure limestones, such as Greer, are useful sorbents in the uncal-
cined form, provided that the CO- partial pressure is not much greater
than the equilibrium for calcination.
The residence time of a sorbent at high temperatures may alter its
sulfation kinetics by varying the sorbent1s pore structure. The effect
of sorbent residence time on its performance should be tested when TG
data are used to project calcium feed requirements.
Large-grained dolomites are not worth considering as sulfur sorbents
for FBC. Even when pulverized ('vLOO ym) and precalcined they are not as
active as 1000 ym particles of other sorbents.
Prediction of Desulfurization Performance
Thermogravimetric rate data can be successfully used to determine
the rate constant of sulfation as a function of sorbent utilization for
calcium-based sorbents. By judiciously selecting operating conditions
that represent conditions in fluidized-bed combustion, the rate constant
can be used to predict sulfur retention in fluidized-bed units. The
agreement between fluidized-bed data and TG projections has been demon-
strated using data collected at 1013 kPa (10 atm) pressure, as well as
at atmospheric pressure.
The TG projections are limited by the availability of complete pilot
plant data (particle size distribution in the bed, fraction of inert par-
ticles in the bed, bed expansion data), the accuracy of pilot plant data
(including fluctuations in coal and sorbent properties and nonsteady-state
operation), the representability of the 20 mg sample used in the TGA of
the bulk limestone, as well as the basic assumptions applied in the
projections.
55
-------
Data obtained from batch fluidized-bed experiments may also be used,.
to determine rate constants of sulfation as a function of sorbent utili-
zation. The rate constants, however, are dependent on the model assumed
for the fluidized bed. Rate constants derived by assuming perfect mixing
of the solids and plug flow of the gas phase agree well with rate con-
stants measured on the TGA. At low levels of sulfur retention, the type
of model assumed has a minimal effect on the derived rate constants from
quasi-steady-state models.
PARTICLE ATTRITION
Perspective on Need
Fluidized-bed combustors operate burning coal in a fluidized bed of
granular, noncombustible particles. These bed particles may consist of
limestone or dolomite, which acts as a sorbent for the sulfur released
during coal combustion. The performance of an FBC system will depend
upon the attrition resistance of these bed material particles.
An understanding of particle attrition in fluidized-bed combustion
processing systems is important for:
• Selecting sulfur sorbents that have the desired attrition.
characteristics, so that operability of the FBC system can
be maintained. (For example, sulfur sorbents resistant to
attrition will be required for FBC processes in which the
fj *
sorbent is regenerated for re-use ' and are generally
desired for processes in which the sorbent is used on a
o
once-through basis. )
• Permitting the prediction of sorbent particle size history
or, in turn, predicting sulfur removal for a given sorbent
and process design
• Permitting the prediction of particle size history for esti-
mating flue gas particulate loading and size distribution
from the fluidized-bed combustor and other process components
(e.g., carbon burnup cell, sorbent regenerator, sorbent pre-
treater, spent sorbent processing unit). Information on flue
56
-------
gas particle loading and size distribution is necessary in
order to assess environmental impact and particle control
technology requirements.
• Aiding in the interpretation of other process phenomena (e.g.,
time element profiles, char or low-grade fuel combustion).
The objectives for the particle attrition work are to develop predic-
tive models that describe sorbent (and fuel) attrition in fluidized-bed
combustion systems. The models will relate attrition to particle proper-
ties, FBC system design parameters, and FBC system operating conditions.
The models provide a basis for incorporating particle attrition charac-
teristics into the assessment of sulfur sorbent selection, sulfur removal
system evaluation, and particulate profile modeling through fluidized-bed
combustion systems. These assessments will be utilized to achieve optimal
plant design for given environmental emission requirements.
Approach
The approach selected to develop an understanding of particle attri-
tion is to:
• Identify sources of attrition in fluidized-bed combustion
systems.
• Select specific attrition mechanisms for study.
• Assimilate available data and propose a model.
• Carry out an experimental program and analysis to develop
a predictive model for the respective mechanisms.
• Formulate a unified attrition model.
• Integrate the attrition model into the sulfur removal sys-
tem and particulate control system models (inclusion in
other system models to be performed as needed - e.g., trace
element profiles).
The work reported includes three aspects:
• Screening tests to gain perspective on the effect of operat-
ing conditions (including interactions) and stone type on
attrition, and to develop experimental techniques. These
57
-------
tests include tests on the effect of stone type, atmosphere,
temperature, pressure, particle size, particle composition
(sulfation, calcination), rate of heating, and time on
attrition; and tests on sampling, particle size measurements,
and the effect of sieving attrition.
• Development of an attrition model for one attrition mechanism
(abrasion or "bubbling-bed" attrition), and experimental
confirmation of the model.
• Identification of future work to be carried out.
The study of attrition in fluidized beds is fragmented and incom-
plete. In other areas, such as ball milling and jet milling, there are
unified theories of attrition. Fluidized-bed attrition is, at best,
described only in part in a limited number of references. Nowhere is
there a comprehensive model incorporating all of the attrition sources
in a fluidized-bed system, nor reference to the mechanism of each source.
The wearing down of particles, called attrition, is variable and not
well understood. If attrition rates are related to various properties of
the particulate solids and fluidization gas and operating conditions, we
should be able to develop an expression describing the rate of attrition
in any given system. Several researchers have studied the effects of
single variables under various conditions, but no general prediction
equations have been formulated. This study is a beginning in defining
a complete fluidized-bed model. The principal sources of attrition are
Identified and examined.
The specific objectives of this study were to:
• Identify the various causes or sources of attrition appli*-
cable to attrition in fluidized-bed combustion systems
• Develop expressions relating attrition rate to design and
operating conditions, and to sorbent properties
• Test the proposed attrition formulas in controlled labora-
tory experiments.
58
-------
Sources of Attrition
The frequently considered source of attrition in a fluidized bed
is the obvious grinding and shattering collisions of particles. There
are several causes of particle wear, which include the following.
Abrasion
In this process, defects, edges and corners are knocked from
particles by low-energy collisions. Abrasion can occur during passage
of a gas bubble through the bed of solids.
High-Energy Collisions
Particles may be accelerated to high velocity - for example, when
entrained in a jet at the grid. The high-velocity particle can strike
another particle or vessel wall and shatter into relatively large
fragments.
O /
Blinichev, Strel'tsov and Lebedeva have distinguished two zones
in a fluidized bed - the lower, which they call the nozzle effect zone,
in which gas jets accelerate large particles to energies sufficient for
shattering; and the upper zone, characterized by intensive mixing and
low-energy impacts that abrade particle surfaces.
Thermal Shock
When cold sorbent particles are added suddenly to a bed of red-hot
solids, there is severe thermal stress on the cold particles. One
expects spalling at the particle surface and perhaps shattering of the
25
entire particle into large fragments.
Chemical Stress
Sorbent particles calcine upon injection into the bed (CaCO_ -»• CaO
+ C02) and then may react with S02 to form CaSO,. These reactions cause
subsequent changes in the lattice structure. This change in the structure
of a particle at its surface hardens particles in some cases, or in
other cases causes internal stresses leading to spalling or weakened
26 27
particle surfaces. '
59
-------
Internal Gas Pressure
When cold limestone or dolomite makeup sorbent is added to a hot
fluidized bed, the resulting calcination generates CO- within the
particle. The internal gas pressure may cause the particle to fracture.
Esso Research Centre in Abingdon, UK, found that a lower calcination
28
rate of fresh limestone results in lower production of fines.
Similarly, water within particle cracks will flash when heated to bed
temperatures. While CO™ pressures are moderate (100.0 kPa equilibrium
at 900°C), steam pressures are high and might explode particles.
Transfer Lines, Rotary Valves, and Cyclones
This is important auxiliary equipment in fluidized-bed combustion
systems. Sorbent breakage rate is related to the circulation rate of
solids and is controlled by equipment design effects on solids impact.
Screening Tests
Purpose
In selecting sorbents and in designing equipment there is a need
for understanding the fundamental characteristics of attrition. Com-
parison of two or more candidate sorbents requires applying the attri-
tion forces that will act in hot fluidization. The purpose of the
current attrition test program is to:
• Identify areas of concern in screening sorbents for
fluidized-bed combustion
• Measure the relative effects of the several sources of
attrition to provide a basis for developing screening
techniques
• Compare the attrition tendencies of candidate sorbents for
fluidized-bed combustion
• Develop techniques for sorbent screening.
Scope
In this test program we have investigated a number of variables over
their expected ranges in fluidized-bed combustion. Testing has included
60
-------
several fluidized-bed systems available at the Westinghouse R&D Center.
The scope of testing and equipment used are summarized in Table 6.
Terminology
Attrition Rate is an obscure term, defined in different ways by
various researchers. Attrition rate is rigorously defined by both a
description of the rate of breakage of each size particle and a descrip-
tion of the fractions of fragment sizes produced by breaking each size
particle. Calculation of these rate and breakage functions is notably
difficult, and extensive data collection is required. In this study we
have defined attrition rate, R, as the rate of disappearance of particles
larger than a stated size. The mass of large particles (coarses) in a
fluidized bed, M, and the attrition rate R are related by
R = ~ M dT '
The extent of particle attrition A is defined as
'"I HM *
dM o
M n M
o M
o
in which M and M, are the masses of coarses before and after an interval
o 1
of attrition.
Attrition in this work is taken to be the process by which particu-
late solid is reduced in size by breaking.
Extent of Attrition is defined as the fraction of sorbent mass
larger than a stated size that is reduced by attrition to fragments
smaller than the stated size.
Effect of Sorbent Type and Atmosphere on Attrition
The purpose of this testing on the 3.5-cm test system was to determine
the dependence of extent of attrition upon various combinations of sorbent
types and reactor atmospheres (gas compositions) at 101.3 kPa (1 atm)
pressure.
61
/ Bdt-/
-------
Table 6 Dwg. 1701B65
RANGE OF VALUES OF TEST VARIABLES IN SEVERAL TEST SYSTEMS
Fluidized Bed
Test System
3.5-cm-diam
Attrition Test
System
10-cm-diam
Attrition Test
System,
Sintered Grid
7 -cm -diam
Plexiglas
Attrition Test
System
Variable Tested
Description
Sorbent Type
Atmosphere
Duration of
fluidization.
Time held at 815°C
Heating Rate
Decree of calcination
Thermal shock
Character of wear
in fluidized-bed
attrition
Particle swelling
Bed pressure
Temperature
Particle diameter
Degree of sulfation
Grid jet
Velocity IU-U .)
nil
Variation of attrition
rate with time
Size of attrited
fragments
Change in limestone
sorbent shape during
fluidization
Range
3 sorbents
3 Gas
compositions
0 to 9 hour
OtolO
riO°C/minand
Uoo°C/min
10% and 100%
10°C/minand
400°C/min
( Photomicro -
graph analysis)
100- 1000 kPa
650 and 815 °C
-710*500 and
-1410 + 1000 urn
0 and 10%
Present and
absent
12. 5 and 25°
0.25 to 647 h
Test Date
Nov. 1976
Jan. 1977
Oct. 1976
Feb. 1977
Feb. 1977;
Mar. 1977;
Apr. 1977
Apr. 1977
Dec. 1977
June 1977
Appendix
Appendix
Dec. 1977
Oct. 1977
62
-------
Our apparatus for these tests was the 3.5-cm-id attrition test cell
system described in Appendix E.
Stones selected are those that have performed well in earlier work.
They are Tymochtee dolomite; Greer, an impure limestone containing
11.3 percent silicon, aluminum, and iron; and Grove, a pure limestone
that is 98 percent calcite. The atmospheres investigated are the two
calcining compositions and a sulfating atmosphere:
Atmosphere
103 kPa, 815°C
Symbol
Cl
C2
ST
Type
Calcining
(low C02)
Calcining
(high C02)
Sulfating
Composition, %
N2
85
50
96.8
co2
15
50
0
so2
0
0
0.2
°2
0
0
3
The procedure was as follows :
1. Grind stone, separate it into sieve fractions, and measure the
size distribution of the 1000 to 1400 urn (12 to 16 mesh) frac-
tion with sieves.
2. Fluidize the 1000-1400-um fraction of the stone in the attrition
test cell with cold nitrogen and note the gas velocity, U .., at
mi
incipient fluidization. From this calculate the stone mass mean
29
particle diameter as defined by Wen and Yu's equation
Re - (33. 72 + 0.0408 Ar)1/2 - 33.7
o
where :
Re
o
Ar
p(T)dp3(pp-p(T))gp~2(T).
3.
Knowing only T and Umf , we can calculate d.
Gradually heat (10°C/min) 70 g of stone in the attrition test cell
to 815°C. Gradual heating avoids attrition from thermal stress.
63
-------
4. Initiate gas flow at 1.2 times the minimum fluidization velocity*
calculated for the stone diameter as calculated in step 2 above,
gas species, and temperature. Maintain gas flow for four hours,
then turn off the gas flow and allow the bed to cool slowly to
room temperature.
5. Weigh the solids removed from the bed. Measure the size distri-
bution of solids by sieve analysis. Weigh the solids on the
exhaust filter. Assay samples of original and reacted stone for
the fraction C02 (LOI) and the fraction of calcium for mass
balance calculations.
Attrition is defined in different ways by different investigators.
Various terms, including decrepitation, elutriation, and attrition, are
used to define the breaking of fluidized-bed particles. In this test
the extent of attrition denotes the rate at which particles smaller than
710 ym are formed. We chose 710 ym as the larger size limit of stone
charged to the system. On a fully calcined solids basis, the extent of
attrition is defined as
extent of attrition = (final bed fines - initial bed fines +
final filter fines) * input mass =
(final bed\ /frac < \L \ /initial Wfrac <\A \ / final Wfrac
-------
The results of these attrition measurements are listed in Table 7,
Table 7
PERCENT OF SOLIDS ATTRITED IN FOUR HOURS
Stone Type
Tymochtee
Grove
Greer
Cl, Calcining
1.842
1.813
0.600
0.554
0.711
0.777
0.846
2.121
3.377
Atmosphere
C2, Calcining
1.003
0.783
0.684
4.405
1.989
1.039
8.326
7.190
3.390
ST, Sulfating
0.4680
0.4500
*
0.5460
0.4360
*
1.9450
1.0160
*
*Denotes no observation.
Table 8 summarizes statistical calculations on these data.
Table 8
SUMMARY OF ATTRITION TEST DATA STATISTICS
Source of Error
Atmosphere (A)
Stone Type (S)
Interactions
Error
d.f. '
2
2
4
15
Sum of Squares
9.298
10.774
7.253
24.118
Mean Square
4.650
5.099
1.821
1.608
F Ratio
2.89
3.17
1.13
65
-------
Compare the F ratios for the data with tabulated F ratios:
F2 IS^0'95) = 3'68
F2 is^0-90) = 2-7° .
These F values lead us to conclude that stone type for the two stones
tested and atmosphere do not affect attrition. There is a chance of only
about 5 percent that we will conclude significant effects when none exist.
Initially, we may say with some certainty that there are no inter-
actions (F = 1.13) between stone type and atmosphere. With regard to
how stone type and atmosphere affect attrition, the conclusion is unclear;
there probably are effects on attrition caused by choice of stone type
and atmosphere; these effects, however, are not shown decisively by these
test data. The test data are plotted in Figure 32 with atmosphere and
stone type plotted as parameters. Some individual data are included to
show the variability of the data.
Accepting the hypothesis that atmosphere and stone type do affect
attrition rate and that their effects are independent (do not interact) ,
we conclude that :
1. Tymochtee and Grove sorbents attrite least in the three atmo-
spheres tested. Greer limestone attrites notably faster than
either Tymochtee or Grove stone under the atmospheres tested.
2. The sulfating atmosphere causes least attrition; this is con-
30
sistent with Exxon's finding that sulfation of dolomite
creates a hard outer shell resulting in less attrition. The
worst attrition occurs in the atmosphere of 50 percent C0? in
50 percent nitrogen.
A separate sample of Chemstone ^ limestone recommended by Engelhard
Minerals and Chemicals Corporation was also tested. Its attrition rate
in a Cl (15 percent C02, 85 percent N~) atmosphere was 0.425 percent/hr,
intermediate between the other sorbents tested.
66
-------
Curve 713774-A
Single Datum
1.5
a i.o
fO
Of
c
.2
0.5
Cl
Mean
C2
Atmosphere
ST
GR = Greer
GV = Grove
TY = Tymochtee
TY
GV
Stone Type
GR
Figure 32 - Attrition Rate Dependence upon Stone Type
and Atmosphere
Effect of Duration of Fluidization. Time Held at 815°C. Heating Rate, and
Degree of Calcination of Attrition
A short series of tests on the 3.5-cm system (Tests Al through A10)
was completed for study of the effects of the duration of fluidization,
the time the stone is held at 815°C, the rate of heating, and the degree
of calcination to see their influence on the extent of attrition.
67
-------
Table 9
DESCRIPTION OF TEST CONDITIONS
Dwg. 1687835
Test
No.
Al
A2
A3
A4
A5
A6
A7
A8
A9
AID
Rate of
Heating
Rapid
Rapid
10°/m
10°/m
10°/m
10°/m
10°/m
10e/m
10°/m
10°/m
Extent of *
Calcination
100%
~ 10%
~ 10%
100%
100%
100%
100%
100%
100%
100%
Hours
Fluidized
0
0
0
0
0
t = 0. 16
t = l
t = 2
t = 4
t = 9
Hours
at815°C
1
0
0
4
1
1.16
2
3
5
10
n--II| ^ T_ A ^
and Gas Velocity ( )
'c
815-
600- N.
20 1 — T X-V-"
.,«
20 L 7^5*
815- >v
60°- /.. \ 0.5 U
20-Z.J I..X- o ™f
815- . .
600 " /._.. V- 0-5 Urf
20.4..S' T..TS.-0 "
815- > .
600- /•• -7\-.1.2Umf
/r->- 0.5 U ; \
20-/C- J mf V-..S.
f« • t hours «J
ta
Tests compared to determine effects of time of fluidization: A6 through A10
hours at 815°C: A4 & A5
rate of heating: Al & A5. A2 & A3
ext. of calcination: A3 & A5, Al & A2
• Calcination was complete except for runs A2 and A3. in which bed
cooling was initiated as soon as the sample reached 815eC
68
-------
Dwg.Z6Z4C61
VO
Table 10
ATTRITION TEST DATA
(Grove Limestone)
Test
Date
10-14
10-15
10-8
10-7
10-18
10-19
10-22
10-28
10-25
10-20
12-2
1-7
Test
No.
A-l
A-l
(Re-
peat)
A-2
A-3
A-4
A-5
A-6
A-7
A-8
A-9
A-10
A-10
Degree
of
Calcin-
ation
100%
100%
-10%
-10%
100%
100%
100%
100%
100%
100%
100%
100%
Rate of
Heating
Rapid
Rapid
Rapid
Gradual
Gradual
Gradual
Gradual
Gradual
Gradual
Gradual
Gradual
Gradual
Mrs at
1
1
0
0
4
1
1.16
2
3
5
10
10
Mrs
Fluidized
0
0
0
0
0
0
0.16
1
2
. 4
9
9
Bed Before/Bed After
To^b
0.37237
0.0010
0.37237
0.0006
0.37237
0.3410
0.37237
0.3640
0.37237
0.0012
0.3723/
0
0.37237
0.0018
0.37237
0
0.37237
0.0001
0.37237
0.0002
0.37237
0.0037
0.3723/
0
Fo/Fb
0.00047
0.00856
0.00047
0.00661
0.00047
0.01852
0.00047
0.00369
0.00047
0.04571
0.00047
0.00502
0.00047
0.00821
0.00047
0.01289
0.00047
0.01681
0.00047
0.01518
0.00047
0.01597
0.00047
0.02673
Mo/Mb
70
44.35
70
43.8296
70
66.9612
70
67.7293
70
43.6613
70
43.7421
70
43.8235
70
43.4447
70
43.3879
70
43.4248
70
43.1842
70
42.6621
Xo/Xb
0.321
0.5296
0.321
0.5421
0.321
0.3435
0.321
0.3263
0.321
0.5475
0.321
0.5354
0.321
0.5276
0.321
0.5662
0.321
0.5686
0.321
0.5686
0.321
0.509$
0.321
0.5370
Filter
Solids, g
Mf
0.0302
0.0111
0
0
0.0069
0.0069
0.0170
0.0514
0.1929
0.1023
0.1403
0.1650
0.4674
Mass Balance
Stone
100.57
99.82
100.27
98.77
99.54
99.74
99.80
99.56
99.35
99.45
98.93
98.84
Calcium
104.60
105.77
102.36
98.35
106.40
104.27
103.02
109.95
110.05
110.24
98.29
103.07
Fines
Formed.
0.892
0.645
1.820
0.322
4.513
0.499
0.894
1.673
1.852
1.779
1.898
3.619
-------
In two of these tests, Grove limestone was heated as rapidly as
possible by pouring cold stone directly into the preheated cell. In the
remaining tests, the Grove limestone was heated in the attrition test
cell at 10°C/min from room temperature to 815°C; this gradual heating
avoided attrition from thermal stress. At 600°C nitrogen flow was turned
to 1/2 U ,. to carry CCL from the bed and aid calcination. After holding
the bed at 815°C for one hour under a gas flow of 1/2 U f, we fluidized
the bed by increasing the gas flow to 1.2 U £. The patterns of con-
mr
trolling temperature and gas flow are shown in Figure 33; in tests Al and
A2, however, the stone was brought to 815°C within several seconds. A
description of test conditions and which tests were compared to draw
various conclusions is listed in Table 9.
Size distribution, masses of bed and filter solids, loss on ignition
(a measure of C0_ loss in calcining), and calcium fraction were measured
before and after attrition testing.
Data from these tests are listed in Table 10. The mass balances are
calculated from:
Fully Calcined Stone Mass Balance
f[" Mass of "I [Mass left] [Liberated]^ ._ ("Initial ]
"[[filter solidsj [ in bed J [ C02 Jj ' [bed massj
M- (1-Y,) + M. (1-Y,) + M Y * M
I t 1 J. O O O
Calcium Mass Balance
[Calcium in Filter and Bed after the Test]
[Calcium Charged to the Bed]
(M,X, + H.X, * M X )
11 J. J. o o
The extent of attrition is calculated from
i
f[Final Bed] [Final Filter] ["initial Bed]l . [initial Bed]
\[ Fines J L Fines J ~ L Fines Jj ' L Mass J
based on the fully calcined stone:
u-V fi + Mf (1-V ff - Mo (1-V fo] * [MO (1-V]
70
-------
Dwg. 7681A36
o
o
E
CD
r'-'Umf
rv—.0.5Umf
Figure 33 - Gas Velocity and Temperature Patterns
in the Attrition Versus Time Tests
where:
Mf
M,
mass of filter solids
final mass of bed solids
M = initial mass charged to the bed
Yf = fraction of CCL in filter solids as determined by LOI
Y. = fraction of CO- in final bed solids as determined by LOI
Y = fraction of CO- in original stone as determined by LOI
X,. = fraction of calcium in filter solids
X = fraction of calcium in final bed solids
ff = fraction of filter solids smaller than 710 ym
f, = fraction of final bed solids smaller than 710 ym
f = fraction of original bed smaller than 710 ym.
Dependence of the Extent of Attrition on Duration of Fluidization
The results of tests A5 through A10, inclusive, infer the dependence
of the extent of attrition on duration of fluidization. The amount of
fines generated in test A5, 0.499 percent, shows that fines are formed in
the absence of fluidization; the processes of heating and calcination
alone cause attrition. In order to show only the effect of fluidization
71
-------
of Grove limestone on attrition, the concentration of fines caused by
heating and calcination of Grove limestone has been subtracted from the
total fines formed. The percentages of fines caused by fluidization alone
are listed in Table 11 and plotted in Figure 34.
31
The attrition theory proposed by Kutyavina and Baskakov models
attrition as a rapid initial rate from the fracture of projections and
defects that decreases as particles become rounded and free from defects.
Finally, the rate of attrition reaches a constant value after six to
eight hours in their tests. Our earlier tests, which extended over a
six-hour interval, indicated a constant attrition rate. The test data
in Table 11 and Figure 34 agree with the Kutyavina and Baskakov model
M k x t a
Total Mass of Proportionality Time Exponent
Fines Formed Constant-
Curve 688663-A
iZ
**
Least - Squares Power Curve
M = 0.917 t°'382( r = 0.889
123456789
Time of Fluidization, hr
Figure 34 - Effect of Duration of Fluidization on Extent of
Attrition in Fluidization of Grove Limestone
at 815°C
72
-------
Table 11
PERCENTAGES OF FINES FORMED DURING ATTRITION TESTING AND
PERCENTAGES ATTRIBUTABLE TO FLUIDIZATION ONLY
(Grove Limestone)
Time of Fluidization
Percent of Fines
Total
Less Fines at t=0
0
0.16
1
2
4
9
9
0.4988
0.8939
1.6730
1.8520
1.7794
1.8978
3.619
0.0000
0.3951
1.1742
1.3532
1.2806
1.3990
3.120
The data from our tests, in least squares analysis, are represented by
M(%) = 0.917 t°'382
with a correlation coefficient of r = 0.89. The average rate of attrition
over the nine-hour interval was 0.23 percent/hr.
Dependence of the Rate of Attrition on the Hours the Stone Was Held
at 815°C
Tests A4 and A5 were designed to indicate the effect on particle
attrition of holding the bed temperature at 815°C without fluidization.
A gas flow of one-half the minimum fluidization velocity was maintained
to assure complete calcination. The results from Table 10 are
Hours Stone Held
815°C
1
4
Percent Attrition
0.499
4.513
Degree of Calcination
100.0
99.8
73
-------
This difference appears to be significant and indicates that holding
Grove limestone at 815°C without fluidization causes attrition. The
degree of calcination is virtually the same in both tests and does not
confound the conclusion.
Dependence of the Rate of Attrition on the Rate of Heating and the
Degree of Calcination
A literature survey of attrition mechanisms has suggested that thermal
shock or (thermal stress) may aggravate attrition. Two mechanisms are
possible:
• Stone may shatter immediately upon thermal shocking
• Stone may be catastrophically weakened by thermal shock.
After fluidization begins the new defects formed through the second
mechanism may fail, increasing the attrition rate. Only the first
mechanism has been tested in these experiments.
In the earlier of these studies (Runs A2 and A3), identical masses
of Grove limestone were heated in the attrition test cell. The first
test involved heating 70 g of stone at 10°C/min to 815°C, then allowing
the mass to cool slowly. In the second test cold stone was suddenly
charged to the hot cell and was cooled in the same manner. No fluidiz-
ing air was fed in these tests.
Mass balances over both tests gave excellent accounting for products
and reactants. Gradual heating at 108C/min caused reduction of 0.32 per-
cent of the stone to smaller than 710 ym. Shock heating at approximately
400°C/min caused 1.82 percent reduction to smaller than 710 pm.
In the later study (Runs Al and A5) the two rates of heating were a
gradual warming of 10 ± 0.1°C/min and a sudden shock heating in which the
stone was poured into the hot cell and came to a temperature of 400°C
within one minute and 800°C in about eight minutes.
Figure 35 shows the temperature history for runs Al, Al repeated,
and A2. The conditions were nearly identical in all three tests and the
temperature curves are similar. Figure 35 actually shows the thermocouple
74
-------
Curve 713775-A
Attrition Test Cell
Thermocouple
= Run Al
a = Run Al repeat
o = Run A2
123456789
Minutes after Adding Cold Stone to the Hot Attrition Ceil
10
Figure 35 - Temperature History when Cold Grove Limestone is Added to
815°C 3.5 cm Attrition Cell (Runs Al, Al Repeated, A2)
temperature, not the stone temperature. The thermocouple temperature was
lowered rapidly by the cold stone dumped on it. As the stone temperature
rose, the thermocouple temperature also began to climb, after about
90 seconds.
The degree of calcination was controlled by allowing time for cal-
cination to go to completion ("vlOO percent) or by cooling the stone
immediately after the heating to achieve approximately 10 percent calci-
nation. Achieving a 10 percent calcination was the goal, and the levels
of 12.4 percent and 5.4 percent, respectively, actually obtained are
considered close enough to the 10 percent goal for our purposes.
75
-------
The results pertaining to the thermal shock effects, as extracted
from Table 10, are listed below.
100% Calcined (nominal)
0 hr Fluidized
1 hr at 815°C
Test
No.
Al
Al
(repeated)
A5
Rate of
Heating,
%nin
Rapid
(400)
Rapid
(400)
Gradual
(10)
%
Attrit.
0.892
0.645
•a
0.4990
% of
CaC03
Actually
Calcined
99.9
99.9
100.0
10% Calcined (nominal)
0 hr Fluidized
0 hr at 815°C
Test
No.
A2
A3
Rate of
Heating
°/min
Rapid
(400)
Gradual
(10)
%
Attrit.
1.820
0.322
% of
CaC03
Actually
Calcined
12.4
5.4
Average = 0.768
These results indicate an interaction between R.. the rate of heating
(°C/min) and C the extent of calcination (%) influence the extent of
attrition A. If K is defined as an unspecified constant, the model for
extent of attrition A is of the form
A =
+ K2) (K3C + K4)
Here k is merely a regrouping of K. The data in the above table when
applied to this model give four equations for the four observations.
(400) (0.999) +
(10) (1) +
(400) + k.j (0.999) +
(10)
(1) +
= 0.768
= 0.499
76
-------
kL (400) (0.124) + k2 (400) + k3 (0.124) + k4 = 1.820
k{ (10) (0.054) + k2 (10) + k3 (0.054) + k^ = 0.322 .
We have transformed the variables RQ and C so that they range from -1 to
o
+1 and rewrite the equations as
1.0 ^ + 1.0 k2 + 0.996 k_ + k, = 0.768
-0.952 kL - 1.0 k2 + 1.0 kj + k4 = 0.499
-0.754 kL + 1.0 k2 + 0.850 k3 + k4 = 1.820
-1.0 kL - 1.0 k2 - 1.0 k3 + k^ = 0.322
with solution
k- = -0.71 coefficient on the interaction RQC
1 o
k2 = +0.83 coefficient on the heating rate effect R
k« » +0.11 coefficient on the degree of calcination effect C
k, » +0.55 equation constant.
These normalized coefficients are comparable and tell the relative effects
of each variable in the absence of fluidization. The positive coeffi-
cients on the rate of heating R (thermal shock) and the degree of cal-
cination C are expected; the negative interaction is unexpected.
The results of these tests suggest avoiding the thermal shock to
sorbent in a fluidized-bed combustor. Thermal shock and its attendant
attrition can be minimized by controlled heating of the sorbent with the
exhaust gases from the combustor.
Change in Limestone Sorbent Shape during Fluidization
After fluidizing Grove 1359 limestone at room temperature in the
7-cm Plexiglas column for 329 hours, we observed that the particles
77
-------
seemed to be about as angular as before. The fluidizing velocity was
46 cm/s, 30 cm/s above the U f of 16 cm/s, and fluidization was vigorous.
The only apparent difference was that the evident dustiness of particles
before fluidization disappeared during fluidization. There was no evi-
dent change in the incidence of points or sharp edges.
We had expected that corners or edges would be knocked off particles
during fluidization. Although 9.5 weight percent of the largest mesh
size (350 to 500 ym; 32 to 42 mesh) was lost in fluidization, it appar-
ently was not lost from edges or vertices. To test the hypothesis that
particle shape remained unchanged in fluidization we measured the shapes
of 55 particles before and 55 particles after fluidization.
The procedure was to measure first the longest dimension of a par-
ticle, then measure the largest perpendicular dimension (Figure 36).
The particle shape was calculated as the ratio of the largest dimension
to the greatest perpendicular dimension. All measurements were taken
from 14X magnification photographs of representative fractions of the
particle population.
Dwg. 7681A33
Figure 36 - Measurement of Perpendicular Dimensions
for Measurement of Particle Shape
78
-------
Figure 37 shows the photos from which we measured the particle
dimensions.
The results of analyzing the shape measuremental data are:
Mean Ratio
Standard Deviation
n
Before
1.6573
0.335
55
After
1.6077
0.354
55
fc^i i *- A = °-443 with 108 d-f; t«- KI ^ * 1-98
calculated ' tabled
The shape ratio change from 1.66 to 1.61 is not statistically significant
as inferred from the statistical t-test of the data.
Reference to Appendix D (Figure D7) shows a rapid initial rate of
attrition, presumably from weak projections and defects being knocked
from particles. For fluidization of uncalcined limestone in a cold sys-
tem, however, the reduction in size mentioned previously apparently
results from loss from flat particle surfaces, not from the sharp edges.
The sharp-cornered, knife-edged particle shape is maintained. This is
comparable to the process of flaking stone as in making arrowheads:
sharp edges and corners are preserved and chips are from flat surfaces
of uniform thickness as postulated in the attrition theory (Appendix D).
These results showed the need for hot testing to see if fluidization
at calcination temperatures (^800°C) preserves particle shape. Such hot
testing was completed in Runs Al through A10, discussed previously, and
is discussed further below.
Character of Wear in Fluidized-Bed Attrition
Following the tests of attrition dependence on time of fluidization,
shock heating, and calcination (Runs Al through A10), the Grove limestone
79
-------
A
1
As Crushed •• Before Fluidization
14 X
After 329 hr of Fluidization
14X
Figure 37 - Grove Limestone Particles before and after Fluidization
at U - U .. = 30 cm/s for 329 Hours
mf
RM-82046
-------
sorbent was photographed. The purpose of this close examination was to
investigate the character of attrition - do particles attrite by
splitting or by spalling tiny chips from the particle surface? - and to
inspect the extent of particle rounding after attrition. The micro-
graphs are shown in Figure 38. With each picture is a description of
test conditions.
Initial inspection of the pictures suggests that particles do not
calcine uniformly. While all of the original stone is a uniform gray,
the reacted stone, particularly after test A5, varies from the dead
white of fully calcined stone, to a gray like the original stone.
Initial inspection of the pictures also indicates that particles
do not split off large pieces (or split in half), but chip away bit-by-
bit from edges, corners, and defects. Compare photos of the original
stone and that from test A10; the rounding of particles from nine hours
of fluidization at 815°C is evident. ANL reported that the "mechanism
for attrition is abrasion (the wearing away of surface material) in con-
trast to the breakup or splitting of particles due to particle-particle
or particle-wall collistions." Kutyavina and Baskakov begin their
paper, "With fluidization, particles are ground by abrasion and splitting.
Abrasion is evidently predominant even for brittle and insufficiently
strong materials." They continue to explain, "Experiments ... showed
that the rate of abrasion decreases over the course of time, with rubbing
off of the uneven parts and a decrease in the number of defects of the
particles." The results of the Westinghouse test described on pages 72
and 73 tend to support this conclusion that the attrition rate is
highest in the beginning of the test when abrasion of projections and
defects would be the greatest. The rate of fines (<710 um) production
R (%/hr) observed in the Westinghouse study reported here varied with
time t (hours) as:
R = dA/dt = d(0.917 t°'382)/dt
= 0.350 t~°-618 '
81
-------
Micrograph, 10X
Test& Description
Original Stone
millimeters
0123
micrograph scale
Micrograph, 10X
Al
Test & Description
10%
Rapid
Ohrs
Ohrs
0.892%
Calcined
Heatup Rate
Fluidization
at815°C
Fines Production
Micrograph, 10X
Test & Description
100*
Rapid
Ohrs
Ihr
0.644%
Calcined
Heatup Rate
Fluidization
at815°C
Fines Production
A2
Micrograph, 10X
Test & Description
Rapid
Ohrs
Ohrs
1.820%
Calcined
Heatup Rate
Fluidization
at815°C
Fines Production
Figure 38 - Sorbent Micrographs - Grove Limestone
82
RM-70617
-------
Micrograph, 10X
Test& Description
10%
Rapid
Ohrs
Ohrs
0.322
Calcined
Heatup Rate
Fluidization
at815°C
Fines Production
Micrograph, 10X
Test & Description
A4 100% Calcined
10° C/min Heatup Rate
0 hr Fluidization
4hr at815°C
4.513% Fines Production
0123
micrograph scale
Micrograph, 10X
Test & Description
A5 100* Calcined
10°C/min Heatup Rate
0 hr Fluidization
Ihr at815°C
0.499% Fines Production
Micrograph, 10X
Test & Description
A6 lOW Calcined
10°C/min Heatup Rate
0.16 hr Fluidization
1.16hr at815°C
0.894% Fines Production
Figure 38 (Continued)
83
RM-70618
-------
Micrograph, 10X
Test & Description
A7 100% Calcined
10°C/min Heatup Rate
1 hr Fluidization
2hr at815°C
1.673% Fines Production
Micrograph, 10X
Test & Description
A8 100% Calcined
Heatup Rate
100%
10°C/min
2hr
3hr
1.852%
Fluidization
at815°C
Fines Production
Test & Description
A9 100% Calcined
Heatup Rate
Fluidization
100%
10°C/min
4hr
5hr
1.779%
at815°C
Fines Production
Micrograph, 10X
millimeters
0 1 2 3
micrograph scale
Figure 38 (Continued)
RM-70619
-------
Micrograph, 10X
A 10 100%
10° c/mm
9hr
10 hr
1.898
Test & Description
Calcined
Heatup Rate
Fluidization
at815°C
Fines Production
Test & Description
A10'
100%
10°C/min
9hr
10 hr
3.619%
Calcined
Heatup Rate
Fluidization
at815°C
Fines Production
Figure 38 (Continued)
Micrograph, 10X
millimeters
0 1 2 3
micrograph scale
85
RM-70620
-------
for which the attrition rate of 0.35%/hr at t = 1 hr falls to half
(0.175%/hr) at t = 3.07 hours. The decreasing rate of attrition is
referenced by Stanley et al.
of course material to fine L:
33
referenced by Stanley et al. in their expression for the rate of loss
and by Mathur and Epstein in their comment, "The grinding rate tended
to drop off with spouting time, as would be expected in any batch
35
grinding operation," and by Gonzales and Otero in their empirical
o f
equation d/dt D = -CD ' in which D is volume/surface diameter and
vs vs vs
C is a constant accounting for system conditions and particle strength.
This rate equation can be solved to give:
D = (D -1'6 + 1.6 ct> -1'1'6 ,
vs \ o J '
which implies a decreasing attrition rate with time, or rearranged to
yield :
-3c (Sjjt)
1-m'
3
" 1-m'
3
1-m1
2+m
»„_<,,. I*2tL] -T— M° _r n-mM [SSE.1 t
dt
where M^ is the bed mass, N is the number of particles in the bed, p is
particle density, and m' is a constant =2.6.
Further evidence that attrition is a result of surface abrasion,
not splitting, is seen in the size distributions of particles before and
after fluidization. Figure 39 shows differential size distribution
curves of material before and after test A-9. Note that there is little
difference in the fractions of particles larger than 710 pm; the change
in distribution occurs in the fines smaller than 710 ym. This result
suggests that tiny chips are being broken off of the larger particles,
not substantially changing the size of the larger particles but greatly
increasing the quantity of fines.
86
-------
Curve 689054-B
1.0
S
o
CO
o>
0.1
ii 0.01
0.001
0.0001
I I
After Testing
Before Testing
T
r
100 1000
Particle Sieve Diameter, pm
Figure 39 - Particle Size Distributions before and after Hot
Fluidization of Grove Limestone, Run A-9
87
-------
The pictures shown in Figure 38 were ranked from "most angular" to
"most rounded" by six observers. Averaged rankings are shown in Fig-
ure 40. As might be expected, the as-received stone was judged to be
angular, the stone fluidized for nine hours (A-l) was judged most
rounded. Aside from these extremes in treatment, however, there was
generally poor discrimination of attrition effects, judging by the photo-
graphs. Figure 41 shows little perceived difference in angularity
between particles fluidized for 0.16, 1, 2, 4 and 9 hours. The extent
of attrition test A-l through A-10 (see Table 10) correlated well with
time of fluidization (see Table 11) but agreed poorly with average
rankings from inspecting the pictures (see Table 12). These rankings
and extents of attrition correlated with r = 0.028, which is virtually
no relation.
Table 12
RELATION BETWEEN EXTENT OF ATTRITION AND DEGREE
OF ROUNDNESS AS JUDGED BY SIX OBSERVERS
Test Number
Ranking ,
Roundness
Attrition,
%
Al
1.83
0.77
A2
5.67
1.82
A3
6.17
0.32
A4
4.00
4.51
A5
3.00
0.50
A6
8.83
0.89
A7
8.17
1.67
A8
7.33
1.85
A9
8.83
1.78
A10
10.00
1.90
It seems that particle rounding is not apparent because the particles
are indeed relatively unchanged in shape. With less than 5 percent of
particle mass attrited, particle appearance is not changed appreciably.
This study suggests that limestone particles subjected to graded
heating, shock heating,* calcination, and fluidization attrite by abrasion
of surface defects, not by splitting; and visual inspection of particles
subjected to less than 5 weight percent attrition cannot discriminate
the extent of attrition very closely.
*Tests conducted after completion of this contract demonstrated that
severe thermal shock causes splitting of particles into large fragments,
88
-------
Curve 689289-A
Rounded 10 r
If
c II
Al AR A5 A4 A2 A3 AS A7 A6 A9 AID
Increasing Angularity
Increasing Roundness
Figure 40 - Mean Rankings of Particle Angularity
Curve 689290-A
S 8
u
|l 6
j!
CT1 U
.S 1- 4
.* n>
c. "5
S 2
qular
-
-
_
~ Test No.
- Mrs. Fluidized
A5
0
A6
0.16
A7
1
A8
2
A9
4
•
A 10
9
Figure 41 - Mean Rankings of Particle Angularity for the
Effect-of-Duration Tests
89
-------
Tests for Particle Swelling
In most attrition tests performed there appears to be an increase
in the fraction of particles on the largest sieve mesh (as exemplified
in Figure 39). One expects a decrease in frequency of the largest par-
ticles because of attrition. If the number of larger particle increases,
it is because smaller particles swelled during treatment, or it is only
an apparent increase in the fraction of large particles because of par-
ticle nonsphericity, as illustrated in Figure 42.
Dwg. 7681A35
These particles negotiated
the sieve . . .
... but some were caught
on a second pass because they
approached the mesh broadside
Figure 42 - Apparent Swelling of Sausage-Shaped Particles
Particle Swelling during Calcination
Comparison of screen analyses before and after calcination of sor-
bent has suggested that coarse particles swell when calcined. Micro-
scopic examination of the particles does not show any evident changes
such as with popcorn or puffed cereals, yet the screen analyses con-
sistently indicate an increase in the number of larger particles.
90
-------
In continued testing of particle swelling, 140 g of 710 to 1000 pm
(16 by 24 mesh) Grove limestone was divided into equal halves. One-half
was fluidized for four hours at 815°C with a gas flow of 1.2 times the
minimum fluidizing velocity. Then both the treated half and the
untreated half were screened to remove particles smaller than 710 ym
(24 mesh).
Particles were photographed at 10X (Figure 43) and analyzed by the
Zeiss (manual) and Leitz (automatic) methods. Results of the photograph
measurement and statistical analysis are listed in Table 13.
Table 13
COMPARISON OF MEAN SIZES OF GROVE LIMESTONE
PARTICLES CALCINED AND UNTREATED
Method Zeiss
Total
Sample Untreated T
Particles 340
Mean Diameter, mm 0.83
Standard Deviation, mm 0.2
Pooled
Std. Dev., mm 0.2
Leitz
reated Untreated Treated
262 62 65
0.94 0.88 0.96
0.2 0.21 0.22
0.215
Statistic t & d.f. 6.69, 600 2.25, 125
Tabulated t, 5% level 1.96
Difference Significant ? Yes
0.98
Yes
While not conclusive with regard to all sorbents, these results for
Grove limestone do support the hypothesis that stone particles do indeed
swell upon calcination.
91
-------
10X Before
10X After
Figure 43 - Micrographs of Grove Limestone before
and after Calcination
92
RM-70884
-------
Some nonspherical particles that pass a sieve on a first screening
may not pass the sieve on a rescreening because elongated particles
approach the screen sideway and do not pass. This appearance of parti-
cle swelling was tested using uncalcined Tymochtee dolomite. We sieved
the sorbent, keeping that fraction that passed a 12 mesh (1410 urn) screen
and was retained on a 16 mesh (1000 ym) screen after 10 minutes shaking
on a mechanical sieve shaker. This fraction was resieved with a further
10 minutes of mechanical shaking. This experiment was replicated using
two Tymochtee samples. One would expect that all particles would again
pass the 12 mesh (1410 ym) screen and not pass the 16 mesh (1000 ym)
screen.
Table 14 lists results of this experiment. There is good replica-
tion of the tests.
Table 14
EFFECT OF RESIEVING A SINGLE SIZE FRACTION OF TYMOCHTEE DOLOMITE
Tyler
Mesh
Opening
ym
DP
1st Replicate
12
16
24
1410
1000
710
2nd Replicate
12
16
24
1410
1000
710
After First Sieving
Grams
on
Sieve
70.000
70.000
Percent
on
Sieve
100.000
100.000
%<
DP
100.000
0.000
100.000
0.000
After Second Sieving
Grams
on
Sieve
0.26
68.14
1.57
0.20
68.07
1.70
Percent
on
Sieve
0.372
97.384
2.244
0.286
97.285
2.429
%<
DP
100.000
99.628
2.244
0
100.000
99.714
2.429
0
93
-------
We conclude from this test that size distribution from identical
successive sievings are not the same and that:
1. There appears to be an increase of larger sizes.
2. There appears to be an increase in smaller sizes, either by
attrition or by particles passing that did not pass the first
time.
These two tests indicate that there are separate, noninteracting
processes that contribute to the observance of sorbent particle swelling.
Calcination results in an increase in particle size, independent of sieve
measurement; the very process of sieving measurement gives the appearance
of particle size increase with uncalcined sorbent. These real and apparent
increases need to be accounted for in precise measurement of particle size
measurement in sorbent reaction processes.
Effect of Bed Pressure on Attrition Rate
It is difficult to state a priori whether varying pressure directly
influences attrition rate, all other variables including (U-U C)/U -
mt ml
being equal. The increased density of fluidizing gas may accelerate par-
ticles faster and increase the extent of collisions. An indirect effect
of pressure may be its influence on degree of calcination:
increased increased decreased rate change in change in
press re ~*" C^2 partial ->• (or extent) of •* particle •> attrition
pressure calcination strength rate
Of course, this secondary effect will vanish with C02~free fluidizing gas
and calcined stone.
The purpose of this experiment was to measure the influence of pres-
sure on the extent of attrition, with other variables held constant.
Test conditions, apparatus, and procedure in this test were as follows:
Sorbent Tymochtee dolomite
Particle diameter 1000 to 1410 ym (12 to 16 mesh)
94
-------
Temperature history
Gas pressure
Gas composition
Gas velocity
Bed diameter
Bed depth
heat at 10°C/m to 815°C, hold for 4 hr, cool
100 and 1000 kPaa (1 and 10 atm absolute)
15 vol % C02 in N2
U = 1.2 U
3.5 cm
mf
3 cm (initial charge of uncalcined stone
70g)
Time of fluidization 4 hours
Figure 44 shows the size distribution curves for two replicate tests
at 100 kPaa (O,0) and two replicate tests at 1000 kPaa (A,D). There
are slight differences between the mean curves in Figure 44, but the
curves are interpreted to be essentially the same. Those size fractions,
comprising over half of the mass of fines (500-700, 350-500, 245-350,
43-61 urn) (24-32, 32-42, 42-60, 250-325 mesh), are statistically the
same for both pressures.
The cumulative results from this test program are:
Pressure, kPaa
100
1000
Percent of Sample Reduced to
Smaller than 710 ym
1st Replicate
0.666
0.534
2nd Replicate
0.381
0.516
A t-test of these data at the 10 percent level gives a calculated t of
0.01 compared with the tabulated t of 2.92, indicating that there is no
significant effect of pressure.
The expression describing attrition rate in the bubbling zone of a
fluidized bed
Rz
U-U
m
mf
95
-------
Curve 690295-B
1.0
_c
"S.
o
c
o>
.O
in
O
i/l
to
(O
0.1
0.01
0.001
aa
"^Initial Charge
100 1000
Dp, Particle Diameter, pm
Figure 44 - Size Distributions of Tymochtee Dolomite after
Hot Fluidization at 100 and 1000 kPa
96
-------
is independent of gas pressure. Any pressure effects are accounted for
in the dependence of U , on gas density; by relating attrition rate R to
the excess gas velocity U-U f, the effects of pressure vanish. The
experimental results reported here confirm that independence of pressure.
Effects of Bed Temperature, Particle Diameter, and Degree of Sulfation on
Extent of Attrition
The theoretical study described in Appendix D concludes that attri-
tion rate in the bubbling zone of a fluidized bed varies inversely with
particle strength and is independent of particle diameter.
As part of our comprehensive study of attrition we chose to study
the effects of three additional variables on the extent of attrition:
1. Fluidized-bed temperature because it affects particle strength
and because it is an important controlled variable in any fluid-
ized-bed combustion process.
2. Particle diameter because of uncertainty in the literature as
to the effect of particle size on attrition rate
3. Degree of sulfation, as there is reason to believe that sul-
fating stone sorbents makes them harder and more resistant to
attrition.
28 ^6
Craig and others and Curran and others have reported that sulfating
sorbent grains increases particle strength and reduces attrition.
Description of the Apparatus and Test Plan
We fluidized Grove 1359 limestone in an attrition test cell 8,7 cm
in diameter. A three-zone furnace heated the cell and bed to the test
temperature. Figure 45 shows the apparatus and Figure 46 outlines the
procedure. All stone was precalcined, carefully sized, and charged to
the attrition test cell to a depth-T-diameter ratio of 1.0 to avoid
slugging. Attrition measurements were based upon the mass of material
that was reduced to a particle size smaller than the mesh size of the
initial charge. Percent attrition is defined as 100 x In (MQ/M,) where
MQ = mass of the original charge and solids in the bed and M, = mass of
97
-------
Dwp. 1693BO)
Attrition
Cell
Ruidized-
Bed Section"
Not to Scale
0.9375
Figure 45 - Attrition Test Cell
98
-------
Oxg. 26-13C6b
c
4000 g Grove Limestone
24 - 32 Mesh
Sieve out and Discard
Fines < 32 Mesh
Sieve out and Discard
Fines < 32 Mesh
c
~ 4000 g Grove Limestone
12 -16 Mesh
Split
7 V
(same processing as with
24 - 32 mesh stone)
(same processing as with
24 - 32 mesh stone)
* The symbol \$- \ means that
1C On j
1 gram of solid materials Is
removed for assay of CO. content.
Figure 46 - Test Procedure for Attrition Testing of Grove Limestone
for Effects of Grain Size, Temperature, and Sulfation
99
-------
solids remaining in the bed after a time interval of fluidization. The
mass of material smaller than the initial mesh size was determined by
measuring the amount of fines in the bed and the fines carried out of the
bed and captured in the off-gas filter.
Our theoretical studies (Appendix D) infer that the degree of
attrition in freely bubbling fluidization is proportional to the pressure-
volume power delivered to the bed in the form of bubbles. This is equiv-
alent to saying that the attrition rate R <* (U-Umf), not a multiple of
U or U/U f. From inspection of bubbling hot beds we chose U-U f = 20 cm/s
as the excess velocity that gives vigorous bubbling without excessive
splashing. In each test we measured the minimum fluidization velocity
U , from a graph of AP across the bed versus the superficial velocity of
hot gas flowing through the bed (Figure 47). The gas mixture used in
these attrition tests was 85 percent nitrogen and 15 percent C0».
We chose a full-factorial design that allows study of three vari-
ables each at two levels for main effects and interactions. The high and
low levels of variables were as shown in Table 15.
Table 15
HIGH AND LOW LEVELS OF THE INDEPENDENT VARIABLES
Variable Symbol (
Temperature T
Particle
Percent
Diameter D
of Possible Sulfation3 S
Low Level High Level
coded as -1) (coded as +1)
615°C 815°C
500 to 1000 to
710 ym 1410 ym
0% 10%b
Based on the amount of calcium in the stone sorbent
Nominal
100
-------
Dwg. 7681A34
AP
U
Figure 47 - Determination of U f from the AP-U Curve
Apparatus
The primary attrition test cell used in this work is pictured in
Figures 48 and 49. The cell is 9.5 cm in diameter and can be fitted
with a variety of perforated or sintered grids. Gas supplied from house
lines or cylinders is metered through any of several flowmeters, pre-
heated outside the cell, and flows through the cell, exiting through a
sintered-metal filter. A three-zone furnace surrounds the cell, allowing
fluidized-bed temperatures of up to 900°C. Thermocouples and pressure
taps allow measurement of bed temperature, bed pressure, bed AP, and
grid AP.
Results of the Tests
The procedure for calculating the fraction increase in surface area
of the stone is shown in the following sample calculation:
Initial mass of stone (649.8 g stone measured) (1-0.0147 frac CO-)
= 640.26 g
on a CCL-free basis
after calcination at
815°C
Mass of stone after (637.55 g stone measured) (1-0.0109 frac C02)
= 630.60 g
treatment, CCL-free
basis
Total loss of C02- 640.26 - 630.60 = 9.66 g
free stone by escape
from the system or
to the filter
101
-------
Dwg. 1700B40
Exhaust
S060REFCO
Manometer
Pressure Gauge
Flowmeter
_ Regulator
Figure 48 - Flow Diagram of Sorbent Attrition Test System,
Cell Diameter is 9.5 cm.
102
-------
Figure 49 - Attrition Test Cell. The Grid is Welded into the
Cylindrical Section near the Center. Cell Diameter
is 9.5 cm id.
103
RM-80597
-------
From measurements described in a later section we estimate the diameter
of eliminated particles to range between 2.8 and 3.5 ym and here take
the diameter of lost particles to be 3 ym. Using the formula for total
surface area of a powder
6 x mass
specific gravity x particle diameter '
we calculate the surface area of lost CCL-free stone to be
(6)(9.66) = 33 2A 2
(1.45) (0.003) J-^'^1 cm •
From specific gravity and degree of calcination measurements we developed
the relation for Grove 1359 limestone
sp. gr. =1.45+2.51 (wt fraction CO )
6 x mass x £(f,/d .)
Surface area of stone in the bed = —^—
specific gravity
= 6 x 637.55 x 8.9715
1.45 + (2.51)(0.0109)
= 23,230 cm2
where Z(f ,/d .) is the sum of mass fractions to mean-particle-diameter
i pi
ratios for the several sieve sizes
Area of stone in bed + fines
lost from the bed = 133,241 + 23,230 = 156,471
Area of initial charge - ' »•««
_. „. , . _ , 156,471 - 23.455 c ,_
Fractional increase of surface - z — ?. -;--- — 2 - = 5.67
area of solids = '
104
-------
The results of the hot attrition tests are listed in Table 16.
Coded values of T, D and S range from -1 to +1, corresponding to
the upper and lower limits of the range of each variable; for example:
T m Temperature, °C - 732.5 . Afc ^.^ T = _1; ^ ^^ T . +1 .
Interpretation of the Results
Analysis of the attrition test results (the two columns at the
right of Table 16) for the factorial model
% Attrition, A = A + A_T + A_D + AT_TD + ACS + A TS
o T D ID b la
+ ADSDS + NDS™ •
Increase in Powder Surface, P = P + P_T + PnD + PTnTD + P0S + PTCTS
O I U 11> b lb
+ PDSDS + PTDSTDS
gives these coefficients in Table 17.
The particle size distributions from before and after fluidization
were interpreted in two ways. The first response (A) is the percentage
of large particles lost to fines by attrition. The second response (P)
is the fractional increase in surface area of the bed contents (including
fines elutriated to the filter or lost from the system).
Both responses show an appreciable effect of temperature on attri-
tion. As temperature increases, the rate of attrition increases, too.
This effect is not surprising; one expects the sorbent strength a to
decrease at higher temperatures, and the rate of particle attrition is
inversely proportional to sorbent strength o.
Both responses show an effect of attrition rate decreasing as the
degree of sorbent sulfation increases. This too is consistent with the
77 *^fi
findings of other workers. '
105
-------
Table 16
THE RESULTS OF HIGH-TEMPERATURE ATTRITION TESTING OF GROVE 1359
Test Conditions
Date
1977
10/4
10/6
10/11
10/13
11/3
11/10
11/15
11/17
Temp
°C
815
815
815
650
650
815
650
650
°P
Mesh
(nominal)
12-16
24-32
12-16
24-32
12-16
24-32
12-16
24-32
%
Sulfa-
tion
12.4
0
0
9.4
13.3
14.7
0
0
Coded
Values For
Smaller
Mesh Size
0.9647
0.9715
0.9534
a 9750
a 9643
0.9750
0.9534
0.9787
Total
Mass
of Stone
g
649.81
488.71
478.87
543.45
556.97
532.10
454.00
386.00
M°. Mass
CCy Free
Stone >
Smaller
Mesh Size3
in the Bed
617.66
468.40
454.82
520.83
530.59
510.76
431.48
372.26
Total
Surface
Area of
Stone in
Bed. m2
2.35
3.28
1.67
3.80
2.02
3.73
1.58
2.62
After 4- Hour Fluidization
Frac. C02
in
Stone
a 0109
a 0182
a 0060
a 1612
a 0247
0.0115
0.0435
0.3430
Frac>
Smaller
Mesh Size
in Bed
0.9576
0.9609
a 9601
a 9748
0.9690
a 9614
a 9461
0.9816
Total
Mass
of Stone
g in Bed
637.55
468.24
472.21
634.31
551.59
523.50
463.90
580.30
M.Mass
ay Free
Stone >
Smaller
Mesh Size3
in Bed
603.86
441.61
449.47
518.65
522.52
497.51
419.80
374.26
Total
Surface
Area of
Stone in
Bed, m
15.65
27.11
13.99
6.54
18.9
12.49
13.15
1.34
% h
Attrition0
2.26
5.89
1.18
0.42
1.53
2.63
2.74
-0.53
S,-S'
s°
Increase in
Surface
Area
5.67
7.27
7.40
0.72
8.36
2.35
7.31
-.49
(a) Mass of CO. - Free Stone > Smaller Mesh Size = (1 - Fraction CO. m Stone M Fraction > Smaller Mesh Size) x ( Mass of Stone)
(b) 100x ln(M° -J- M')
-------
Table 17
FACTORIAL MODEL COEFFICIENTS DESCRIBING
ATTRITION OF GROVE 1351
1st Response,
% attrition as loss of
course particles
2nd Response,
fractional increase
in surface area
A AT AD
+2.1 +0.9 -0.2
P PT PD
+4.8 +0.9 +2.4
ATD
-1.1
PTD
-1.5
As
-0.4
ps
-0.6
ATS
-0.2
PTS
-1.1
ADS
+0.3
PDS
+0.4
ATDS
0.8
PTDS
+0.4
The effect of particle diameter on attrition rate is not clear from
these results. The analysis indicates that particle diameter has no
significant effect on the rate of loss of large particles, but the
analysis evidences a strong effect of particle size on the rate of crea-
tion of new surface. These results could be consistent (if big particles
generate smaller fragments without significantly changing the size of the
original particles). The conclusion of a significant effect between par-
ticle size and rate of surface formation, however, does not agree with
our model (Appendix D). One is tempted to disregard this effect in sup-
port of a theory already formulated; but the magnitude of the effect is
too great to ignore, and we recommend further study of the effect of par-
ticle size on attrition rate.
Among the interactions that of temperature and particle size is
notable in both responses. We have no explanation for this interaction.
These results suggest the advantage of keeping the bed temperature
down to reduce attrition. Similarly, this study confirms the avail of
partially sulfating sorbent to increase its strength and lessen attri-
tion losses.
107
-------
Conclusions
• The rate of Grove limestone attrition increases with fluidized-
bed temperature.
• The rate of Grove limestone attrition decreases with an increas-
ing degree of stone sulfation.
• Results are uncertain as to the effect of particle size on
limestone attrition rate during fluidization.
Severe Attrition Caused by a Jet at the Grid
This section describes severe attrition caused by an inadvertent jet
on the grid of the 10.3-cm fluidized bed. In assembling the attrition
test cell for testing on one day, we neglected to seal the grid at the
cell wall. Inspection of the cell after testing showed a slit leak about
3 cm long (see Figure 50). Fluidizing the cell showed jetting from this
leak and uneven gas distribution.
Dwg. 7681A32
Gas Flow
Figure 50 - Leak in 10.3-cm Fluidized Bed
108
-------
Conditions in the test were:
Particle size range 95.1%<1410 ym
0.2% < 710 urn
Temperature 815°C
Time of fluidization 4 hours
Minimum fluidization velocity 35.1 cm/s
Nominal gas flow velocity 45.8 cm/s
Gas composition 85% NZ, 15% C02
These operating conditions are not irregular; only the presence of the
jet along the wall was different from usual conditions.
Results of this test and an earlier test where no jet was detected
are shown in Figure 51. This figure presents the same data on two sets
of coordinates: the outside figure is plotted against an arithmetic
ordinate (vertical axis) that emphasizes the differences in higher fre-
quencies of particles. The inset curves are plotted against a logarith-
mic axis that details the lower frequencies. The As are a plot of
starting material for both attrition tests, the Os show the size distri-
bution of limestone after attrition testing with the grid jet, the O s
represent material size after a test without the grid jet. The cumulative
size distributions are shown in Figure 52.
The first and most obvious conclusion from these results is that
the jet at the grid caused severe attrition. There was a decrease in
the two largest classes of particles; these coarse particles were
attrited. Conversely, there was an increase in all smaller sizes:
Original stone
Fluidized 4 hours,
Fluidized 4 hours,
Percent of mass smaller than:
710 um
0.212
no grid jet 0.75
grid jet 6.17
1000 urn
5.37
6.63
21.79
1410 um
95.1
93.69
97.62
109
-------
o
i
*
1 r
A Original Stone
o Fluidized 4hrs; jet at grid/
o Fluidized 4 hrs: no jet
at grid
100
1000
40 60 80100
200
Particle Diameter, d , pro
400 600 1000
Figure 51 - Size Frequency Plots for Grove Limestone; Ordinate
Shown in Both Arithmetic and Logarithmic Coordinates
Curve 6909U-A
99.9
4-hour rluidi
zation with a jet
the grid
4-hour fluidization, no
grid jet
Particle Diameter, d , mn
P
Figure 52 - Cumulative Distributions for Grove Limestone
110
-------
A second observation is that the fraction of coarsest particles
decreased under the severe grid jet conditions. In virtually all tests
the fraction of coarsest particles increases, presumably from swelling.
Under the conditions of acute attrition, any swelling effect was over-
come by attrition of coarses, reducing their weight fraction from 4.9 to
2.3 percent. This suggests the influence of the grid design and the
importance of controlling high-energy jets. Such effects have been
eliminated in the bubbling-zone-attrition tests by use of sintered-metal
grid.
Model of Attrition in the Bubbling Zone of a Fluidized Bed
A major part of the experimental attrition work in this study was
development of a model describing the rate of attrition in the bubbling
zone above the influence of grid jets. The results of this study are
explained in detail in Appendix D.
The conclusions gained from this work are:
1. Attrition in the bubbling zone of a fluidized bed abrades par-
ticles, forming fine fragments. The attrition rate is defined
as:
_ Mass of fines formed/second
Mass of coarse bed solids
2. In the bubbling zone the attrition rate, R, is proportional to
the excess bubble velocity, [U(superficial gas velocity) -
U ..(minimum fluidization velocity)].
3. The attrition rate, R, in the bubbling zone is proportional to
the bed depth, Z, at any point. The overall production rate
of fines (g formed/s) is proportional to the square of the total
bed depth.
4. The bubbling zone attrition rate, R, is proportional to particle
density, p , inversely proportional to particle strength, a .
s s
5. The attrition rate in the bubbling zone is initially high; the
rate decreases with time to a steady state.
Ill
-------
6. The above conclusions are expressed symbolically as
RZ
U-U -
mf
[F(t) +1]
2
z
g a
6c s
m
; m
7. Research data published in the literature confirmed conclusion 3,
above, and the implied conclusion that attrition rate is not
affected by particle diameter.
8. Experimental results from our laboratory confirm conclusions 2
and 5, above.
9. Attrition rate in the bubbling zone of a fluidized bed can be
controlled not only by choice of a weak or strong granular mate-
rial, but by specification of bed depth, gas velocity, and
particle diameter as it affects U f.
Size Distribution of Fragments Attrited from Particles in the
Bubbling Zone of a Fluidized Bed
The purpose of this work was to measure the particle size of attrition
fragments elutriated from a bed of fluidized sorbent and compare the
measurements with the particle size distribution of fragments remaining
in the bed.
In the bubbling zone of a fluidized bed we have observed formation
of predominantly very fine fragments rather than large chips. Here we
have investigated the extent of very fine particles formed by bubbling-
zone attrition.
The apparatus we used was the 7-cm-id attrition test cell diagramed
in Figure 53. The sintered grid eliminates grid jets and allows bubbling
through the entire bed depth. The high freeboard permits splashing par-
ticles to fall back to the bed. Particles elutriated from the bed are
trapped in the Balston filter.
112
-------
Rotameter
.J
i
'
Pressure
Taps
Plenum
Filter
Plexiglas Fluidized Bed
6.99cm ID x 91.44cm High
, Sintered -Metal Distributor
Plate
— (]>) Pressure Gauge
S & K 31827
Rotameter
Valve
' Regulator
'House Air
Figure 53 -• Flow Diagram for Room-Temperature Fluidized Bed
Figure 54 - Photomicrograph of Elutriated Grove Limestone Recovered
from the Balston Filter, Magnification 500X
113
RM-74061
-------
Our test procedure was to charge the attrition test cell with 300 g of
uncalcined 32-x-42-mesh Grove 1359 limestone and measure the minimum
fluidization velocity U f from a velocity-AP curve. We then fluidized
the stone in air at U ,. + 25 cm/s. After 15 minutes of fluidization we
mr
disassembled the filter, measured the size of filter solids, and sieve-
analyzed the bed contents. Another interpretation of this test is
described in Appendix D following Equation 26.
The measurement of fine particle size was made from filter solids
collected in a separate fluidization. After fluidizing Grove limestone
in the same apparatus at U + 20 cm/s for 23 hr, we photographed fines
collected from the Balston filter. Figure 54 shows one such photograph
at 500X magnification. Measurement of all grains intersected by a line
drawn across the photograph yielded an estimate of the geometric mean
particle diameter at 2.77 um. This seemed small, and we checked the
*._.. _J._.. —UJ.^,. „„—.._,. analysis. The geometric mean diam-
eter of filter solids by Coulter Counter measurements was 3.45 um, which
compares well with the photographic measurements.
Results of the fluidization measurements for 15 minutes at U ,
ml
+ 25 cm/s are listed in Table 18. The specific surface contributions
for the several sieve size ranges after fluidization are presented in
Table 19. Inspection of the row of S. values shows that 90 percent of
the new specific surface formed is on particles elutriated from the bed.
Of the mass fractions of particles lost or formed, 4.2 percent were lost
in the largest size fraction, -32 + 42 mesh; 3.6 percent gained in the
broad midrange from 42 mesh to the pan (<43 pm in the bed); and 0.6 per-
cent gained in the filter or were lost as 3 um microfines. Thus,
14 percent of the attrition resulted in 3 um microfines on a mass basis.
Very small particles generated in fluidization may present problems
with downstream equipment or escape through particle control equipment
(filters). Continued awareness and analysis for microfines is needed,
as their presence can affect air quality or downstream system components.
114
-------
Table 18
DEPENDENCE OF SOLIDS SPECIFIC SURFACE ON TIME OF FLUIDIZATION
Dwg,l692B53
U-U . = 25cm/s
Time Interval, hr
Total Fluidization Time, hr
Mesh
42
60
115
250
325
Pan
Filter
Lost3
D: , Mean
Diameter,
cm
0.042
0.030
0.018
0.009
0.0052
0.0036
0.0003
0.0003
Specific Surface, cm /ga
% Increase in Specific Surface
Start
0
1/4
1/4
1/4
1/2
1/2
1
1
2
2
4
4
8
Mass of Solids on Sieve, g
330
0
0
0
0
0
0
0
53.91
0
316.188
10.857
a 429
0.418
0.048
0.016
1.336
0.709
101.46
88.2
315.815
10.366
0.328
0.148
0.010
0.009
2.487
0.837
130.25
141.6
311.631
12.516
0.510
0.181
0.013
0.009
3.984
1.156
171.69
218.5
309.081
12.869
0.599
0.122
0.012
0.009
5.691
1.617
220.92
309.8
303.903
15.849
0.723
0.186
0.025
0.009
7.643
1.662
266.54
394.4
300.957
16, 621
0.906
0.136
0.019
0.014
9.002
2.345
312.97
480.5
4.3 SORBENT REGENERATION
Westinghouse has performed an engineering evaluation of sorbent
regeneration for both atmospheric and pressurized FBC. This evaluation
has been reported in Regeneration o£ Calcium-Based S02 Sorbents for
Fluidized-Bed Combustion; Engineering Evaluation, Report No. EPA-600/
7-78-039 (March, 1978). No new work has been completed on sorbent
regeneration since that report was published.
In the report the economics of FBC power plants operated with regen-
eration are projected on the basis of current estimates of regeneration
process performance. Coal-feed reductive decomposition is the regenera-
tion process considered for AFBC; three regeneration schemes (two
reductive decomposition processes and a two-step process) are evaluated
for PFBC. Estimated costs of FBC power plants with regeneration are
compared with costs of FBC plants using once-through sorbent (no
115
-------
Table 19
DISTRIBUTION OF PARTICLE SURFACE AREA FOR VARIOUS PARTICLE SIZES
AFTER 15 MINUTES FLUIDIZATION OF GROVE LIMESTONE AT 25°C
Sieve Mesh Range
Mean Particle Size in
Size Range, dpi, cm
Mass of Solids in Size
Range, Mj
Contribution to Specific
Surface, Sb
Contribution to Specific
Surface before Fluidiza-
tion, cm2/g
Increase in Specific
Surface, AS^
-32 + 42
0.042
316.188
51.65
53.91
2.26
-42 + 60
0.030
10.856
2.48
0
2.48
-60 + 115
0.018
0.4287
0.16
0
0.16
-115 + 250
0.009
0.4175
0.32
0
0.32
-250 + 325
0.0052
0.0475
0.06
0
0.06
-325
0.0036
0.0162
0.03
0
0.03
Filter
0.0003
1.3359
30.55
0
30.55
Losses
0.0003
0.7094a
16.21
0
16.21
Total
101.46
53.91
52.07
Calculated from mass balance.
Surface area at solids in size range v total mass of solids;
p E M. 2.65 x 330 '
CThis is calculated by assuming spherical particles; not rigorously measured.
-------
regeneration). The economic feasibility of the regenerative system
depends on several variables, including, in particular, the sulfur
concentration achievable in the regenerator off-gas, the reduction in
fresh sorbent feed rate possible through regeneration, and the cost of
fresh sorbent and of solid residue disposal. The performance required
for the regenerative FBC system to achieve economic feasibility is pro-
jected, and critical development needs are discussed. An integrated
regeneration system for both AFBC and PFBC, capable of achieving the
performance necessary, has yet to be demonstrated experimentally.
The sulfur recovery system is, in general, the dominant subsystem
in the regeneration process. The economic and environmental potential,
and the technical feasibility of regeneration are uncertain. The regen-
erative AFBC power plant shows more economic potential than does the
regenerative PFBC plant, because of the higher levels of SO- achievable
from the regenerator in the atmospheric pressure case. Major technical
uncertainties are associated with the transport of high-temperature
solids between the combustor and the regenerator. The reduced rate of
sorbent consumption and solid residue disposal resulting from regenerative
operation is accompanied by a reduction in power plant efficiency
and an increase in the consumption of coal and other auxiliary fuels.
117
-------
5. SULFUR OXIDE CONTROL - ALTERNATIVE SORBENTS
Westinghouse has performed an evaluation of alternative sulfur
sorbents for FBC. This evaluation was reported in Alternatives to
Calcium-Based SC>2 Sorbents for Fluidized-Bed Combustion; Conceptual
Evaluation, Report No. EPA-600/7-78-005 (January, 1978).3 No new
work has been performed on alternative sorbents since that report was
published.
The report gives results of a conceptual engineering evaluation to
screen supported metal oxides as alternatives to natural calcium-based
sorbents (limestones and dolomites) for S02 control in atmospheric and
pressurized FBC processes. We evaluated alternative sorbents, using
three acceptance criteria.
• SO™ removal capability in the combustor, predicted b>
thermodynamics
• S0» concentrations achievable in the regenerator off-gas, accord-
ing to thermodynamics
• SO- concentrations of the regenerator, achievable on the basis
of the material and energy balances.
The evaluation identified 14 potentially acceptable sorbents for AFBC
and 11 for PFBC. We prepared cost estimates to project the maximum
acceptable loss rates for the alternative sorbents because of attrition
and/or deactivation. Loss rates must be less than 0.1 percent of bed
inventory per hour in order to compete economically with natural
calcium-based sorbents, even if maximum thermodynamic performance were
obtained. U.S. resources of some minerals may be of extreme importance
for many of the alternative metal oxide sorbents considered.
118
-------
Unless the high levels (near maximum) of SO™ or E^S projected in
this report can be achieved in the regeneration process, the alternative
sorbents will be economically unfeasible. At present, too little is
known about the cost of sorbent preparation. Feasible process technology
for sorbent preparation should be proposed and the economic feasibility
assessed. The availability of the sorbent materials and the support
materials requires more modeling in order properly to assess all of
the market factors. The area of industrial expansion into alternative
sorbent preparation and distribution must also be explored.
The environmental impact (spent sorbent dispsoal, trace metals
release, impact on NO , CO, particulates, etc.), of alternative sorbents
X
may lead to the elimination of certain candidate materials. The complex
sociological interactions of these materials must be explored before
development is committed.
119
-------
6. REFERENCES
1. Keairns, D. L., et al., Fluidized Bed Combustion Process Evaluation -
Phase II - Pressurized Fluidized Bed Coal Combustion Development,
Report to EPA, Westinghouse Research Laboratories, Pittsburgh, PA,
EPA-650/2-75-027c, September 1975, NTIS PB 246-116.
2. Newby, R. A., N. H. Ulerich, D. F. Ciliberti, and D. L. Keairns,
Effect of S0~ Emission Requirements on Fluidized-Bed Combustion
Systems, Preliminary Technical/Economic Assessment. Report to EPA,
Westinghouse Research and Development Center, Pittsburgh, PA,
August 1978, EPA-600/7-78-163.
3. Newby, R. A., and D. L. Keairns, Alternatives to Calcium-Based SO-
Sorbents for Fluidized-Bed Combustion: Conceptual Evaluation.
Report to EPA, Westinghouse Research and Development Center,
Pittsburgh, PA, January 1978, EPA-600/7-78-005.
4. Newby, R. A., S. Katta, and D. L. Keairns, Regeneration of Calcium-
Based Sorbent for Fluidized-Bed Combustion: Engineering Evaluation.
Report to EPA, Westinghouse Research and Development Center,
Pittsburgh, PA, March 1978, EPA-600/7-78-039.
5. Keairns, D. L., D. H. Archer, R. A. Newby, E. P. O'Neill, E. J. Vidt,
Evaluation of the Fluidized Bed Combustion Process, Vol. I. Report
to EPA, Westinghouse Research Laboratories, Pittsburgh, PA,
December 1973, EPA-650/2-73-048a, NTIS PB 231-162.
6. O'Neill, E. P., D. L. Keairns, W. F. Kittle, A Thermogravimetric
Study of Limestone and Dolomite - The Effect of Calcination Condi-
tions, Thermochimica Acta, 14: 209; 1976.
7. Ulerich, N. H., E. P. O'Neill and D. L. Keairns, The Influence of
Limestone Calcination on the Utilization of the Sulfur-Sorbent in
120
-------
Atmospheric Pressure Fluid-Bed Combustors. Final report to EPRI,
Westinghouse Research and Development Center, Pittsburgh, PA, Con-
tract RP720-1, February 1977, EPRI FP-426.
8. Ulerich, N. H., E. P. O'Neill, and D. L. Keairns, A Thermogravimetric
Study of the Effect of Pore Volume - Pore Size Distribution on the
Sulfation of Calcined Limestones, Thermochimica Acta, 26; 269-82;
1978.
9. Ulerich, N. H., R. A. Newby, and D. L. Keairns, Sorbent Requirements
for a Gulf Coast Lignite-Fixed Atmospheric Fluid Bed Combustion Power
Plant. Final report to EPRI, Westinghouse Research and Development
Center, Pittsburgh, PA, Contract RP1179-1, October 1978.
10. O'Neill, E. P., N. H. Ulerich, R. A. Newby, and D. L. Keairns,
Criteria for the Selection of SO™ Solvents for Atmospheric Pressure
Fluid Bed Combustors. Final report to EPRI, Westinghouse Research
and Development Center, EPRI Contract RP721, January 1979.
11. Beecher, D. T., et al., Energy Conversion Alternatives Study (EGAS)
Vol. III. Report to NASA, Westinghouse Research Laboratories,
Pittsburgh, PA, NASA (R-134942), Washington: National Science
Foundation; 1976, NTIS PB 268558.
12. Hoke, R. C., L. A. Ruth, and H. Shaw, Combustion and Desulfurization
of Coal in a Fluidized Bed of Limestone, IEEE-ASME Joint Power
Generation Conference, Miami Beach, FL, September 15-19, 1974.
13. Moss, A., Chem. Eng. (Birmingham University), 23: 24; 1972.
14. Reduction of Atmospheric Pollution, Final Report, Vol. 2. Office
of Air Programs, National Coal Board, London, UK, September 1971,
PB 210-674.
15. Burdett, W. A., Sulphation of CaO by S02 and SO.,, Central Electricity
Generating Board, London, UK, May 1979, Note R1M/N1052.
121
-------
16. Yang, R. T. , C. R. Krishna, and M. Steinberg, Fluidized-Bed Coal
Combustion with Lime Additives. The Phenomenon of Peaking Sulfur
Retention at a Certain Temperature, Ind. Eng. Chem. , Fund am. ,
16 (4):465-467; 1977.
17. Hartman, M. and R. Coughlin, AICHE J., 22: 490; 1976.
18. Hubble, B. R. , et al The Formation of Mg3Ca (SO^)^ during the
Sulfation Reaction of Dolomite, Journal of the Air Pollution
Control Association, 27 (4): 343-346; 1977.
19. Hubble, B. R., S. Siegel, L. H. Fuchs, and P. T. Cunningham,
Chemical, Structural, and Morphological Studies of Dolomite in
Sulfation and Regeneration Reactions. Proceedings of the
4th International Conference on Fluidized-Bed Combustion
McLean, VA, December 1975.
20. Borgwardt, R. H. , Kinetics of the Reaction of S02 with Calcined
Limestone, Env. Sci. & Tech., 4 (1): 59-63; 1970.
21. Hatfield, J. D., Y. K. Kim, R. C. Mullins, and G. H. McClellan,
Investigation of the Reactivities of Limestone to Remove Sulfur
Dioxide from Flue Gas. Report to Office of Air Pollution,
Tennessee Valley Authority; 1971.
22. Coutant, R. W. , et al., Investigation of the Reactivity of Limestone
and Dolomite for Capturing SO- from Flue Gas. Report to NAPCA,
Battelle Memorial Institute, Columbus, OH, November 20, 1970,
NTIS PB 196 749.
23. Fields, R. B. and J. F. Davidson, Reaction of S0? with Limestone
in a Fluidized Bed; Estimation of Kinetic Data from a Batch Experi-
ment, Cambridge University, Cambridge, UK, paper presented at AIChE
Meeting, Miami, November 1978.
24. Blinichev, V. N., V. V. Strel'tsov, E. S. Lebedeva, An Investigation
into the Size Reduction of Granular Materials during their Processing
in Fluidized Beds, Int'l Chemical Energy 8 (4): 615-18; October 1968.
122
-------
25. Jonke, A. A. et al., Annual Report on a Development Program in
Pressurized Fluidized-Bed Combustion, Argonne National Laboratories,
Argonne, IL, July 1976, ANL/ES-CEN-1016.
26. Paige, J. I., J. W. Tron, J. H. Russell, and H. J. Kelly, Sorption
of SCL and Regeneration of Alkalized Alumina in Fluidized-Bed
Reactors. Report of Investigations 7414, U.S. Bureau of Mines,
August 1970.
27. Chemically Active Fluidized Bed Process, Monthly Technical Narra-
tive No. 20, Foster Wheeler Energy Corporation, Livingston, NJ,
January 24 - February 20, 1977, prepared March 14, 1977.
28. Craig, J. W. T. et al., Chemically Active Fluidized Bed Process
for Sulphur Removal during Gasification of Heavy Fuel Oil, Second
Phase. Report to EPA, Esso Research Centre, Abingdon, UK,
November 1973, EPA-650/2-73-039.
29. Wen, C. Y., and Yu, Y. A., Mechanics of Fluidization, Chem. Eng.
Prog. Series, 62: 100-111; 1966.
30. Kinzler, D., Exxon Research and Engineering Co; Linden, NJ,
personal communication, April 23, 1976.
31. Kutyavina, T. A., and A. P. Baskakov, Grinding of Fine Granular
Material with Fluidization, Chemistry and Technology Fuel Oils,
8 (3): 210-13; March - April 1972.
32. Jonke, A. A., A Development Program on Pressurized Fluidized-Bed
Combustion, Monthly Progress Report, Argonne National Laboratories,
Argonne, IL: 29-35; June 1976, ANL/ES-CEN-F092.
33. Stanley, D. A., L. Y. Sadler, III, D. R. Brooks, and M. A. Schwartz,
Production of Submicron Silicon Carbide Powders by Attrition Milling,
Fine Particles, Second International Conference, Boston, MA.
Princeton, NJ: Electrochemical Society, Inc.; 1974: 331-36.
34. Mathur, K. B., and N. Epstein, Developments in Spouted Technology,
Can. J. of Chem. Eng., 52 (2): 129-45.
123
-------
35. Gonzales, V., and A. R. Otero, Powder Technology, 7 (3): 137-43.
36. Curran, G. P., et al., Formal report No. 5 to EPA, Project 550,
Consolidation Coal Company, April 1977, EPA 600/7-77-031.
37. Encyclopedia of Chemical Technology, Second Edition, R. E. Kirk and
D. F. Othmer, eds., New York: Intersciences Publishers; 1963.
124
-------
APPENDIX A
SULFUR OXIDE REMOVAL DATA BASE AND MODEL
The feed ratio of calcium to sulfur required for desulfurization in
fluidized-bed combustion depends on the system design, the operating
conditions, the particular calcium-based sorbent, and the sorbent particle
size used. In assessments of the overall FBC system it has been conve-
nient in the past to select Ca/S feed ratios of about 3/1 for AFBC and
about 2/1 for PFBC as base cases; and, over the years, these estimates
have acquired a more permanent status, without reference to any partic-
ular sorbent, system design, or operating conditions. Desulfurization
in the bed, however, is essentially a competition between two processes:
once sulfur dioxide is released in the bed it may either remain in the
gas and escape or react with calcium oxide (CaO) or carbonate (CaCO.,)
in the bed to form calcium sulfate (CaSO,). The balance between these
processes can be drastically altered by enhancing the probability of
escape (shallow bed, high gas velocity) or by changing the reaction rate
of the sorbent (e.g., decreasing the sorbent particle size, raising the
bed temperature at atmospheric pressure).
In order to predict the Ca/S molar rate required for a given level
of desulfurization, a model of the desulfurization system that encom-
passes the effects of the relevant variables is required. Such a model
must succeed in correctly projecting the effects of key variables on the
system. In particular, since the following effects have been observed
in experimental combustors, it should demonstrate
• That an optimum temperature for desulfurization exists at
atmospheric pressure using sorbent particles 500 pm or larger
in size
125
-------
• That sorbent utilization improves at higher pressures
(VLOOO kPa)
• That there is no marked temperature effect at higher pressures
(VLOOO kPa).
Attempts to model fluidized-bed desulfurization to the extent of
predicting Ca/S molar feed ratios may be: a) statistical correlations
based on analysis of fluidized-bed combustion data (Babcock & Wilcox
ite
A3
Al A2
[B&W], Battelle ); b) models that attempt to use sulfation rate data
obtained in the Laboratory (Argonne National Laboratories [ANL],
Westinghouse, National Coal Board [NCB]; or c) fundamental models
in which the gas/solid reaction model is matched with the fluid dynamics
of the bed (no successful model known).
The correlation models are generally untrustworthy outside the
range of data correlated, and, since the objective is to project desulfur-
ization under novel constraints, they are not useful here. Of the other
models, the simplest and most direct is that developed at Westinghouse.
It may be noted that recent fundamental gas-solid modeling by
Hartman and Coughlin has yielded a sufficiently close prediction of
effects observed in practice to be considered as a starting point for
the development of a fundamental model: development and analysis of a
reasonable body of self-consistent rate data has not yet been reported.
THE WESTINGHOUSE MODEL
The model used here has been described previously (Appendix E of
Reference A4). The following assumptions are made in deriving the model;
1. Release of sulfur from coal as sulfur dioxide (SO,,) occurs with
equal probability at any place in the bed.
2. Sulfur dioxide released at a distance, h, below the upper bed
surface passes in plug-flow through a homogeneous bed of sorbent
of height h. If a mean reaction rate, K, can be assigned to the
absorption of gas in the bed, the fraction of the sulfur
126
-------
-Kt
retained is 1 - e , where t is the gas residence time in the
bed of height h. Integration for all points h along the bed
height yields an expression for the fractional sulfur removal,
1 —KZ
R, R=l-— (1-e ), where Z is the residence time of gas
JxZ.
entering the bed from below.
3. The mean rate constant for gas absorption in the bed, K, is
obtained at any fixed mean utilization of the sorbent from
thermogravimetric data reduced to unit gas concentration. The
rate constant for any given sorbent type and chemical form will
depend on the mean sorbent utilization, sorbent particle size,
bed temperature, bed pressure, and gas composition in the bed.
The model of sulfur generation in the bed assumes that sulfur is
released from coal with equal probability at all places in the bed.
While this is a reasonable general assumption, it implies that, for
example, 10 percent of the sulfur is released within the top 10 percent
of bed height. Assume that all the sulfur generated in the bed below the
top 10 percent is captured with 100 percent efficiency, and the overall
desulfurization target is 97 percent. Then the top 10 percent of the
bed must capture the sulfur liberated within it with at least 70 percent
efficiency, despite the small remaining bed height and effective gas
residence time.
As the emission standard becomes more stringent, the reaction
tant term, KZ, i
lowing table shows:
constant term, KZ, required between SO- and CaO increases as the fol-
Rate Constant/Rate
Constant at 80%
Effective Sulfur Dioxide (at constant gas
Removal Standard, % residence time)
80 1
90 2
95 4
97.5 8
98.75 16
127
-------
The rate constant for limestone sulfation required for varied sulfur
removal efficiencies is projected in Figure Al as a function of gas
residence time. In practice, the rate constant could be varied by chang-
ing the sorbent particle size or bed temperature. Alternatively,
bed residence time could be adjusted by varying gas velocity or bed
depth.
If all of the sulfur is actually liberated in the lower two-thirds
of the bed, then a different picture emerges. The top third of the
bed behaves only as an absorber into which SO,, is fed from the bottom.
The reaction rates required for desulfurization with this kind of sulfur
generation pattern are illustrated in Figure A2.
A3
A generalized sulfur generation model has been published by ANL,
and this could be incorporated into the model discussed here, if desired.
The choice of a particular sulfur generation pattern, however, in any
case is speculative.
Operating conditions that would disturb the uniform sulfur genera-
tion pattern are: the location of coal feed points and their number, the
coal particle size, and the extent of vertical mixing in the operating
fluidizing mode. Thus, fine particles of coal fed near the base of a
tall bed at low fluidization would probably cause disproportionate
release of SO. near the base of the bed and improve sulfur removal rela-
tive to the prediction of a uniform sulfur generation model.
TG DATA
The use of TG data to model the fluidized-bed sulfation depends on
several assumptions, which are listed below:
1. The rate-limiting process is governed by diffusion within the
sorbent itself.
2. The residence time of the sorbent in the bed does not affect
its reactivity (although the degree of sulfation does affect
reactivity).
128
-------
Curve 69*4387-A
NJ
100
90
c
o
1 80
CM
O
00
£ 70
c
o
t3
I 60
50
40
y_ Bed Height, Expanded
Gas Velocity, Interstitial
I I I I I I I I
i i i i i i 11
6 8 10 20
Sulfation Rate Constant, s
40 60 80 100
-1
200 400 600 1000
Figure Al - Required Reaction Rate as a Functicm of Sulfur Removal
Efficiency and Gas Residence Time (t)
-------
Curve
to
c
4g
*—•
03
CO
v.
£
"c
JS
oo
O
O
(O
01
45
40
35
30
^ 25
20
oj 15
10 -
5 -
Residence time t = l s
Uniform
Generation
Model
Lower Bed
Generation
Model
I
1
75
80
85
90
95 100
SCL Removal Required
Figure A2 - The Impact of Sulfur Generation Pattern and
Desulfurization Requirement on Required
Reaction Rate
130
-------
3. The fluidized-bed atmosphere surrounding the limestone particles
is oxidizing with respect to the CaS/CaSO, transition.
The implication of the first assumption is that while mass transfer
(e.g., of S02 from the gas phase to the sorbent particle) in the fluidized
bed may be much greater than in the TG apparatus, mass transfer dominates
the actual reaction rate over a small extent of reaction. Once the cal-
cium utilization has reached about 10 percent, the rate is usually dom-
inated instead by inter- and intragranular diffusion of the solid. (In
some pressurized cases comparatively low mass transfer depresses the
observed TG rate over a significant portion of the rate curve). The same
internal pore structure must be provided in the sorbent during the TG test
as would prevail during a fluidized-bed combustor operation. Thus, the
calcination conditions under which the sorbent pore structure is generated
must be carefully controlled before the TG sulfation experiment if the TG
data are to provide a valid simulation of the effect of sorbent utiliza-
tion on reaction rate.
Concerning the second assumption above, the mean residence time of
the sorbent may be much longer in a fluidized bed (up to 24 hours) than
in a TG experiment (^2 hours), unless the latter is carried out under
conditions of low-SO_ concentration (0.05%). The sorbent pore structure
may be modified by this long exposure to temperature, and, unfortunately,
this may either enhance or retard the sulfation rate, depending on the
particular sorbent. This effect is not accounted for in the TG-based
kinetic model.
It is often postulated that the existence of local reducing areas
in the bed may cause sulfur capture as CaS, which is subsequently oxidized
to sulfate. The assumption that this mechanism is unimportant in deter-
mining the overall rate of reaction is justified by the fact that oxida-
tion of calcium sulfide (CaS) in limestone is limited in the same way that
utilization of CaO is limited in sulfation. Oxidized sulfided limestone
contains an inner layer of sulfided stone which has not been oxidized.
131
-------
Since CaS is not found in the product stone from FBC, it is clear that
there is no significant additional reaction rate component from sulfida-
tion of the sorbent.
The TG data, which yield rate as a function of sulfur loading, or
sorbent utilization, vary greatly according to the sorbent properties
and combustor operating conditions. The variables that affect the reac-
tion rate and can be studied on the TGA are
1. Type of sorbent
2. Form of sorbent (e.g., limestone, lime, hydrated lime)
3. Particle size of sorbent
4. Calcination conditions (in particular the temperature and
partial pressure of CO. during calcination)
5. Residence time (time at temperature)
6. System pressure
7. System temperature
8. Percent excess air in the system
9. SCL partial pressure and other gas composition variables.
In making a projection of the effect of operating conditions on the
Ca/S molar feed ratio, it is advisable to use TG data obtained under
experimental conditions that correspond closely to those that will be
encountered in the fluidized bed.
For any set of design input:
Bed height
Gas velocity
Bed voidage
Bed emulsion and bubble phase volume
Coal:
Ash content (heating value)
Sulfur content
132
-------
the rate data can be used in the model to give Ca/S molar feed ratio
projections. A data base of over 300 atmospheric-pressure and 70 pres-
surized TG runs exists at Westinghouse over the range of conditions
listed in Tables Al and A2.
Table Al
RANGE OF ATMOSPHERIC TG SULFATIONS
Stone(s)
Ames
1359
Size,
U.S. Mesh
% Excess
Air
Temp . ,
°C
Comment
, Brownwood, 16/18 20 815 Varied
, Greer, Carbon 5-100/200
calcination
1337, Western,
Mississippi
Bellefonte
Greer
Ames
Ames, Brownwood
Greer, Carbon,
1359
Ames, 1359,
Greer, Carbon
1359 (2263)*
Carbon
Tymochtee
Canaan, Kaiser
16/18
35/40
35/40
5/6-35/40
20
20
20
20
780-950
750-940
900
850
16/18
-325
16/18
100/200
20
20
2-16% 0,
t
20
815
800-950
, 815
815
Varied calcination
Calcined at 900°C
in 60% CO^
Scattered particle
sizes
Calcined at 900°C
in 60% CO.,
*Residence time varied (0.1-0.5% S0_ in sulfating atmosphere)
133
-------
Table A2
RANGE OF PRESSURIZED TG SULFATIONS 1013 kPa (10 atm)
Stone
Size,
U.S. Mesh
% Excess
Air
Temp . ,
°C
Comment
Canaan
1337
1337
Greer
Greer
Greer
Greer
Greer
Greer*
Tymochtee
1359 (2263)*
35/40
35/40
8/10
35/40
6/8
40/100
100/200
16/18
14/18
16/18
16/18
300
300
20
300
300
300
300
15-100
20
Calc. 815
(0.15 atm
C02) 0.75-
16% 02)
20
843
843-1010
900
843-1010
900-1010
900-1010
950-1000
815
815
815
815
Half calcined
Carbonated
FB Calcine 815, 15%
co2
*Residence time varied (0.05-0.5% S02 in sulfating atm),
PROJECTION METHOD
Sorbent feed projections are obtained using TG rate data. The rate
data are fitted with polynomial equations for use in the computer-
generated projections.
The average rate constant for SO- sorption in the bed, K, that is
needed to maintain a given level of sulfur retention, R, is calculated
at a defined gas residence time (Z - expanded bed height
the equation
.. v
interstitial gas velocity' Usin8
R - 1 -
KZ
,, -KZ,
(1 - e )
134
-------
Note that K is related to the first-order rate constant for the reaction
CaO + S02 + 1/2 02 -> CaS04, K , by
~ 6 + (1 - 6)e
where
K1 = f(a)
6 = volume fraction of bed bubbles
e = bed voidage in emulsion phase
F = fraction of emulsion volume occupied by inerts.
The sorbent utilization, a, at which the reaction rate constant, K ,
applies is determined from TG data. A mass balance on the TG system
gives the molar rate of reaction, -r— , in terms of the reaction note con-
stant as
da K* C £TG
dt P(l - ETG)
where
p = solids density, mole Ca/cc
C = mole S0»/cc in TG reaction gas.
The polynomial equations representing TG rate data a = f(Tjr) are then
used to calculate the sorbent utilization. The required Ca/S molar feed
ratio is then defined by
Ca/S = | .
CONFIRMATION OF THE MODEL
The model has been used previously to show that TG data accurately
demonstrate important features of desulfurization phenomena in fluidized
beds. Desulfurization phenomena that have been observed in fluidized
A4
beds and demonstrated on the TG include:
• The occurrence of an optimum temperature for desulfurization in
AFBC with sorbent particle sizes of greater than 500 urn
• No marked temperature effect at higher pressures (^1000 kPa)
135
-------
• Improved sorbent utilization at higher pressures (^1000 kPa)
• Improved sorbent utilization with precalcination.
The specific data that the TG has been used to model include data from
the NCB, ANL, B&W, Pope, Evans and Robbins (PER), and Westinghouse.
Model projections of the Ca/S molar feed ratios required for vari-
ous levels of desulfurization in AFBC, as a function of limestone type,
are compared to the data collected from the ANL and British Coal Research
(BCR) fluidized-bed units for limestone 1359A7 in Figure A3. Conditions
for the fluidized-bed runs were:
101.3 kPa (1 atm), Limestone 1359
490-630 ym particles in the feed
0.79-0.85 m/s (2.6-2.8 ft/sec) velocity
788-798°C
0.61 m (2 ft) bed height
3% 02> 15% CO- in the flue gas
To obtain the projections, TG rate data from sulfation at 815°C in 0.5%
S02, 4% oxygen, and nitrogen were utilized. The sulfations were carried
out with 420-to-500-ym particles of limestone, calcined at 815°C in 15%
CO- and nitrogen. The gas residence tiiae (as determined by input bed
height and velocity) was 0.66 s. ANL operated with s. gas residence time
of 0.74 s. This longer residence time may account for the slightly
lower Ca/S molar ratio requirements in the ANL 1359 data.
An example of the TG projection of pressurized results obtained in
the Exxon miniplant is shown in Figure A4. Projections of the desulfur-
ization obtainable at varied Ca/S molar feed rates using Dolomite 1337
A7
are compared to results obtained in Run 27 of the miniplant. More
A9
recent results from Runs 68-73 designed to investigate higher levels
of sulfur removal could also be projected with accuracy (Figure A5).
AMENDMENTS TO THE MODEL
Future amendments to be made on the model include a) integration of
sorbent particle size distribution, b) including the impact of attrition
136
-------
Curve 691792-A
o
t>
3
TS
as
"c
OJ
Co
Fluid-Bed
Operating Conditions:
latm, 101. 3kPa
420- 300 \un limestone particles
Bed Height 4 ft
Velocity 6 ft/s
815 °C
~ 20* excess air
Carbon Limestone
Greer Limestone
Limestone 1359
ANL best fit of data collected
for Limestone 1359 (1971)
Curve 712993-A
100
90
2
"Ej
oo
70
Dolomite 1337
950°C, U%C02, 3.5%02
2.9 s residence time (expanded bed height/
superficial velocity)
1000 -1190 urn particles ITG Run P99)
2000 - 2380 urn particles (TG Run p 104)
Exxon Results (Run271
829-931 °C
2.9-3.5s residence time
13-17%C02 in flue gas
1.5-4%02in flue gas
Dolomite 1337 (840 -2380 Mm)
1234567
Ca/ S Molar Ratio
Figure A3 - Predictions of the Ca/S Feed Ratios Required
for Desulfurization Using Westinghouse TG
Data for Three Sorbents
0.2 0.6
1.0 1.4 1.8
Ca/S Molar Feed
2.2 2.6 3.0
Figure A4 - Comparison of Pressurized
TG Projections with Data
from the Exxon Miniplant
-------
Curve 690412-A
o
"c
ce.
100
u>
oo
t 95
13
LO
«P
90
85
60
Curve
TG Projections
Dolomite 1337
950°C. 14% C02
2.2 s residence time
420-500 urn (TG Run P 96)
1000-1190 urn (TG Run P99)
,2. 3.5% 02
Exxon Results
/ Dolomite 1337, (840-2380um)"
933-947 °C
1.6-2.2 s residence time
16-22% 0)2 in flue gas
2 -4%09 in flue gas
J I I L
j _ I
i
j L
t>3
CO
•s
o>
o>
.0
\—
o
•s
o
75
I I
Lowellville Limestone
TG Data
Calcined Nonisothermally upto815°C m 15% C0
Sulfated at 815 °C in 4% 0> 0.5% S0
Bed Temperature 774-858°C
3% S Coal
3-5.3% Excess 0
12-14%C02mFlueGas
i 30 -
20 -
10 -
.2 .6 1.0 1.4 1.8
Ca/S Molar Feed Rate
2.2
2.6
3.0
10 20 30 40 50
Utilization Calculated from TG Data (0.1% Ca/min Rate Criterion)
Using B&W Feed Particle Size Distribution
Figure A5 - Comparison of TG Projections for
Greater than 90% Sulfur Removal with
Exxon Data for Dolomite 1337
Figure A6 - Comparison of Sorbent Capacity
Obtained from Westinghouse TG Data
with B&W Fluid-Bed ResultsA10
-------
and elutriation on particle size distribution, and c) discerning the
effect of different mass transfer rates on the TG and in fluidized beds.
Instead of using a TG curve based on one particle size representing
the estimated average particle size in the bed, it would be preferable
to use a composite TG curve representing the sorbent particle size dis-
tribution within the bed. Westinghouse has already calculated such a
composite curve for one particular case - the atmospheric-pressure
A10
experiments of B&W, with excellent results.
The Westinghouse TG data for different sizes of Carbon limestone
were combined to derive a projected composite utilization for the size
consistency in the bed of the Babcock and Wilcox 3 ft by 3 ft (0.9 m by
0.9 m) AFBC unit. The average operating parameters of 20 B&W runs (8 ft
[2.4 m] bed, 1.5 ft [0.5 m]/s, 50% sulfur removal) using Carbon limestone
were used to estimate a molar reaction rate on the TGA (0.1% Ca/min)
that corresponds to the average bed rate constant. A weighted average,
based on the B&W sorbent feed size distributions, of the sorbent utiliza-
tion obtained on the TGA at 0.1 percent calcium reacting per minute was
calculated for TG runs on finely divided size fractions of Carbon lime-
stone at temperatures and gas compositions similar to those in the B&W
combustor. The results of this comparison are shown in Figure A6.
Westinghouse feels that the agreement is excellent. This technique to
account for particle size distributions is time consuming, however, for
many TG experiments are required to cover the full range of particle
sizes. A fundamental model that would permit the construction of the
TG curves for different particle size ranges from a base curve would be
of inestimable value.
As another amendment, a more fundamental correction to the model
would result from generation of a mean rate value for the bed given the
residence time distribution of the sorbent in the bed. This mean rate
would consider particle attrition within the bed and elutriation from
the bed.
The third required amendment would adjust the initial rates of
reaction to account for mass transfer effects in the fluidized bed. To
139
-------
some extent this effect can be calculated from batch fluidized-bed
results. Mass transfer should only be significant in cases for sorbent
utilization of 10 percent or less, corresponding to Ca/S molar feed
ratios of 10/1 or higher. Such high sorbent feed rates lie outside the
range of practical interest, except for generative systems.
CONCLUSION
Thermogravimetric rate data can be successfully used to determine
the rate constant of sulfation as a function of sorbent utilization for
calcium-based sorbents. By judiciously selecting operating conditions
that represent conditions in fluidized-bed combustion, the rate constant
can be used to predict sulfur retention in fluidized-bed units. The
agreement between fluidized-bed data and TG projections has been demon-
strated using data collected at 1013 kPa (10 atm) pressure, as well as
at atmospheric pressure.
The TG projections are limited by the availability of complete
pilot plant data (particle size distribution in the bed, fraction of
inert particles in the bed, bed expansion data), the accuracy of pilot
plant data (including fluctuations in coal and sorbent properties and
nonsteady-state operation), the representability of the 20 mg sample
used in the TGA of the bulk limestone, as well as the basic assumptions
applied in the projections. The modeling assumptions and the limita-
tions implied by each are outlined in Table A3.
140
-------
Table A3
LIMITATIONS OF TG PROJECTIONS
Model
Assumption Limitation Comment
Diffusion Control Mass transfer may
influence initial
reaction rate
First-Order Reaction No account for
sorbent sintering
Uniform Sulfur No variation in
Generation sulfur generation
pattern is
accounted for
All
Generation and Sorption • Ash can absorb Has been included
of Sulfur All Occur in sulfur in the model
Main Bed by Limestone
• Sulfur can be
generated and
absorbed outside
of bed
141
-------
REFERENCES
Al Attig, R. C., et al., Additive Injection for Sulfur Dioxide Control,
A Pilot Plant Study, APTP-1176, The Babcock and Wilcox Company,
Alliance, OH, 1970, NTIS PB 226 761.
A2 Liu, C Y., Personal Communication.
A3 Jonke, A. A., et al., Reduction of Atmospheric Pollution by the
Application of Fluidized-Bed Combustion, ANL ES-CEN-1002, Argonne
National Laboratory, Argonne, IL, 1970.
A4 Keairns, D. L., et al., Fluidized-Bed Combustion Process Evaluation -
Phase II - Pressurized Fluidized-Bed Coal Combustion Development.
Report to EPA, EPA-650/2-75-027c, Westinghouse Research Laboratories,
Pittsburgh, PA, 1975, NTIS PB 246 116.
A5 Bethell, F. U., D. W. Gill, and B. B. Morgan, Mathematical Modeling
of the Limestone-Sulfur Dioxide Reaction in a Fluidized Bed
Combustor, Fuel 52; 1973.
A6 Hartmen, M., and R. W. Coughlin, Reaction of Sulfur Dioxide with
Limestone and the Grain Model, AIChE J., 22(5); 1976.
A7 Jonke, A. A., et al., Reduction of Atmospheric Pollution by the
Application of Fluidized-Bed Combustion, ANL ES-CEN-1004, Argonne
National Laboratory, Argonne, IL, 1971.
A8 Hoke, R. C., et al, Studies of the Pressurized Fluidized-Bed Coal
Combustion Process, Exxon Research and Engineering Co., Linden, NJ,
September 1977, EPA-600/7-77-107.
A9 Hoke, R. C., et al, A Regenerative Limestone Process for Fluidized
Bed Coal Combustion and Desulfurization, Monthly Report Nos. 97-99
to EPA, Exxon Research and Engineering Co., Linden, NJ, March-May
1978.
142
-------
A10 Lange, H. B., and C. L. Chen., S02 Absorption in Fluidized Bed
Combustion of Coal-Effect of Limestone Particle Size, EPRI
FP-667, The Babcock & Wilcox Company, Alliance, OH, 1978.
All Ulerich, N. H., R. A. Newby, and D. L. Keairns, Sorbent Requirements
for a Gulf Coast Lignite-Fired Atmospheric Fluid Bed Combustion Power
Plant. Final report to EPRI, Westinghouse Research and Development
Center, Pittsburgh, PA, Contract RP1179-1, October 1978.
143
-------
NOMENCLATURE
FBC Fluid-bed combustion
AFBC Atmospheric fluid-bed combustion
PFBC Pressurized fluid-bed combustion
TGA Thermogravimetric analysis
TG Thermogravimetric
B&W Babcock and Wilcox
ANL Argonne National Laboratories
NCB National Coal Board
PER Pope, Evans and Robbins
h bed height, expanded
R Sulfur removal, fractional
Z gas residence time, expanded bed height/interstitial gas velocity
K average rate constant for SO sorption in the bed
K first-order rate constant for the reaction
CaO + S02 + 1/2 02 —+• CaSO^
f, volume fraction of bed bubbles
e bed voidage in emulsion phase
F fraction of emulsion volume occupied by inerts
o sorbent utilization, fractional
C mole SO./cc in TG reaction gas
P solids density, mole Ca/cc
CTG TG voidage
d particle diameter
144
-------
APPENDIX B
SORBENT INFORMATION AND TG RATE DATA
145
-------
Set 1
The Effect of Temperature on Pressurized
Limestone Sulfation
146
-------
.2 •
.1
«*
RUN »P20 TCA tl
GREEK LIMESTONE, 2380/3560 u«
HEATED 1O C/MIN TO REACTION TEMPERATURE
CALCINED AT 894 C IN 4.31 C02; 1S.SZ 02; N2 AT
SULFATED AT 894 C , 1 L/MIN AT lOatn
IN 0.381 S02; 4.3* CO2; 15.8* O2; N2
.3
Is too
TIME/MINUTES
ISO
.1
RUM IP22 TGA II
GREER LIMESTONE, 2330/3360 un
HEATED 10 C/MtN TO REACTION TEMPERATURE
CALCINED AT 1010C IN 4.31 C02; 15.81 02;
SULFATED AT 1010C , 1 L/MIN AT lOatn
IN 0.38Z S02; 4.3Z C02; 15.81 02; N2
N2 AT lOatm
100
-125
L50
TIME/MINUTES
.3
.2
.1
.*•***
/ RUN »P21
* CREER LIMESTONE
/ HEATED 10 C/MI.1
* CALCINED AT 954
* SULFATED AT 954
f IN 0.331 S02;
* *
*
*
TCA 11
, 2380/3360 un
TO REACTION TEMPERATURE
C IN 4.3Z C02; 15. 8* 02; N2 AT lOat
C , 1 L/MIN AT lOatn
4.3Z C02; 15.81 O2 ; N2
25 50 75 100 125 150 1?5
TIHE/HIHUTL'S
s
H
H
u
.1
RUN *P23 TGA II
CREER LIMESTONE, 420/500 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 1010C IN 4.3Z CO2; 15.BZ 02; N2 AT lOati
SULFATED AT loioc , i L/MIN AT I0at«
IN 0.38Z S02; 4.3Z C02; 15.8Z 02; N2
"to to to
.. tr .
TIME/MINUTES
too
tfo teo
-------
.6
.5
.4
.2
.1
RUN IP24 TGA M
CREER LIMESTONE, 149/420 u»
SEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 899 C IN 4.31 CO2; 15.8X 02; N2 AT lOatB
SULFATED AT 899 C . 1 L/MIN AT lOatm
IK 0.381 S02; 4.31 CO2; 1S.8Z 02; N2
.7
.5
.4
.3
.1
"toJo to to to to Jo Jo Jo too
TINE/MINUTES
RUN »P26 TGA '1
GREEK LIMESTONE, 149/400
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 1010C IN 4.31 C02; 15.8Z 02; N2 AT lOatn
SULFATED AT 1010C , 1 L/MIN AT lOatm
IN 0.38% S02; 4.3Z C02; 15.81 O2; N2
"to to to to too tilt*o t
TIME/MINUTES
60
OO
.7
.6
.5
w
z
2 .3
H
O
.2
.1
,* *
RON IP25 TCA »1
CREER LIMESTONE, 149/420 u«
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 954 C IN 4.31 CO2 ; 15. 8Z O2 ; N2
SULFATED AT 954 C . 1 L/MIN AT lOatn
III 0.38Z S02; 4.3Z CO2 ; 15.81 O2 ; N2
S
lOatn
.1
"to to to Jo too t3o t*o
TIME/MINUTES
160
..'**
RUN fP27 TGA II
GREER LIMESTONE, 2380/3360 u«
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 1010C IN 4.35 CO2; 15.8» O2; N2 AT lOatn
SULFATED AT 1010C . 1 L/MIN AT lOatm
IN 0.38: S02; 4.3Z C02; 15.8t O2; N2
"IT
So
TIMI/MIIDTES
"to too tzo
-------
RUN IP49 TGA »1
CREER LIMESTONE, 74/149 un
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 950 C IN 4.3* C02 ; 1S.8Z 02; N2 AT lOatn
SULFATED AT 950 C , 1 L/MIN AT lOatm
IN 0.381 S02; 4.3Z C02; 15. 81 02;
N2
io
TIME/MINUTES
loo
.5
.*
.3
.2
.1
-t20
X""*
RUN 176-122 TGA ifl
CREER LIMESTONE, 420/500
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 843 C IN 4.3* C02; 15. SZ 02; N2 AT lOatm
SULFATED AT 843 C , 2 L/MIN AT lOatm
IN 0.33J 502; 4.3Z C02; 15.81 02; N2
-to"
to 4o
TIME/MINUTES
loo
120
o
M
H
H
U
.6
* *
o
M
I-
RUN IP52 TGA tl
CREER LIMESTONE, 74/149 un
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 1000C IN 4.3Z C02; 15.SZ O2; N2 AT lOatn
SULFATED AT 1000C , 1 L/MIH AT lOatn
IN 0.38J S02; 4.3% C02; 15.8* 02; N2
"to fo Jo to to to to to iotoo
TIME/MINUTES
/ RUN #76-123 TGA fl
« CREER LIMESTONE, 420/500
* HEATED 10 C/MIN TO REACTION TEMPERATURE
* CALCINED AT 896 C IN 4.3JS CO2 ; 15.8Z O2 ; N2 AT lOatn
, SULFATED AT 896 C , 2 L/MIH AT lOatm
IN 0.18Z S02; 4.3Z CO2; 15,flZ 02; N2
"I lo is JO J5 30 35 io *5 50 55 60
TIHE/MINOTES
-------
a
H
H
* *
RUN *76-124 TCA II
GREEK LIMESTONE, 420/500
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 866 C IN 4.JZ CO2 ; 15.81 O2 ; N2 AT lOatl
SOLFATED AT 866 C , 2 L/NIH AT lOatm
IN 0.581 S02; 4.31 CO2 ; 15.81 02; N2
•to"
•tr
120
TIME/MINUTES
Ul
O
.6
.5
.1
0
RUN 176-125 TGA <1
GREEK LIMESTONE, 420/500
SEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 955 C IN 4.31 C02; 15.8* 02; N2 AT lOatn
SULPATED AT 955 C , 2 L/MIN AT 10«tm
III 0.38T S02; 4.31 C02; 15.SI 02; N2
"to
TIME/MIMUTES
leo
"foo
.5
.4
.2
.1
RUN 176-126 TGA »1
GREER LIMESTONE. 420/500
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 928 C IN 4.3Z CO2; 15.81 02; N2 AT lOati
SULFATED AT 928 C , 2 L/MIN
IN 0.38Z S02; 4.31 C02; 15.81 02; N2
loo
120
TIME/MINUTES
J
S
.4
.3
.2
.1
.* *'
.* *
RUN 176-127 TGA fl
GREEK LIMESTONE. 420/500
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 978 C IN 4.3X C02; IS.81 02; N2 AT lOatn
SULFATED AT 978 C , 2 L/MIN AT 10at>
IN 0.381 S02; 4.3Z CO2; 15.81 02; N2
"to to
TIME/MIHDTIS
loo tzo 1*0
-------
.6
.5
a
u
< •*
•4
FRACTION SU
K> U»
. 1
0
1
* *
*
*
*
*
*
*
*
*
** RUN #76-128 TGA 11
* GREEK LIMESTONE, 420/500
* HEATED 10 C/MIN TO REACTION TEMPERATURE
* CALCINED AT 980 C IN 4,31 C02 ; 15. 8Z 02; N2 AT lOat
* SULFATED AT 980 C , 2 L/MIN AT lOattt
* IN 0.38* S02; 4.3X C02 ; 15.81 02; N2
*
*
*
*
*
*
*
*
k
0 20 30 40 50 60 ?0 60
TIME/MIJIUTES
-------
Set 2
The Effect of Temperature on Pressurized
Dolomite Sulfation
152
-------
.6
.5
.4
.3
Z .2
.1
x"1
RUN »P7 TRA II
DOLOMITE 1337, 420/500 um
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 843 C IN 4.3* CO2; 15. S? 02; N2 AT lOatn,
SIII.FATED AT 84 3 C . 1 L/MIN AT lOatm
IN 0.38Z S02; 4.31 C02; 15.9% 02; N2
-to
.7
.6
.5
.4
-i .3
.2
.1
RUN »P10 TRA »1
DOLOMITE 1337, 420/500 um
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 871 C IN 4.31 C02; 15.81 02;
SULFATED AT 871 C , 1 L/MIN AT lOatm
IN 0.38* S02; 4.3* C02; 15.8% 02; N2
N2 AT lOatn
•*T
•to
CO
TIME/MINUTES
TIME/MINUTES
. 7
.6
.5
.4
.3
.2
.1
RUN IPS TCA II
DOLOMITE 1337, 420/500 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 927 C IN 4.3* C02; 15.81 O2; N2 AT 10«t«
SULFATED AT 927 C , 1 L/MIN AT lOatu
IN 0.38J S02; 4.31 C02; 15.81 O2; N2
"iO 80 100
TIME/MINUTES
120 140 160 1
80
. 7
.6
.5
.4
. 3
.2
.1
RUN IP11 TOA 01
DOLOMITE 1317. 420/500 un
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 982 C IM 4.351 CO2 ; 15. 82 02;
SULFATED AT 982 C , 1 L/MIN AT lOatm
IN 0.387 S02; 4.3Z CO2; 15.8Z 02; H2
N2 AT 10,-itn
-tr
-tb-
tb-
Too
izu
TIME/MINUTES
-------
RUN »P13 TCA II
DOLOMITE 1337, 420/500 u«
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 982 C IN 4.3Z C02; 15.8Z 02; N2 AT 10at»
SULFATED AT 982 C , 1 L/MIN AT lOatm
IN 0.38Z S02; 4.31 C02; 15.8X O2;
N2
to to to~
"tootlo tlfl 160 180
RUN IP 15 TGA II
DOLOMITE 1337, 420/500 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 954 C IN 4.3* C02; 15.8Z 02; N2 AT lOatm
SULFATED AT 954 C , 1 L/MIN AT lOatm
IN 0.38Z S02; 4.3Z C02; 15.8* O2; N2
to
TIME/MINUTES
too tsl
TIME/MINUTES
200
.8
.7
.6
riON SULFATED
• * •
Ut *• Ui
0.2
«
h
.1
0 I
*
*
*
*
*
Ik Rl'N IP14 TCA 11
* DOLOMITE 1337, 420/500 uo>
* HEATED 10 C/MIN TO REACTION TEMPERATURE
* CALCINED AT 982 C IN 4.3* C02; 15.8* 02; N2
* SULPATED AT 982 C , 1 L/MIN AT lOatn
t IN 0.38T. S02; 4.3Z C02 ; 15.81 02; N2
20 tO tO iO 100 120 140 160
.6
.4
AT lOatm
O
(H
H
O
.2
.1
,.* *'
RUM IP16 TGA II
DOLOMITED 1337, 420/500 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 899 C IN 4.3Z C02; 15.81 02; N2 AT lOi
SULFATED AT 899 C , 1 L/MIN AT 10«t«
IN 0.38: S02; 4.3Z C02; 15.8Z 02; N2
TIME/MINUTES
—to te-
TIME/MIHUTES
loo tzo
-------
.7
.6
.5
.4
« .3
RUN f?17 TCA 11
DOLOMITED 1337, 420/500 u«
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 927 C IN 4.3* C02 ; 15.8X 02; N2 AT
SOLFATED AT 927 C ; 1 L/MIN AT lOatm
IN 0.38Z S02; 4.31 C02j 15.81 02; N2
.2
.1
(Ji
Ui
"tototoSo"
toot20
TIME/MINUTES
RUN IP18 TCA II
DOLOMITED 1337, 420/500 um
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 1000C IN 4.3Z C02; 15.8Z 02; N2 AT 10at«
SULFATED AT 1000C , 1 L/MIN AT 10«tm
IM 0.38: S02; 4.3Z C02; 15.8* 02; H2
to to too"
TIME/MINUTES
20
160
-------
Set 3
Effect of 0£ Concentration on
Atmospheric-Pressure Desulfurization of Dolomite
156
-------
RUN i»333 TCA #2
TYMOCHTEE DOLOMITE. 1000/1190 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 14.3* C02; N2; 0.2 L/MIN.
SOLFATED AT 815 C , 0.6 L/MIN
IN 0.5Z S02; 16Z 02; N2
tooIso
TIME/MINUTES
"ISO
.1
RUN 1334 TGA II
TYMOCHTEE DOLOMITE, 1000/1190 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 14. 3* CO2 ; N2; 0.2 7,/MIN.
SULFATED AT 815 C , 0.6 L/MIN
IN 0.5Z S02; 10.5t 02; N2
f&—fro
"to to Ioo 120 t«o t?o tso Joo
TIME/MINUTES
.9
.8
***•
,**'
.7
.6
.5
.4
.3
.2
.1
RUN 1335 TGA 12
TYMOCHTEE DOLOMITE, 1000/1190 un
HEATED 10 C/MIH TO REACTION TEMPERATURE
CALCINED AT 815 C IN 14.3Z C02; N2 ; 0.2 L/MIN.
SULFATED AT 815 C , 0.6 L/MIN
IN 0.5Z S02; 21 02; N2
-to"
"too"
-155-
ioo
TIME/MINUTES
.* '
RUN »336 TCA 12
TYMOCHTEE DOLOMITE, 1000/1190 un
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 81S C IN 14.31 C02; N2; AT 0.2 L/MIN.
SULFATED AT 815 C , 0.6 L/MIN
IN 0.5% S02; UZ 02; N2
-155-
TIME/MINUTES
200
-------
CO
RUN 1148 TCA 12
TYHOCHTEE DOLOMITE. 1000/1190 un
HEATED 10 C/MIH TO REACTION TEMPERATURE
CALCINED AT 81S C IN 14.31 CO?; N2
SULFATED AT 815 C , 0.6 L/MIN
IN O.SZ S02; 7* 02; N2
ft to"
"toSoloo
TIME/MINUTES
1201*0160
-------
Set 4
Effect of C>2 Concentration on
Pressurized Desulfurization of Dolomite
159
-------
.**•
! RUN »P46 TGA *1
TYMOCHTEE DOLOMITE, 1000/1190 un
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 1.5T C02; N2 AT lOa
SULFATED AT 815 C , 1 L/MIN AT lOatm
IN 0.5T. S02; 7X 02; N2
20
30
TIME/MINUTES
to
RUN »P71 TCA *1
TYMOCHTEE DOLOMITE, 1000/1190 un
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN l.SX CO2; N2 AT lOatm
SULFATED AT 815 C , 1 L/MIN AT lOattn
IN 0.5Z S02; 21 02; N2
"to to to"
TIME/MINUTES
100
120
RUN fP72 TGA II
TYMOCHTEE DOLOMITE, 1000/1190 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 1. 5 J C02: 112 AT lOatn
SULFATED AT 815 C , 1 L/MIN AT lOatn
IN 0.51 S02; 0.751 02; N2
to Jo io to to
TIME/MINUTES
"to to to Jo
RUN JP76 TCA *1
TYMOCHTEH DOLOMITE, 1000/1190 um
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 1.5Z C02; N2 AT lOatm
SULFATED AT 815 C , 1 L/MIN AT lOatn
IN 0.5X S02; 161 02; N2
to to to Jo to
to Jo to too
TZHE/NIHDTES
-------
RUN *P79 TGA 11
TYtlOCHTEE DOLOMITE, 1000/1190 un
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 1.57 C02; N2 AT lOatm
SULFATED AT 815 C , 1 L/MIN AT lOatm
IN 0.5Z 502; 10.531 O2; N2
to to Jo to to
TIME/MINUTES
~ioto So to
-------
Set 5
Effect of Excess Air on
Uncalcined Limestone Sulfation
162
-------
.9
.8
.7
o
C ..
z
2 .«
fr4
U
£ .3
.2
.1
0 •
RUN IP28 TKA l»l
GREER LIMESTONE, 1000/1190 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 8.7* CO2; 10.51 02; N2
SULFATED AT 815 C , 1 L/MIN AT lOatm
TN 0.5Z S02; 8.71 C02; 10.5Z 02; N2
.3
U>
TIME/MINUTES
RUN IP29 TCA M
GREER LIMESTONE, 1000/1190 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED1 AT 815 C IN 15Z C02; 2.71 02; N2 AT lOato
SULPATED AT 815 C , 1 L/MIN AT lOatn
IN 0.5Z S02; 151 C02; 2.77. O2: !I2
.2
.1
.**•*
* *
,****'
RUN *"30 TCA *1
CREER LIMESTONE, 1000/1190 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 5.8t C02; 141 02; N2 AT lOatm
SULFATED AT 815 C , 1 L/MIN AT lOacm
IN 0.5Z S02; 5.8Z C02; 14Z 02; N2
-tr
•*r
-fo
TIME/MINUTES
TIME/MINUTES
-------
Set 6
Effect of Sorbent Residence Time on Desulfurization
164
-------
.8
.7
.6
o
w
H
<
ft. .5
j
a
VI
z .4
o
-3
.2
.1
.**
RUN »P54 TGA tl
GREER LIMESTONE; E-0-10, 1000/1190 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 15X C02; N2 AT lOatm
SULFATED AT 815 C . 1 L/MIN AT lOatm
IN 0.5Z S02; 4J O2 ; N2
to to toSotoo
TIME/MINUTES
.3
.2
.1
RUN *P58 TGA tl
GREER LIMESTONE; E-0-10, 1000/1190 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 151 C02; N2 AT lOatn
SULFATED AT 815 C , 1 L/MIN AT lOatm
IN 0.51 S02; 4Z 02; M2
10 20 JO 40 JO
TIME/MINUTES
o
o
O o
z
o
M
H
.6
.5
.4
.3
.1
RUN »P57 TCA »1
CREER LIMESTONE; E-O-10, 1000/1HO um
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 15* C02; N2 AT lOatin
SULFATED AT 815 C , 1 L/MIN AT lOatm
IN 0.0961 S02; 4Z 02; !J2
"Jo to to So too tzo tlo teo
TIME/MINUTES
, 7
.6
.5
2 .4
H
. 3
RUN »P6<) TCA Jl
GREER LIMESTONE; E-O-lf), 1000/1190 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IS 15* C02; N2 AT lOatm
SULFATED AT 815 C , 1 L/MIN AT lOatm
IN 1.53: S02; 4* 02; N2
to to to to too t75 tlo
TIME/MINUTES
tao J
oo
-------
.2
o
u
H
o
M
H
..***
.1
RUM JP64 TGA II
LIMESTONE 1359; STOHE 2263, 1000/1190 am
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 1.51 CO2; K2 AT lOatn
SDLFATED AT 815 C , 1 L/MIH AT 10at«
IN 0.51 S02; 41 02; N2
RUN »P68 TGA II
LIMESTONE 1359; STONE 2263, 1000/1190 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 151 C02; N2 AT lOatn
SULFATED AT 815 C , 1 L/MItl AT lOatn
IN 0.096Z S02; 4t 02; N2
„***
to"
~}o So
TIME/MINUTES
to to Jo to to So to"
•to—t
00
TIME/MINUTES
.3
.2
•J
9
.1
.1
RUN »P65 TCA tl
LIMESTONE 1359; STONE 2263. 1000/1190 un
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT fl!5 C IN 1.5X C02; N2 AT 10 a tin
SULFATED AT 815 C . 1 L/MIN AT lOatn
IN 0.3! S02; 41 02; N2
.*****'
,«•*
ToJo
TIME/MINOTES
•lo
RUN *P69 TCA |»1
LIMESTONE 1359; STONE 2263, 1000/1190 un
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 1.5Z CO2; N2 AT lOatn
SULFATED AT R15 C , 1 L/MIN AT lOatm
IN 0.05Z S02; 4% O2; N2
-Jo
to"
"too{20
TIME/MINUTES
-------
.7
.6
.5
* *
.2
RUN 1339 TGA 12
GREER LIMESTONE; E-O-1O, 1000/1190 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 15Z C02; N2
SULFATED AT 815 C , 0.6 L/MIN
IN 0.5Z S02; 4t 02; N2
.1
"to to Jo to to to"
TIME/MINUTES
•to—Jo—$o—t
00
, **
RUN I486 TCA 1)2
GREER LIMESTONE; E-0-10, 1000/1190 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 15* CO2; N2
SULFATED AT 815 C , 0.6 L/MIN
HI O.lt S02; 4Z 02: N2
-to
tlo(to
240
TIME/MINUTES
.6
.5
.4
.3
.2
.1
>**
RUN 0487 TCA »2
GREER LIMESTONE; E-0-10, 1000/1190 un
HEATED 10 C/KIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 15Z C.02 ; N2
SULFATED AT 815 C , 0.6 L/MIN
IN 0.05Z S02; 47 02; N2
o
TIME/MINUTES
~teo ioo
.2
a
M
I"
.1
. * *
RUK *52« TCA n
LIMESTONE 1359; STONE 2263, 1000/1190 um
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN K.5Z CO2; N2
SULFATED AT 815 C , O.f L/MIN
IN 0.5* S02; 47 02; N2
TIME/MINUTES
-------
.6
.5
•«
2 .3
.2
RUN 1536 TCA 12
CREER LIMESTONE: E-0-10, 1000/1190 un
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 151 C02; N2
SUIFATED AT 815 C , 0.6 L/MIN
IN 0.31 SO2; *l O2; N2
.1
"Jo to to to too tto t*o teo
TIME/MINUTES
OO
.2
.J
P
.1
RUN 1569 TGA 12
LIMESTONE 1359; STONE 2263, 1000/1190 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 151 C02; N2
SULFATED AT 815 C . 0.6 L/MIN
IN 0.11 502; 4Z O2; N2
-fe-
tr
TINB/MIHDTES
-------
Set 7
Large-Grained Dolomite Performance
169
-------
.2
•J
.1
RUN IPS TGA II
CANAAN DOLOMITE, 420/500 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 843 C IN 4.31 CO2; 1S.8Z 02; N2 AT lOatm
SULFATED AT 843 C , 1 L/MIN AT lOatm
IN O.SZ S02; 4.3Z C02; 1S.8Z 02; N2
"to to Jo to So 60"
TIME/MINUTES
"toSo
VJ
o
.5
.4
.3
o
H
U
U.
.2
.1
RUN J318 TGA 12
KAISER DOLOMITE, 74/149 urn
HSATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 900 C IN 60* C02; N2; 0.2 L/MIN
SULFATED AT 815 C . 0.6 L/MIS WHIN.
IN 0.5X S02; 4* 02; N2
"to to io to to io
30 40 SO
TIME/MINUTES
RUN 1317 TCA t2
CANAAN DOLOMITE, 74/149 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 900 C IN 60* CO2; N2; 0.2 L/MIN.
SULFATED AT 815 C , 0.6 L/MIN
IN 0.5% S02; 41 02; N2
-tiT
-to
TIME/MINUTES
-------
Set 8
Data for Comparison with Batch Fluid-Bed Results
171
-------
.4
.3
.2
.1
.3
.**'
.2
RUN 1354 TCA ft
AMES LIMESTONE; NU-1, 1000/1410 am
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 15Z CO2; N2
SULFATSD AT 815 C , 0.6 L/MIN
IN 0.5* S02; 4Z 02; N2
RUN *366 TCA 12
LIMESTONE 1359; NU-5, 1000/1410 un
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 151 C02; N2
SULFATED AT 815 C . 0.6 L/MIN
IN 0.5Z S02; 4Z 02; N2
.1
to Jo"
"too £20 1*3 teo
~50
100
zso
3oo
TIME/MINUTES
TIME/MINUTES
N)
.3
.2
.1
.3
.2
*
RUN 1358 TCA *2
BROWNUOOD TEXAS LIMESTONE, 1000/1190 am
HEATED 10 C/MIS TO REACTION TEMPERATURE
CALCINED AT 815 C IN 15Z CO2; N2; 0.2 L/MIN.
SULFATED AT 815 C , 0.6 L/MIS
IN 0.5Z SOZ: *Z 02; !*2
.1
too
TIME/MINUTES
"Joo
RUN 1368 TGA 12
CARBON LIMESTONE; NU-S (GRAY PART.); 1000/1410 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 15Z C02; N2; 0.2 L/MIN.
SULFATED AT 815 C . 0.6 L/MIN
IN 0.5Z S02; 4Z 02; N2
is—Jo
too til t53 t?5 $00
TIME/MINUTES
-------
.3
P
w
H
.2
»**
.1
RDS *713 TGA *1
PENRITH LIMESTONE, 710/840 UBI
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 850 C IN 213 O2 ; N2
SULFATED AT 350 C , 0.6 L/MIN
IN 0.25Z S02; 21* 02; N2
"to ts loo trs tso iTs Joo
TIME/MINUTES
.**'
RUN *A35 TGA «2
MISSISSIPPI LIMESTONE; L-9, 1000/1410 un
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN 15* C02; N2
SULFATED AT 815 C , 0.6 L/MIN
IN 0.5Z S02; 4Z 02; N2
to to to So too tlo t«o tto tso
TINE/MINUTES
.2
tt.
rJ
.1
RUN »420 TGA *2
BELLEFOHTE LIMESTONE; L-2, 1000/1410 um
HEATF.D 1(1 C/MIN TO REACTION TEMPERATURE
CALCINED AT 815 C IN N2; 0.2 L/MIN.
SULFATED AT 815 C , 0.6 L/MIN
IN 0.53 S02; 4Z 02; N2
"too tso loo
TIME/MINUTES
ISO
JOO
-------
Set 9
Data for Projections of Pilot Plant Performance
174
-------
Ui
RUN *P96 TGA tl
DOLOMITE 1337. 420/500 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 950 C IN 14Z C02; 3.5Z 02; N2 AT lOatn
SULFATED AT 950 C , 1 L/MIN AT lOatn
IN 0.51 S02; 3.5Z 02; N2
.9
.8
.7
to lo io to t
oo
tin t
*o
TIME/MINUTES
.4
.2
.1
0
RUN JP104 TGA #1
GIBSONBURC LIMESTONE, 2000/2380 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 950 C IN 14* C02; 3.5% 02; N2 AT lOatn
SULFATED AT 950 C , 1 L/MIN AT lOatm
IN 0.5% S02; 3.5% O2; N2
RUH IF99 TCA II
GIBSONBURC LIMESTONE, 1000/1190 urn
HEATED 10 C/MIN TO REACTION TEMPERATURE
CALCINED AT 950 C IN 141 C02; 3.5Z 02; N2 AT lOatl
SULFATED AT 950 C , 1 L/MIN AT 10«tn
ID 0.51 S02; 3.51 02; N2
"to Jo Jo to to to to So fa t
TIME/MINUTES
00
•to"
80 10
[40
TIME/MINUTES
-------
Table Bl
Dwg. 8531028
SORBENT SUPPLIER INFORMATION
Sorbent
Ames Limestone
Bellefonte
Limestone
Brown wood
Limestone
Carbon Limestone
Greer Limestone
Grove Limestone
"limestone 135?"
Mississippi
Limestone
Penrith Limestone
Canaan Dolomite
1337 Dolomite
Kaiser Dolomite
Tymochtee Dolomite
Quarry/Mine
The Puskarich
Quarry
Bell Mine
White Mines Inc.
Juarrv
UwelMHt Quarry
Greer lOuarryt Mine
Stephens City Quarry
Alton Mine
Duff Quarry
Address
Puskarich Limestone Co
IBS Panda Road S £. 1801 3141
Carrollton. Ohio 4*15
ct-. Paultinlin
216-627-5«15
216-7*- 3685 (Cluarryl
Warner Company
Bellelonte. Pa. 16823
ct: Louis Yost
600 Greene
iU-355-47ol
White Mines Inc.
P. 0. Bo» 500
Brownwood. Texas !6101
cl; James Bitter
915-646-8526
Carbon Limestone Co
Lmeltvllle. Ohio 44436
ct: Mr Reed
Mr. W. S. Foster
216-536-6275
Greer Limestone Co.
Greer Building
Morgantown, W. Vs. 26505
ct; Mr. Western
3M-2%-25«
M.J. Grove Lime Co.
Division of Flintkote Co.
P. 0. Box 656
Frederick. Md. 21701
ct: M.E Barger
301-662-1181
Mississippi Lime Company
7 Alby Street
Alton. Illinois 62002
ct: jerry Lepchemke
618-465-7741
Cambridge University
Dept of Engineering
Pembroke Street
Cambridge, England CB2 JRA
ct: Blake Fields
Plizer Materials, piqments
and Metals Division
Daisy Hill-Road
P.O. BO>667
Canaan Conn. 06015
Ct: G. M. HiCkS
216-526-12
-------
Table B2
SORBENT ANALYSIS
Sorbent
Measured Grain
Size (um)
Chemical Analysis (% weight)
Ca as
CaC03
Mg as
MgC03
Al as
A1203
Fe as
Fe203
Si as
Si02
Na as
Na20
K as
K20
Ames Limestone
Beliefonte Limestone
Brownwood Limestone
Carbon Limestone
Greer Limestone
Limestone 1359
Mississippi Limestone
Penrith Limestone
Canaan Dolomite
Dolomite 1337
Kaiser Dolomite
Tymochtee Dolomite
29 + 9
330 + 180;
<50
72 + 31
47 + 11
11+2
24 + 7
^ 110
14-540 ooliths
-
400 + 75
43 + 11
600 + 180
15 + 1
84.6
95.7
95.1
90.1
67.8
96.4
98.1
96.7
55.7
53.2
55.5
51.1
2.00
0.98
1.11
1.42
2.32
1.5
0.46
2.15
40.9
46.0
43.7
42.3
6.0
0.59
0.94
2.6
9.8
<0.4
0.12
-
0.19
0.076
0.011
2.0
3.7
0.24
0.97
2.0
3.7
1.0
0.10
-
2.7
0.17
0.17
0.97
8.0
1.2
1.8
3.4
15.4
<0.1
0.47
-
0.49
0.41
0.043
3.77
0.17
0.10
0.026
0.046
0.56
0.008
0.030
-
0.024
0.043
<0.002
0.075
0.90
0.16
0.17
0.42
2.29
0.10
0.039
-
0.48
0.053
<0.004
0.68
-------
APPENDIX C
FLUIDIZED-BED DATA
179
-------
Table Cl
FLUIDIZED-BED CONDITIONS AND ASSUMPTIONS FOR CALCULATIONS
Sorbents Calcined at 815°C in 15% CC>2 for 4 hours
Sulfated at 815°C in 0.5% S02, 4% 02
p = 2.71 g/cc assumed
Run
NU5
NU6
NU7
NU8
L-6
00 L-ll
0
Limestone
1359
Carbon
Brownwood
Ames
Belief onte
Mississippi
Total Gas Flow,
cc/s
905
915
915
876
983
925
Fraction SO2 j Bed Charge,
in Feed Gas I g
j
0.0053 56.96
0.0053 64.0
0.0053 54.29
0.0053 100
0.0049 50
0.0049 38.3
Mole Ca
in Bed
0.8461
0.7427
0.8495
0.8465
0.7313
0.6373
FB Particle
Radius* cm
0.05595
0.05265
0.04715
0.04585
0.0514
0.0425
TG Run
366
231
358
354
416
435
TG Particle
Radius, cm
0.05595
0.05265
0.0548
0.0603
0.0514
0.0425
-------
RUN #NU5
S02 FRACTION OF EFFLUENT
TIME (MINUTES)
0.000000
0.000110
0.000210
0.000500
0.000790
0.000950
0.000300
0.000710
0.000190
0.000000
0.000000
0.000010
0.0000)0
0.000050
0.000020
0.000040
0.000030
0.000030
0.000050
0.000080
0.000050
0.000060
0.000070
0.000060
0.000030
0.000070
0.000070
0.000070
0.000070
0.000100
0.0000*0
0.000100
0.000130
0.000150
o.ooouo
0.0001SO
0.000120
0.000190
0.000190
0.000200
0.000210
0.000230
0.000220
0.000290
0.000290
0.000350
0.000290
0.000270
0.000300
0.000310
0.000310
0.000320
0.000390
0.000350
0.000350
0.000410
0.000420
0.000410
0.0004*0
0.0005SO
0.000510
0.0005)0
0.000560
O.OOOS90
0.000520
0.0005SO
0.000600
0.000i50
0.000620
0.000700
0.000710
0.000720
0.0007(0
0.0007«0
O.OOOISO
0.000(90
0.000950
0.000(10
0.000950
0.000990
0.000910
0.000990
0.001000
0.001020
0.001100
0.001120
0.001050
0.001130
0.001090
0.001100
0.001100
0.001150
0.00090C
0.000150
0.000110
0.000050
0.000060
0.000020
0.000010
0.000020
0.000
0.250
0.500
0.730
1.000
1.250
1.500
1.7JO
2.000
2.250
13.000
13.250
13.500
13.750
14.000
14.250
14.500
14.750
15.000
15.250
15.500
15.750
16.000
16.250
16.500
16.750
17.000
17.250
17.500
17.750
IS. 000
18.250
19.000
19.250
19.500
19.750
20.000
20.250
20.500
20.750
21.000
21.250
21.500
21.750
22.000
22.250
22.500
22.750
23.000
23.250
21.500
23.750
24.000
24.250
24.500
24.750
25.000
25.250
25.500
25.750
26.000
26.250
26.500
26.750
27.000
27.250
27.500
27.750
2(.000
2(.250
28.500
28.750
29.000
29.250
29.500
29.750
30.000
30.250
30.500
30.750
31.000
31.250
31.500
31.550
31.750
32.000
32.750
12.500
32.750
33.000
13.250
33.500
11.750
14.000
34.250
34.500
34.710
35.000
181
-------
RUN #NU6
S02 FRACTION OF EFFLUENT
TIME (MINUTES)
0.000000
0.000070
0.000300
0.000070
0.000040
0.000050
0.000030
0.000000
0.000000
0.000060
0.000060
0.000060
0.000060
0.000100
0.000060
0.000150
0.000110
0.000090
0.000090
0.000110
0.000130
0.000110
0.000140
0.000180
0.000120
0.000180
0.000270
0.000140
0.000200
0.000150
0.000180
0.000180
0.000190
0.000220
0.000210
0.000220
0.000210
0.000280
0.000260
0.000270
0.000270
0.000270
0.000300
0.000300
0.000310
0.000350
0.000350
0.000350
0.000390
0.000400
0.000410
0.000410
0.000450
0.000450
0.000500
0.000510
0.000530
0.000560
0.000590
0.000580
0.000630
0.000660
0.000660
0.000700
0.000710
0.000730
0.000750
0.000800
0.000820
0.000860
0.000900
0.000900
0.000910
0.000950
0.001010
0.000160
0.001000
0.001950
0.000470
0.000180
0.000390
0.000300
0.000100
0.000110
0.000080
0.000060
0.000050
0.000020
0.000000
0.000
0.500
0.750
1.000
1.250
1.500
1.750
2.000
13.000
13.250
13.500
13.750
14.000
14.250
14.500
14.750
15.000
15.250
15.500
15.750
16.000
16.250
16.500
16.750
17.000
17.250
17.500
17.750
18.000
18.250
18.500
18.750
19.000
19.250
19.500
19.750
20.000
20.250
20.500
20.750
21.000
21.250
21.500
21.750
22.000
22.250
22.500
22.750
23.000
23.250
23.500
23.750
24.000
24.250
24.500
24.750
25.000
25.250
25.500
25.750
26.000
26.250
26.500
26.750
27.000
27.250
27.500
27.750
28.000
28.250
28.500
28.750
29.000
29.250
29.290
29.500
30.000
30.250
30.500
30.750
31.000
31.250
31.500
31.750
32.000
32.250
32.500
32.750
0.000
182
-------
RUN #NU7
S02 FRACTION OF EFFLUENT TIME (MINUTES)
0.000000 0.000
0.000000 11.500
0.000080 13.500
0.000080 13.750
0.000090 14.000
0.000090 14.250
0.000110 14.500
0.000110 14.750
0.000110 15.000
0.000120 15.250
0.000145 15.500
0.000150 15.750
0.000160 16.000
0.000180 16.250
0.000200 16.500
0.000210 16.750
0.000230 17.000
0.000260 17.250
0.000280 17.500
0.000300 17.750
0.000320 18.000
0,000330 18.250
0.000350 18.500
0.000370 18.750
0.000390 19.000
0.000440 19.250
0.000470 19.500
0.000450 19.750
0.000540 20.000
0.000560 20.250
0.000570 20.500
0.000550 20.750
0.000630 21.000
0.000650 21.250
0.000680 21.500
0.000670 21.750
0.000710 22.000
0.000720 22.250
0.000770 22.500
0.000730 22.750
0.000770 23.000
0.000840 23.250
0.000820 23.500
0.000805 23.750
0.000910 24.000
0.000910 24.250
0.000930 24.500
0.000970 24.750
0.001000 25.000
0.001040 25.250
0.001080 25.375
0.001100 25.500
0.001110 25.750
0.001090 26.000
0.001070 26.250
0.001010 26.500
0.001000 26.000
0.000720 27.000
0.000200 27.250
0.000100 27.500
0.000080 27.750
0.000060 28.000
0.000050 28.250
0.000050 28.500
0.000020 29.000
0.000020 29.500
0.000010 30.500
0.000000 32.500
183
-------
RUN #NU8
SO FRACTION OF EFFLUENT TIME (MINUTES)
0.000000 0.000
0.000000 14.750
0.000030 15.000
0.000060 16.000
0.000110 16.250
0.000040 16.500
0.000080 16.750
0.000100 17.000
0.000130 17.250
0.000120 17.500
0.000120 17.750
0.000150 18.000
0.000300 18.250
0.000150 18.500
0.000160 18.750
0.000300 19.000
0.000230 19.250
0.000290 19.500
0.000340 19.750
0.000310 20.000
0.000310 20.250
0.000400 20.500
0.000350 20.750
0.000350 21.000
0.000500 21.250
0.000480 21.500
0.000500 21.750
0.000520 22.000
0.000520 22.250
0.000560 22.500
0.000680 22.750
0.000680 23.000
0.000620 23.250
0.000590 23.500
0.000690 23.750
0.000640 24.000
0.000650 24.250
0.000640 24.500
0.000720 24.750
0.000630 25.000
0.000720 25.250
0.000730 25.500
0.000790 25.750
0.000810 26.000
0.000810 26.250
0.000790 26.500
0.000810 26.750
0.000900 27.000
0.000930 27.250
0.000920 27.500
0.000940 27.750
0.001070 28.000
0.001010 28.250
0.001050 28.375
0.001050 28.500
0.001170 28.750
0.001010 29.000
0.001040 29.250
0.000920 29.500
0.000980 29.750
0.000970 30.000
0.000990 30.250
0.000340 30.500
0.000130 30.750
0.000070 31.000
0.000000 32.000
0.000000 0.000
184
-------
RUN #L6
S02 FRACTION OF EFFLUENT TIME (MINUTES)
0.000000 0.000
0.000000 0.125
0.000010 2.000
0.000010 A.000
0.000010 6.000
0.000100 7.000
0.000500 8.000
0.000800 9.000
0.001100 10.000
0.001380 11.000
0.001560 12.000
0.001760 13.000
0.001900 14.000
0.002030 15.000
0.002200 16.000
0.002390 17.000
0.002520 18.000
0.002630 19.000
0.002770 ' 20.000
0.002850 21.000
0.002920 22.000
0.003010 23.000
0.003110 24.000
0.003190 25.000
0.003280 26.000
0.003350 27.000
0.003400 28.000
0.003510 29.000
0.003570 30.000
0.003610 31.000
0.003680 32.000
0.003710 33.000
0.003780 34.000
0.003800 35.000
0.003830 36.000
0.003890 37.000
0.003910 38.000
0.003950 39.000
0.003990 40.000
0.004000 40.750
0.000050 41.000
0.000030 42.000
0.000019 43.000
0.000015 44.000
0.000013 45.000
0.000011 46.000
0.000010 47.000
185
-------
RUN #L11
S02 FRACTION OF EFFLUENT TIME (MINUTES)
0.000000 0.000
0.000000 0.075
0.000000 1.000
0.000000 3.000
0.000010 5.000
0.000050 7.000
0.000110 9.000
0.000190 11.000
0.000260 13.000
0.000370 15.000
0.000480 17.000
0.000630 19.000
0.000820 21.000
0.001100 23.000
0.001390 25.000
0.001680 27.000
0.001950 29.000
0.002240 31.000
0.002480 . 33.000
0.002670 35.000
0.002860 37.000
0.003000 39.000
0.003040 41.000
0.003170 43.000
0.003280 45.000
0.003340 47.000
0.003430 49.000
0.003410 51.000
0.003590 53.000
0.003630 55.000
0.003840 57.000
0.004000 59.000
0.001500 60.000
0.000310 61.000
0.000190 62.000
0.000110 63.000
0.000090 64.000
0.000050 65.000
186
-------
APPENDIX D
A MODEL FOR PARTICLE ATTRITION BY ABRASION
IN THE UPPER ZONE OF A FLUIDIZED BED
INTRODUCTION
This study of particle attrition in fluidized beds was carried out
to aid in predicting makeup solids requirements and particle control in
exhausts in fluidized-bed energy systems.
Fluidized-bed fossil fuel processing systems are now being commer-
cialized. These systems provide for compact, nonpolluting combustion or
1 2
gasification of coal and other solid or liquid fuels. ' Studies of
fluidized-bed processes show that bed particles inevitably wear down into
fine dust. This wearing down of particles, called attrition, is variable
and not well understood.
If attrition rates are related to various properties of the particu-
late solids and fluidizing gas, and operating conditions, we should be
able to develop an expression describing the rate of attrition in any
given system. Several researchers have studied the effects of single
variables under various conditions, but no general prediction equations
have been formulated.
The examples used in this paper are generally related to fluidized-
bed combustion of fossil fuels. A fluidized-bed combustor consists of
0.5-to-2-mm particles of calcined limestone that captures sulfur dioxide
or dolomite (S02) as soon as it is formed
800-850°C
»
800-850°C
CaCO- »• CaO + C02 calcination
CaO + S02 + -z 02 • »CaSO^ desulfurization
187
-------
The specific objectives of this study were to
1. Identify the various causes or sources of attrition, focusing
study on the cause of attrition occurring in fluidized-bed
combustors and gasifiers
2. Develop an expression relating the attrition rate due to
abrasion in the bubbling zone (one of the sources of attrition)
to operating conditions and material properties
3. Test the proposed attrition formula in controlled laboratory
experiments.
HYPOTHESES
What are the Sources of Attrition in a Fluidized Bed?
The frequently considered source of attrition in a fluidized bed is
the obvious grinding and shattering collisions of particles. There are
several causes of particle wear, however, which include:
1. Abrasion. In this process defects, edges, and corners are
knocked from particles by low-energy collisions. Abrasion can
occur during passage of a gas bubble through the bed of solids.
2. High-Energy Collisions. Particles may be accelerated to high
velocity, for example when entrained in a jet at the distribu-
tion plate. The high-velocity particle can strike another
particle or vessel wall and shatter into relatively large
fragments.
o
Blinichev, Strel'tsov, and Lebedeva have distinguished two zones
in a fluidized bed: the lower, which they call the "nozzle
effect" zone in which gas jets accelerate large particles to
energies sufficient for shattering; and the upper zone,
characterized by intensive mixing and low-energy impacts which
grind particle surfaces.
188
-------
3. Thermal Shock. When cold particles are added suddenly to a bed
of red-hot solids, there is severe thermal stress on the cold
particles. One expects spalling at the particle surface and
4
perhaps shattering into large fragments.
4. Chemical Stress. Sorbent particles calcine, then react with
S02; calcium oxide (CaO) forms calcium sulfate (CaSO,), with
subsequent changes in the lattice structure. This change in
particle structure at its surface hardens particles in some
cases, or in other cases causes internal stresses leading to
4-6
spalling or weakened particle surfaces.
5. Internal Gas Pressure. When cold limestone or dolomite makeup
sorbent is added to a hot fluidized bed, the resulting calcina-
tion generates carbon dioxide (CO.) within the particle. Esso
Research Centre in Abingdon, UK, found that a slower calcination
rate of fresh limestone results in lower production of fines.
Similarly, water within particle cracks will flash when heated
to bed temperatures. While C0_ pressures are moderate (100.0 kPa
equilibrium at 900°C), steam pressures are high and can explode
particles.
6. Transfer Lines and Cyclones. These are not a part of the fluid-
ization process but are generally included in a fluidized-bed
system. This breakage rate is related to the circulation rate
of the solids and is controlled by equipment-design effects on
solids impact.
8
Kutyavina and Baskakov explain, "With fluidization, particles
are ground by abrasion and splitting. ... Abrasion is evidently predomi-
nant even for brittle and insufficiently strong materials."
9
Similarly, Wei describes two mechanisms of particle attrition:
"grinding" or the abrasive removal of a layer of
crystallites and matrix from the skin, and "shat-
tering" of the deep disintegration of the matrix
material.
189
-------
The former mechanism leaves behind a large particle
somewhat reduced in size and a pile of very fine parti-
cles; the latter mechanism leaves an assortment of frag-
ments from the very small to the very large. The former
is controlled by the hardness of the crystallites and the
abrasion resistance of the matrix; the latter is con-
trolled by the impact elasticity of the matrix and the
imperfections in the structure.
Doheim, Ghaneya, and Rassoul observed with fluidized iron ores in a non-
reacting system that the primary mechanism of attrition is by abrasion,
not breakage. Jonke of Argonne National Laboratory (ANL), observed "that
the mechanism for attrition is abrasion (the wearing away of surface
material), in contrast to the break-up or splitting of particles due to
particle or particle-wall collision." Blinichev and others report
that the wear of hard fluidized particles is by abrasion; soft materials
12 8
split, then abrade. Forsythe and Hertwig, Kutyavina and Baskakov,
13
and Zenz make the same observation.
In this appendix discussion is limited to only the first source of
attrition: abrasion, or grinding caused by rising gas bubbles in a fluid-
ized bed. In most fluidized beds several attrition mechanisms will act.
In this study, grid (distribution-plate) jets were eliminated in the
environmental work by using a porous, sintered-metal grid. Temperature
and chemical effects were avoided by operating at room temperature.
Relation between Gas Belocity U and Minimum Fluidization Velocity in
Attrition Testing
Meaningful measurement of attrition requires that tests be carried
out under some specified conditions; that is to say, values of parameters
(fixed variables) be specified.
Attrition testing is often carried out at some arbitrary gas veloc-
ity, U. In our earlier testing we measured attrition with the gas-
velocity parameter held at a constant multiple of the minimum fluidization
velocity U .. Exxon R&E has reported early testing at a constant gas
mf
velocity; some samples (low Umf) fluidize vigorously at a given velocity,
others (high U c) move only slowly. Exxon reports fluidizing all samples
mf
at a constant multiple of the minimum fluidizing velocity, U = 1.6 U .
190
-------
We propose that the rate of attrition for this mechanism of a
fluidized solid is affected by the rate at which energy is supplied to
the fluidized beds.
N
mass of fines
formed per mass
of bed solids
per unit time
effects of
all other
variables
[Q]
rate of energy
influx to the
bed per unit
mass of bed
solids
exponent N
expected to
be Ril.O
1.
This is a statement of Rittinger's law of grinding when N
The energy supplied by flowing gas to the bulk of a fluidized bed
(in the abrasion attrition mechanism) is virtually all pressure energy
that converts to kinetic energy of particles. In contrast, the energy
at the grid jets is kinetic energy. In the current series of tests the
grid is a sheet of sintered metal and grid jets are eliminated.
The specific* pressure energy causing attrition in the bulk of a
fluidized bed is that fraction of the gas flow causing bubbling and
particle collisions, namely the Pressure-Flow Rate energy per unit bed
mass
Energy - (U - Umf)AAP/M
U
(U - U -)A -
mf 8
(U - Umf )A
AZ
(u - u ,)
PbZ/M
Z/M
U
mf
A
AP
superficial gas velocity
min. fluidizing velocity
bed cross-sectional area
bed pressure drop
M = bed mass
g = gravity acceleration
g = Newton's law factor
p, = bed bulk density
V
Z
bed volume
bed depth
*Per unit mass of bed solids.
191
-------
It is the energy potentially available for attrition that is held con-
stant in these tests; if U - U f is 30 cm/s, for example, this energy is
-8-
-------
The goal of the following development is to formulate an expression for
the rate of particle attrition in the freely bubbling regime (upper zone)
of a fluidized bed. We begin development of a formula for attrition rate
by factoring the definition to include the rate of particle collision
-. _ grams of fines formed
sec x gram of coarses
grams of fines formed
|_ particle collision J
collisions
sec x gram of coarsesJ
(2)
The mass of fines formed per particle collision will depend upon
particle shape and strength
and the energy of the
sion. For example, consider
the spherical particle with
volume Vn = -7- D and mass
TT 36 p
M = T P D • Tne particle
p 6 s p
is traveling at some veloc-
ity U taken to be propor-
tional to the bubble
velocity, U, , related to the
b
bubble diameter, D, , by the
15
Davies-Taylor relation as
M35A04
Particle Mass \
Un« Area formed)
= C,
Flakes Have
Constant
Thickness
VolumeV,
(3)
Figure Dl - Dimensions of a chip
abraded from a bed
particle
The constant of proportionality is C.
U
and the particle velocity is
(4)
193
-------
During a collision, the particle's energy is transferred into creat-
ing new fracture surface area Af with efficiency C_. The fracture energy
of the solid is a (=) erg/cm , and the new surface area formed is
Af = C2 PS I °P3 Cl DP2/2 Sc ' '
Combining this with the expression for U gives
A IT Cl °2 ps 8 Dp Db
Af 54 go (5)
The chip volume and mass, and the fracture area, are related by
Vchip = C3 Af (6)
M . . = c, p A; ,
chip 3s f
where n = 1.5 for knocking corners from a cube and n = 1.0 for chipping
flat flakes from a rounded particle. We take n =1.0 and interpret C
to be a measure of chip thickness
f V
C, = -C. P (=) cm .
•J Af
I
For the remainder of the discussion, C, is taken to be constant for the
quasi-steady-state interval, typically one-to-four hours for attrition
in the bubbling regime. If we combine equations (5) and (7) and assume
that two chips are formed in each collision.
2 3
grams of fines formed 2 ' jr ps 8 p b
particle collision 1 2 3 27 go * (8>
This is the first factor needed in equation (2).
194
-------
The second factor in equation (2) may be factored further into
collisions _ (collisions/bubble) (bubble/sec)
sec x gram of coarses (gram of coarses) ' ^yj
Dwp. 6439A72
The number of collisions caused
by a bubble rising up the bed of
height
C4 x
x
=
as
Z is given by (Figure D2)
bubble area x bed depth
particle
unit bed volume
„ ir -2 _ 1-e
C, 7- D, Z r-
4 4 b ir _3
T u
o p
- Z D2
X K
C,. (1-e) 4 — TT , (10)
fSsr
1 \ iK
\\ 1 1 \
fctt
\\'\
A V
A \\
^ ll
\ \ 1 .
N .
Projected Bubble Area = A&
^ Area in Which Collisions are
Caused by Movement of the
Bubble = 0^
^ If —
'AH o o
*=> <=> Grid Holes
<=> o cs
vp Figure D2 - Volume of particles dis-
turbed by a rising
bubble
where e is bed porosity. The number of bubbles formed per second is
(U - U f) A 3
—Si where V, - -r D^ is bubble volume.
V, bob
b
Although not immediately evident, there are several circumstances
that suggest the influence of another variable:
• The dimensional constant C' (=) length
• The dimensionless group C'/Z where Z is the bed depth
0 Occurrence of the gravitational constant g, which implies
something like hydrostatic pressure in the fluidized bed
and, therefore, inclusion of the bed depth Z.
Consequently, in equation (8) the factors C.Z replace Ci; rearrangement
gives
r * z "i. ic2 c
[u-UmfJ 3C1C2
g Z Ps
8,, o
(11)
195
-------
Combining the constants in this expression gives
r__Kj^i
LU - v J
= C x
Z2P
s
g 0
6
(12)
STROUHAL BOND
NUMBER NUMBER
It is physically consistent that the Bond number N should occur; it is
the ratio of gravitational force to surface force. Gravitational force
is necessary to press particles together, and the surface force is what
resists attrition. The rate of attrition being proportional to (U-U ..)
mf
is realistic: no attrition should occur for U < U , as the bed is
mf
static; (U-U ,) is proportional to the rate of energy input to the
mt
fluidized bed.
There is no universal agreement on the rate of attrition being
proportional to the excess bubbling velocity U-U • Gonzales and Otero
assume the relation dD /dt = - C D in which D is particle diameter and
P s P e
C and m are constants. They take C « U where U is gas velocity in the
bed and conclude from experimentation that the constant e = 0; that is,
attrition rate is independent of U and, in turn, independent of U-U ..
17 mf
Merrick and Highley, on the other hand, invoke Kittinger's law of size
reduction by abrasion, assume the rate of input of energy to the bed to
1 /1M
be proportional to (U-U £), and deduce that R = - rj — = K (U-U ,-) . which
mt M dt mf
agrees with the proposed model, equation (12).
A possible pitfall lies in interpreting the Merrick and Highley
equation for particle diameter D , dD /dt = - -r K (U-U ,) D and conclud-
p p J mf p
ing that the attrition rate decreases as particle diameter decreases.
This equation is derived from the first-order law describing the mass of
the bed M, dM/dt = - K (U-U f) M. This equation does not imply that
attrition rate depends on particle size. By substituting M = p ~ D
1 6 p
one derives dDp/dt = - — K (U-U f) D , which again does not suggest a
rate dependence on particle size.
196
-------
Data from Other jkmrces Indicate that Attrition Rate is Independent
of Particle Size
Equation (12) indicates that the rate of attrition for fluidized
9
abrasion of particles does not depend on particle size. Wei and others,
in discussing the attrition in jet mills, note the absence of experimental
data:
The strength of an impact of a catalyst on the
steel wall or on another catalyst can be measured by its
kinetic energy (l/2)mv2. In the absence of conclusive
experimental data, one may theorize several alternatives:
that the rate of attrition should be proportional to the
impact energy, or that there should be a threshold energy,
or that the attrition rate should increase faster than the
impact energy. This point is neglected in the literature.
Since particles of all sizes travel at nearly the same
speed in the fluidized bed, the major variable in kinetic
energy is the particle mass, roughly the cube of the par-
ticle diameter. On the other hand, the energy required to
break a particle is proportional to the amount of new sur-
face area formed; if the fracturing pattern is geometri-
cally similar, the energy varies as the square of the
diameter. In this case, the ratio of impact energy to
fracturing energy requirement is proportional to particle
diameter. Therefore, the larger the particle diameter,
the greater is its rate of attrition. The mass fragmen-
tation rate and pattern of catalyst cracking in a jet mill
experiment yield useful information about the attrition
mechanism, but this could be significantly different from
what is actually happening in a commercial F.C.C. Thus,
one should be extremely careful in the interpretation of
the laboratory data.
It is important to notice that Wei is discussing the high-energy colli-
sions of particles in a jet where particles shatter, not the low-energy
abrasion in which chips are worn from a particle's surface.
Others interpret their data to suggest an effect of particle diam-
eter on the rate of attrition, but scrutiny of those findings indicates
18
no particle-size effect on attrition rate. Tarman and Punwami of IGT
provide useful data. In studying the attrition of siderite at room
197
-------
temperature, they observed that attrition rate varies inversely with
initial particle size at a superficial gas velocity of 52 cm/s. Note
that the superficial gas velocity U, not excess bubbling velocity
(U-U -), was held constant. If we apply the IGT-developed expression
19
for U ._ (Babu and others ) ,
mf
1/2
Jjj- {(25.252 + 0.0651 Ga)1/2 - 25.25}
g P
Ga
°P P
where:
gas viscosity (g/cm-s)
g = gravity acceleration (cm/s )
2
p = solid density (g/cm )
s 3
p = gas density (g/cm )
o •
D = particle diameter (cm),
to Tarman and Punwami's conditions, we have the data listed in Table Dl,
Table Dl
ANALYSIS OF IGT DATA
Initial Particle
Diameter, pm
133
173
205
246
352
436
* y = 0.0017 g/cm-s
p =3.0 g/cm
S o
p = O.OC13 g/cm
o
Umf>
cm/s
3
5
7
10
20
30
U-U ,,
mf
cm/s
49
47
45
42
32
22
• Assumed conditions
198
-------
These are plotted in Figure D3, which shows an inverse relation between
U-U c and particle size.
ml
80
70
n 60
50
i. 40
E 30
Curve 718701-A
20
Best Fi t Inverse
Relation Fit to
Data
O Tarman and
Punwami's Data
I i I
TOO 200 300 400 500
Particle Diameter, pm
Figure D3 - Variation of U-Umf with particle diameter for
siderite fluidized at U = 52 cm/s (1.7 ft/sec)
Applying the attrition equation developed in this study, equa-
tion (12),
to the conclusion from Figure D3,
(U - U ) « D'1
mf p
for constant U infers that
R = D
-1
for constant U within the range of Tarman and Punwamifs data.
199
-------
The above development shows that we expect Tarman and Punwami's
result. The attrition rate R varies inversely with particle diameter D
in this special case where U-Umf varies with D^. The attrition rate R
mf
mf p
would not be expected to vary if U-U f had been held constant.
Gas velocity is sometimes expressed in terms of the dimensionless
20
and unfortunately chosen "fluidization number," U/U . Doheim, Ghaneya,
10
and Rassoul studied attrition of iron ore fluidized with hydrogen above
a woven-mesh grid. They conclude that "the amount of fines generated by
attrition is larger for the coarse iron ore ... attrition increases with
particle size." Their comparisons are made at constant U/U = 2, not
at constant power input to the bed for which U-U is constant. Analysis
of their data for attrition without chemical reaction gives the values
of r for R « (U-U Jr listed in Table D2.
mf
Table D2
INTERPRETATION OF DOHEIM, GHANEYA, AND RASSOUL'S DATA
Temp . ,
°C
25
450
D
P
Range
250-315
315-400
250-315
315-400
Mean
280
355
280
355
R,
mg/min
21.0
30.7
43.1
58.2
20.5
25.5
10.8*
15.13*
[Roc(U-Umf)r]
1.74
0.89
U-U - 2U - - U - - U ,.
mf mf mf mf
*Calculated assuming fluidization with hydrogen.
As this analysis shows, the dependence of attrition rate R on U-U
is variable, averaging r = 1.3. If we allow for variance in the data,
this is interpreted to fit the model of R « (U-U ,) * . We conclude
mi
that particle size does not influence attrition rate directly but
indirectly as it affects U f, which in turn changes U-U -, a variable
controlling R.
mf
200
-------
Data from Other Sources Indicate that Attrition Rate Is Proportional to
Bed Depth. Z
Our model for attrition rate R at depth Z is the bubbling zone of a
fluidized bed as given by equation (12). The model relates attrition
rate R as proportional to bed depth Z as measured downward from the bed
surface (Figure D4). Assuming
that (U-U ,) is independent of
mi
bed depth Z, we can calculate
L, the total rate of mass loss
from a bed containing mass M
of solids by integrating equa-
tion (12) over the entire bed
mass.
D«n. 6439A73
'••.' : •
Figure D4 - Depth Units
m
L = f R d (mass of bed, m) (13)
J R (bulk density of bed) d (volume of bed, v)
(14)
/.. (particle density \ ,, , . . , ,„„.
R \ 1-bed porosity ) (bed "oss-sectional area) dz (15)
Azdz
(16)
(17)
201
-------
Equation (17) states that the total rate of fines production by attri-
tion (gram/second) is proportional to the square of bed depth h. This
is substantiated by experimental results from the Esso Research Centre,
Abingdon, UK.
Researchers at Esso reported:
The relative importance of all variables affecting bed
loss rate of BCR 1691 is summarized in the equation
below. This was derived from further analysis of the
fresh bed test results.
1 1 7
223.31 x h
t°'44 x (T-750)1'80 x s°'41
L = total loss rate from bed (g/min) (= R x bed mass)
h = bed depth (inches)
t = bed age (hours)
T = bed temperature (deg. C) (750 deg. C is taken as
CaCOo decomposition temperature)
S = bed sulphur content including inherent sulphur
(% by weight)
This equation shows the approximately square relation-
ship between losses and bed depth as observed previously.
The loss rate of 3.1 g/min calculated from the above
equation for cycle test conditions is in fair agreement
with the measured rate of 4.5 g/min.
The Westinghouse model and the Esso data agree that the rate of solids
loss (g/m) is proportional to the square of total bed depth, h; this in
turn substantiates that local attrition rate R is proportional to the
depth in the bed, Z.
21
Equation (20) is further confirmed by ANL's tests in which bed-
depth-to-bed-diameter ratio, h/D, was increased from 1.3 to 2.13, probably
all within the bubbling, nonslugging regime (Table D3).
202
-------
Table D3
ANL ATTRITION TESTS OF CALCINED TYMOCHTEE DOLOMITE
h/D
1.30
2.13
Fluidizing-Gas
Velocity, ft/s
2
2
% Loss - Rt = A
1/2 hr
0.5
0.9
1 hr
0.7
1.2
2 hr
0.9
1.7
5 hr
1.6
3.0
7 hr
2.3
3.9
10 hr
3.0
5.3
ANL's data give ratios of Rt h/D=2>13 * Rt h/D=1>30 of 1.80, 1.71,
1.89, 1.88, 1.70, 1.77, which average 1.79. These results fit closely
with Esso's observation that the loss rate L (g/min) is proportional to
(bed depth)2'17
2.17
(18)
In Table D3, R is an overall bed attrition rate time averaged between 0
and time, t. We note that R for a bed of depth h, R(h) is given by
R(h)
L(h)
L(h)
Vp (1-e) Ahp (1-e)
P P
= K
L(h)
(19)
where K is some proportionality constant. For two different beds of
depths h. and h.
R (h2)t R (h2) KL(h2)/h2
2.17
/h.
h.
,1.17
2.17
(20)
i -i ~i
Hence, from Esso's observation of L « hA" , we expect ANL's data to
correlate by
R (h2)t R (h2/D)t /h.
a.17
h2/D
a.17
(21)
203
-------
and, indeed, the mean Rt ratio of 1.79 agrees well with
2 is^'17
ril) -1-73- <»>
We concluded that steady-state attrition in the freely bubbling zone
of a fluidized bed (above the grid-jet region) is indeed described by
equation (12)
STROUHAL BOND
NUMBER NUMBER
The Rate of Attrition Decreases with Time to a Steady State
The preceding expression for attrition applies to a steady-state
condition in which the rate of attrition is constant. Several workers
have reported that the rate of attrition decreases with time:
29
• Vaughan at Battelle Columbus Laboratories (2/7/77)
reported, "there is an initial attrition period where the
attrition loss is relatively high; this is perhaps due to
the corner rounding or other stabilization effects."
23
• Curran at Conoco observed a reduction in attrition rate
when fluidizing sulfied dolomite, described by R - At
with b ranging between -0.19 and -0.32. He attributed the
rate decrease to sintering and densificatioa of the stone.
• Merrick and Highley postulate that as attrition progresses
finer particles will spend part of the time in voids between
larger particles, and during this time no attrition will
occur. Thus the attrition rate will change as time
progresses.
204
-------
24
Mathur and Epstein remark on spouted beds:
The grinding (attrition) rate tended to drop off with
spouting time, as would be expected in any batch grind-
ing operation.
25
Stanley and others report the equation form for specific
surface in attrition milling using a mechanical mill:
s = s
ultimate
in which a and t are constants, S n . is the limit-
' ultimate
ing specific surface.
We can rearrange this equation to
dt (Sultimate ~ S) = ~ a (Sultimate ~ S) •
which implies decreasing attrition rate with time.
Q
• Kutyavina and Baskakov report the equation for total fines
formed (M -M) as
o
m"
M - M = k tm
o
in which k and m are constants. This equation form reflects
a falling attrition rate. They explain, "The rate of abrasion
decreased over the course of time, with rubbing off of the
uneven parts and a decrease in the number of defects of the
particles."
12
• Forsythe and Hertwig explain the decrease in attrition rate
with time as being caused by smoothing of rough edges, elimina-
tion of rough particles, and cushioning by the increasing frac-
tion of fines.
18
• Tarman and Punwami propose that fines fluidize between larger
particles, thus lubricating and reducing abrasion.
It seems well established that the attrition rate for initially
angular particles decreases with time. The mathematical description of
205
-------
this decrease is uncertain in the literature, and it seems improbable
that it could be predicted. Exponential and power models seem reasonable
yet they are guesses. In a fluidized bed in the bubbling upper zone, the
proposed attrition model is such that the rate of attrition decreases
monotonically to the steady-state rate of equation (12). Defining the
decreasing function F(t) by
F(t) -0 , t
dF(T)
dt
F(t)
0 < t <
0 < t <
the model for transient and steady-state attrition is
(23)
This model then predicts the form of the attrition rate curve and the
extent of attrition curve to be those shown in Figure D5. This is the
final current form of the model proposed to describe attrition by abra-
sion in a FBC.
Curve 696084-A
Steady State
<:
c
_o
'^
&
<
"o
"c
o>
*-
X
Time, t
Time, t
Figure D5 - Forms of the attrition rate curve R(T) and its
t
integral, the extent of attrition A(t) = f R(t)dt,
o
206
-------
Experimental Results Agree with the Attrition Rate Model
The model defined by equation (23) and pictured in Figure D5 is based
on a variety of observations from different researchers. The observations
have been unified into a single statement [equation (23)] under the condi-
tions of consistency in dimensions and physical reality. Equation (23),
however, is at this point a theory, unproved.
This experimental program tested two hypotheses of the attrition
theory; first, that the attrition rate decreases steadily to a constant
rate; and second, that attrition rate is proportional to U-TJ ... We did
mi
not test other hypotheses.
Figure D6 shows the apparatus for these experiments. The sintered-
metal distribution plate eliminates jets and the attendant attrition from
high-velocity collisions. The manometer allows measurement of a pressure
drop across the bed of granular solids and, in turn, the minimum fluidiza-
tion velocity U at which bubbling in the bed of solids first begins.
The test procedure was to charge the apparatus with Grove limestone
about 7 cm deep, then measure the minimum fluidization velocity U .*
mf
The bed was then fluidized in air at room temperature for time intervals
after which gas flow was stopped and the bed solids were sieved into size
fractions and weighed.
Test Results Substantiate the I_dea_ that Attrition Rate Decreases
to Constant Rate
The purpose of the first series of experiments was to test the hypoth-
esis that attrition rate decreases to a constant value as indicated in
the theoretical development [equation (23)] and depicted in Figure D5.
The experimental approach was to fluidize a mass of stone over an interval
of time long enough to determine the shape of the rate curve.
*The value of Umf is found by increasing superficial gas velocity U
through the solids and recording pressured drop through the solids.
At U,^ the bed weight equals the pressure force supporting the bed
and Ap remains constant with increasing U.
207
-------
We used the apparatus shown in Figure D6. A mass of 32 by 42 mesh
limestone was first fluidized for 15 minutes, then wet sieved for measure-
ment of the mass of coarse (>355 \im or 42 mesh) stone remaining. We
repeated this procedure with increasing time intervals for a total
fluidization time of 647 hours. Results are listed in Table D4.
The rate of attrition or rate of formation of fines is the first-
order rate process defined by
n _ 1 dM
R ~ ~ M d^
where M is the mass of coarse granular solids in the bed. The percent of
attrition is the mass of fines formed per unit mass of coarse bed solids
(xlOO) or
t
A
o
/I rlM
~ M dF dt '
and with M = M at t = 0, the above expression integrates to
o
M
A = % extent of attrition - 100 In ^°-
This is the value listed in the last row of Table D5 and is the measure
of the extent of attrition.
Figure D7 is a graph of the extent-of-attrition data from Table D4;
this graph of A against time looks like a logarithm graph so we chose a
regression model for curve fitting:
A = K In (1 + ^x + k2x2 + k^3 + k^x + k5x5)
Regression analyses gave the values
K = 1.4688 k3 = 1.99893 E - 04
k, = 3.2658 k. = 4.83055 E - 07
1 4
k = -2.71103 E - 02 k = 3.78150 E - 10
208
-------
DM;. 261JC29
Table D4
DATA FROM THE 647-HOUR ATTRITION OF GROVE 1359 LIMESTONE, NOVEMBER 1977, 25°C
Time Interval
Length
of Time
TotalTime.hr
Wt>42, M
Wt<42 F*
D
Ifb
Filter
Solids
Gross
Tare
IFT
Accumulated
Total Fines
Z FK + Z FT+M
D T
Total Solids
Possible
Total Fines
M°
100 In ~
Start
0
0
316.1467
0
0
4,2434
4:2434
0
0
3LU467
0
0
1
1/4
1/4
313.35
2.4884
2.4884
4.5517
4:2434
0.3083
2.7967
316.1467
2.Z967
0.79
2
1/4
1/2
311.1658
1.8614
4.3498
4.5672
4.2434
0.3228
4.6726
315.8384
4.9809
1.59
3
1/2
1
309.0195
1.7926
6.1424
4.5971
4.2434
0.3537
6.4961
315.5156
7.1272
2.28
4
1
2
306.9038
1.7081
7.8505
4.6510
4.2434
0.4076
8.2581
315.1619
9.2429
2.97
5
2
4
304.8045
1.6492
9.4997
4.6935
4.2434
0.4501
9.9498
314.7543
11.3422
3.65
6
4
8
301.5885
2.7149
12.2146
4.7445
4.2434
0.5011
12.7157
314.3042
14.5582
4.71
7
8
16
298.9002
2.1062
14.3208
4.8255
4.2434
0.5821
14.9029
313.8031
17.2465
5.61
8
16
32
296.6960
1.5707
15.8915
4.8769
4.2434
0.6335
16.5250
313.2210
19.4507
6.35
9
32
64
294.4493
1.5438
17.4353
4.9463
4.2434
0.7029
18.1382
312.5876
21.6974
7.11
10
64
128
290.1494
3.4671
20.9024
5.0762
4.2434
0.8320
21.7352
3L1.8846
25.9973
8.58
11
128
256
287.15
2.0223
22.9247
5.2205
4.2434
0.9771
23.9018
3L1.0518
28.9967
9.62
12
200
456
284.238
1.7891
24.7222
5.3659
4.2434
1.1225
25.8447
310.0747
31.9083
o
10.64
13
191
647
280.8914
2.1689
26.1911
5.4215
4.2434
1.1781
28.0692
308.9606
35.2553
11.82
to
o
vo
Calculated Estimate. Not Measured
-------
Rota meter
Pressure
Taps
Plenum
Dwa. 6419A42
Exhaust
Balston
Filter
Plexiglas Fluidized Bed
6.99 cm ID x 91.44 cm High
-Sintered-Metal Distributor
Plate
—© Pressure Gauge
S & K 31827
Rota meter
Valve
Regulator
House Air
Figure D6 - Flow diagram for room-temperature fluidized bed
210
-------
Table D5
SUMMARY OF DATA COLLECTED AND CALCULATED
VALUES FOR BED-SOLIDS MASS HAD SAMPLES
NOT BEEN REMOVED.
C02-FREE BASIS
Hours
Fluidized
Start
1
3
10
30
100
200
Measured
Mass
g
483.3
468.3
456.5
455.3
448.3
447.6
444.4
443.6
434.4
437.4
414.5
% C02
()=meas'd
[]=est'd.
(0.5)
[0.4]
(0.35)
[0.4]
(0.42)
[0.42]
(0.20)
[0.25]
(1.47)
[2.35]
(0.59)
Mass
C02-Free
Stone, g
481.08
466.43
454.90
453.48
446.40
445.72
443.51
442.49
428.01
[427.1]
412.05
Mass C02-Free
Stone, Corrected
as if Sample Was
Not Taken
481.08
466.43
454.90
—
447.80
—
445.58
—
431.00
—
415.81
% Attrition
100 In
(481.08/M)
0
3.09
5,60
7.17
. 7.67
10.99
14.58
Curve 7ISW-A
Initial Particle Size 350-495
h, bed depth 18cm
200 300 400 500
Tlmeof Fluidization, hours
Figure D7 - Change of Extent of Attrition with Time
in Bubbling Fluidization
211
-------
Figure D8 shows its derivative (% rate of attrition). These results,
spanning 647 hours of vigorous fluidization, demonstrate that the rate
of attrition does indeed decrease continually to a constant steady state.
Notice how the rate R falls precipitously from its initial value of
4.80 %/hr at t=0 to 1.81 %/hr at t = 0.5. The rate approaches its final
value of R = 0.0055 %/hr asymptotically. The rate reaches 99 percent of
its final value (4.80 - 0.99 {4.80 - 0.0055} = 0.0534) after 25 hours of
fluidization.
These results, more than ever, underscore the necessity for report-
ing attrition results on a consistent basis. Individual researchers
need to keep the time interval of fluidization constant where possible
or otherwise account for the effect of time.
Hot Attrition Testing Further Confirms that Attrition Rate Decreases to
a Constant Rate
The previous tests at room temperature showed that attrition rate
decreases with time until a constant rate is reached. We extended these
experiments at combustor temperature to confirm that the rate decreases
to a steady state under process conditions also.
Our apparatus consisted of a 8.57 cm-id attrition test cell fitted
with a sintered-Inconel grid. We calcined 500-to-710 ym Grove limestone
to <0.5% C00, then fluidized it vigorously at 815°C and U=U , + 20 —.
^ mt s
After one hour of fluidization we removed the solids, sieved out fines
smaller than 500 ym, weighed the coarse solids larger than 500 ym, and
returned them for continued fluidization. This process was repeated
several times with increasing intervals of fluidization time. Samples
were removed for CO- assay. Figure D8 and Table D5 summarize results
and calculations. The spots in Figure D8 (•) are measured masses of solids
all other masses shown were calculated on the assumption that the curves
are proportional.
The percent extent of attrition is calculated by
M°
% attrition, A = 100 £n ~
M
212
-------
where M° is the starting mass of solids. Calculated data from Figure D8
give these values of A plotted in Figure D9. Differentiation of this
curve provides values of A, the percent attrition rate, plotted in the
same figure. Values of A and R measured in these tests are listed in
Table D6.
The effect of temperature on attrition rate is revealed by comparing
results of both hot and cold attrition testing. Earlier cold tests were
carried out with uncalcined 355-to-500 ym Grove 1359 limestone at a veloc-
ity of U r + 200 cm/s and a temperature of 25°C. The hot tests were
mf
performed with calcined 355 to 500 ym Grove 1359 limestone at a lower
velocity of U .. + 20 cm/s and at 810°C. In the hot tests the resistance
ml
to attrition, particle strength, was apparently much less, and the solids
attrited notably faster. The characteristic values for the two tests
are shown in Table D7.
Curve 697006-A
8
t/T
E
481.08
Fictitious bed mass for
fluidization without
sample removal
447.80
445.72
442.49 J
• Denotes measured mass
o Denotes calculated mass
Temp=815±5°C
Not to Scale
428.01
427.1* 412.05
10
Time, hours
30
100
200
Figure D8 - Summary of Data Collected and Calculated Values for Bed
Solids Mass Had Samples Not Been Removed. See Table D5,
213
-------
Table D6
VALUES AT ATTRITION EXTENT AND RATE MEASURED FOR
24-32 MESH CALCINED GROVE LIMESTONE FLUIDIZED AT
810 °C WITH U = U .. + 20 cm/s
mt
Hours Fluid! zed
0
1
3
10
30
100
200
Extent of Attrition, A
%
0
3.09
5.60
7.17
7.67
10.99
14.58
Rate of Attrition, R
%/hour
3.86
2.18
0.71
0.052
0.0407
0.0407
Curve 697005-A
A =6.68+0.04071-6.6 e
25
50
75 100 125
Time, hours
5 10 15
Time, hours
J L
150 175
200
Figure D9 - Extent of Attrition and Attrition Rate for Grove 1359
Limestone Fluidized at U = U , + 20 cm/s and 875°C
mf
214
-------
Table D7
COMPARISON OF ATTRITION RATES AND TIME INTERVALS
REQUIRED TO REACH STEADY STATE
Test Temperature,
°C
25
810
Transient Interval
Required to Approach
within 99% of Steady
State, hr
211
7.95
Steady-state Rate
of Attrition,
%/hr
0.0056
0.0407
It is evident that attrition proceeds much faster at the higher
temperature, presumably because hot calcined stone is weaker than cold
uncalcined stone. At the high temperature the transient is much briefer.
These results suggest the caution needed in inferring attrition
rates in a hot system from cold attrition rate data.
Test Results Substantiate Idean that Attrition Rate is Proportional to
the Excess Bubbling Velocity U-
The development of an attrition equation was based on the hypothesis
that the rate of formation of new particle surface, and in turn the rate
of attrition (grams attrited /grams of coarse solid/hour) , is proportional
to the excess fluidizing velocity U-U . We have tested this by measur-
ing the specific surface of bed solids as a function of time at two
different values of U-U ...
mr
The experimental apparatus was that shown in Figure D6 . We measured
the size distribution of 330 g of crushed Grove limestone, then determined
its minimum fluidization velocity from a velocity - AP curve. The stone
was then fluidized in air at room temperature at U = U . + 25 cm/s for
mi
increasing time intervals up to a total time of eight hours. After each
interval of fluidization we measured the size distribution of bed solids.
Then we repeated the entire test, but at U - U
lists results from this test.
mf
12-1/2 cm/s. Table D8
215
-------
Table D8
Dwg.l692B53
DEPENDENCE OF SOLIDS SPECIFIC SURFACE ON TIME OF FLUIDIZATION
U-Umf = 25cm/s
Time Interval, hr
Total Fluidization Time, hr
Sieve
Mesh
42
60
115
250
325
Pan
Filter
Lost3
Dj , Mean
Diameter.
cm
0.042
0.030
0.018
0.009
0.0052
0.0036
0.0003
0.0003
Specific Surface, cm /ga
%Increasein Specific Surface
Start
0
1/4
1/4
1/4
1/2
Mass of
330
0
0
0
0
0
0
0
53.91
0
316.188
10.857
0,429
0.418
0.048
0.016
1.336
0.709
101.46
88.2
315.815
10.366
0.328
0.148
0.010
0.009
2.487
0.837
130.25
141.6
1/2
1
1
2
2
4
4
8
Solids on Sieve, g
311.631
12.516
0.510
0.181
0.013
0.009
3.984
1.156
171.69
218.5
309.081
12.869
0.599
0.122
0.012
0.009
5.691
1.617
220.92
309.8
303.903
15.849
0.723
0.186
0.025
0.009
7.643
1.662
266.54
394.4
300.957
16,621
0.906
0.136
0.019
0.014
9.002
2.345
312.97
480.5
U-U , = 12.5cm/s
mf
Time Interval, hr
Total Fluidization Time, hr
Sieve
Mesh
42
60
115
250
325
Pan
Filter
Lost a
DJ. Mean
Diameter.
cm
0.042
0.030
0.018
0.009
0.0052
Specific Surface, cm /ga
Increase in Specific Surface
Start
0
1/4
1/4
1/4
1/2
1/2
1
1
2
2
4
^ 4
8
Mass of Solids on Sieve, g
330
0
0
0
0
0
0
0
53.91
0
296.667
31.001
1.266
0.468
0.024
0.009
0.032
0.533
69.36
26.7
294.024
33.047
1.451
0.552
0.036
0.033
0.129
0.728
76.27
41.5
293.343
32.904
1.563
0.617
0.095
0.039
0.441
0.998
89.62
66.2
289.512
35.423
1.972
0.703
0.118
0.036
0.967
1.269
108.05
100.4
281.286
42.124
1.817
0.696
0.222U
0.030b
1.573
2.252
144.64
168.3
278.797
40.065
2.163
3.022
0.287
0.023
3.102
2.541
187.31
247.5
aSp. surf = 6 Z (Mj-5- Dj)/Mop
b Recorded datum of 0. 128 rejected and replaced with 0.030 by interpolation .
216
-------
The rate of increase in specific surface* was calculated by fitting
a regression curve to the percent increase in specific surface vs . time
data and differentiating. Results of this procedure are listed in
Table D9. Specific surface was calculated from the particle size
distribution.
Table D9
CALCULATION OF RATE OF INCREASE IN SPECIFIC SURFACE
U-U
mi.
cm/s
25
12.5
Regression Line
S(t)
loe.oit °'A9°
35.86t °-632
QP o i
dt " b
51.94t -0'510
22. 7t ~°'638
C ' / O 1
25 12.5
2.29t -°<128
The mean value of this ratio function S' /S'
/D \.
interval 0 <_ t <_ 8h.
is 2.003 over the
The hypothesis being tested is that the rate of increase in specific
surface is proportional to the excess fluidization velocity; that is,
dS/dt - S1 = (u - U C)P
mr
where p % 1. The average value of p calculated from Table D9 data is
si
; 2.003
25
12.5
1.002 .
This estimate of p is based on two data points (U-U , - 25 and 12.5) and
mr
does not have an error estimate. The calculated 1.002 is close to the
hypothesized exponent of 1.0, and we conclude that the rate of generation
of specific surface is proportional to U-U _. This is, in turn, related
mr
to the rate of attrition.
*Specific surface, denoted by S, is defined as the total particle surface
area per unit volume of particulate solids. It has the units of
cm
~l
217
-------
J o
The tests described in this section demonstrate that (U-U „) «—.
mf7 dt'
the rate of formation of new particle surface area. Rittinger's law
j C
used in developing the Westinghouse model states that -r^ « input power
by means of gas flow. For attrition in the bubbling bed input
power « (U-U ,.). It follows that
mi
JO J Q
R a —- • —— oc input power; input power = (U-U £) .
at at mt
It follows then that R « (u~umf): tne attrition rate in the bubbling
regime of a fluidized bed is proportional to the energy velocity above
the minimum, U-U ...
mi
Discussion
The expression developed to describe the rate of attrition, R, in
the bubbling zone of a fluidized bed is
lm
^z2
m % 1 . (23)
JS
Experiments described in this appendix demonstrate that
first: R « (U-U ,.) and
ml
second: R « [F(t) + 1] ; F—»-0 and F1 *0 as t—»»
F' < 0 for 0 <_ t * °° ;
and other experimental data, as described earlier in this paper, demon-
strated that
third: R « z .
The remainder of the theory, inferred by analogy and induction from
other systems, proposes that
fourth: R « I/a (soft materials attrite more easily than hard
materials)
fifth: Rap g/g (attrition rate is proportional to the
gravity force on the bed of solids)
218
-------
The transient decrease in attrition rate, described by F(t), points
out the need for caution in describing attrition rates. The average
attrition rates, as reported in the literature, will vary with the time
interval for averaging.
The attrition rates indicated by short-term tests may indicate a
much higher rate than will be encountered in a commercial system. After
hundreds or thousands of hours of fluidization, particles should attrite
at the lower steady-state rate, and the rate of solids loss should be
much less.
The increase in attrition rate with bed depth suggests design of
deeper or shallower beds for control of fines production. Similarly,
the effect of U-U ,. on attrition rate suggests control of U , where U,
mr mi
the superficial gas velocity, is fixed. The minimum fluidization veloc-
ity, U r, can be controlled by selection of particle size.
mr
Most effort in controlling (usually minimizing) attrition in
fluidized-bed sulfur sorbent systems has been an extensive search for
the ideal stone that attrites slowly because it is strong. The results
of this study show that we can also control attrition rate in the bubbling
zone by judicious specification of equipment and operating conditions.
REFERENCES
1. Keairns, D. L., ed., Fluidization Technology, Washington, B.C.:
Hemisphere Publishing Company; 1976.
2. National Research Council, Committee on Processing and Utilization
of Fossil Fuels, Reports on Assessment of Low- and Intermediate-Btu
Gasification of Coal and Assessment of Advanced Technology for Direct
Combustion of Coal, Washington, D.C.: National Academy of Sciences;
1977.
3. Blinichev, V. N., V. V. Strel'Tsov, E. S. Lebedeva, An Investigation
into the Size Reduction of Granular Materials during their Processing
in Fluidized Beds, Int'l Chemical Engrg, 8(4): 615-618; October 1962.
219
-------
4. Snyder, R., et al., Annual report on a Development Program on
Pressurized Fluidized-Bed Combustion, Argonne National Laboratory,
Argonne, IL, July 1976, ANL/ES-CEN-1016, p. 189.
5. Paige, J. I., J. W. Town, J. H. Russell, H. J. Kelly, Sorption of
SO- and Regeneration of Alkalized Alumina in Fluidized-Bed Reactors
Bureau of Mines Report of Investigations 7414, August 1970, p. 32.
6. Chemically Active Fluidized Bed Process, Monthly Technical Narra-
tive No. 20, January 24-February 20, 1977, Foster Wheeler Energy
Corporation, Livingston, NJ, Prepared March 14, 1977.
7. Craig, J. W. T., et al., Chemically Active Fluidized Bed Process for
Sulphur Removal during Gasification of Heavy Fuel Oil, Second Phase
Report to EPA, Esso Research Centre, Abingdon, UK, November 1973,
EPA-650/2-73-039.
8. Kutyavina, T. A., and A. P. Baskakov, Grinding of Fine Granular
Material with Fluidization, Chemistry and Technology of Fuel Oils,
8(3): 210-13; March-April 1972.
9. Wei, J., L. Wooyoung, and F. J. Krambeck, Catalyst Attrition and
Deactivation in Fluid Catalytic Cracking, Chemical Engineering
Science, 32(10): 1211-18; 1977.
10. Doheim, M. A., A. A. Ghaneya, and S. A. Rassoul, The Attrition
Behavior of Iron Ores in Fluidized-Bed Reactors, La Chimia E
L'Industria, 58(12): 836-40; December 1976.
11. Jonke, A. A., A Development Program on Pressurized Fluidized-Bed
Combustion, Monthly Progress Report, Argonne National Laboratory,
Argonne, IL, June 1976, ANL/ES-CEN-F092, pp 29-35.
12. Forsythe, W. L., Jr., and W. R. Hertwig, Attrition Characteristics
of Fluid Cracking Catalysts, I&EC, 41(6): 1200-06; June 1949.
13. Zenz, F. A., Find Attrition in Fluid Beds, Hydrocarbon Processing,
pp 103-5, February 1971.
220
-------
14. Regenerable Sorbents for Fluidized Bed Combustion, Quarterly
Progress Report to EPA, No. 5, Exxon Research & Engineering Com-
pany, Linden, NJ, January 1-March 31, 1977.
15. Davidson, J. F., and D. Harrison, Fluidization, New York:
Academic Press; 1971, p. 18.
16. Gonzales, V., and A. R. Otero, Formation of VO. Particles in a
Fluidized Bed, Powder Technology, 7(3): 137-43.
17. Merrick, D., and J. Highley, The Effect of Particle Size Reduction
on Elutriation from a Fluidized Bed with Feed from a Wide Size Dis-
tribution, AIChE Symposium Series No. 137, V. 40: 366-78.
18. Tarman, P. B., and D. V. Punwami, Development of the Steam-Iron
System for Production of Hydrogen for the Hygas Process, Interim
Report No. 2 to ERDA, IGT, Chicago, IL, July 1, 1974-June 30, 1975,
FE-1518-34.
19. Babu, S., B. Shah, and A. Talwalker, Fluidization Characteristics
of Coal Gasification Materials, AIChE 69th Annual Meeting, Chicago,
IL, November 28-December 2, 1976, p. 33.
20. Catchpole, J. P., and G. Fulford, Dimensionless Groups, I&EC:
46-60; March 1966.
21. Jonke, A. A., A Development Program on Pressurized Fluidized-Bed
Combustion, Monthly Progress Report, Argonne National Laboratories,
Argonne, IL: 29-35; June 1976, ANL/ES-CEN-F092.
22. Vaughn, D. A., et al., Fluidized Bed Combustion Industrial Applica-
tion Demonstration Project, Special technical report to ERDA on
Battelle's Multi-Solids Fluidized-Bed Combustion Process, Battelle
Laboratories, Columbus, OH, February 7, 1977, ERDA Contract
E(49-18)-2472, p. 20.
221
-------
23. Curran, G. P., et al., High-Temperature Desulfurization of Low-Btu
Gas, Formal report to EPA, No. 5, Project 550, Consolidation Coal
Company, Pittsburgh, PA, Series Period July 1, 1973-January 31,
1976. EPA 600/7-77-031, April 1977.
24. Mathur, K. B., and N. Epstein, Developments on Spouted Bed Techno-
logy, Canadian Journal of Chem. Engrg., 52(2): 129-45; April 1974.
25. Stanley, D. A., L. Y. Sadler, III, D. R. Brooks, and M. A. Schmarty,
Production of Submicron Silicon Carbide Powders by Attrition
Milling, 2nd International Conference on Fine Particles, Boston, MA,
October 7-11, 1973, pp 331-36.
222
-------
LIST OF SYMBOLS
2
A. - area of new fracture surface (cm )
b * exponent in Curran formula
GI = ratio of particle velocity to bubble velocity
C~ * efficiency of changing kinetic energy to surface energy
C» * ratio of chip thickness to bed depth
Cl - coefficient in equation (6), a measure of chip thickness (cm)
C, « ratio of disturbed bed cross-section area to bubble cross-
section area
C_ * coefficient in Gonzales and Otero equation (cm /s)
C = 1 Cl C2 C3 C4 te{luation <12)]
D » fluidized-bed diameter (cm)
D, - diameter of a spherical bubble (cm)
D * diameter of a bed particle
P
Ff - function or "function of" or "the function"
2
g - gravity acceleration (980 cm/s )
2
g « Newton's law conversion factor (lg«cm/dyne*s )
3 2
G = Gallileo number -D p (p - p ) g * y
a p g s g
h * total bed depth (cm)
-,k2,k3,k,,k - coefficients in regression model (varied)
L - total loss rate from bed by attrition [equation (18)] (g/min)
= R x bed mass
223
-------
m" - exponent in Kutyavina and Baskakov equation
m * exponent in Gonzales and Otero equation
m1 = exponent in equation (24)
M = mass of coarse bed solids (g)
M ,VL = value of M at the beginning and end of a time interval (g)
2
N = Bond number « g Z p * g a
DO S C
N = Strouhal number - RZ * (U - U -)
st mt
n = exponent in equation (6)
r = exponent in Table D2
R = attrition rate per unit mass of bed (g/g-s) = L * bed mass
2
S = specific surface of granular solids (cm /g)
t = time (s)
U = superficial gas velocity (cm/s)
U, -= bubble rise velocity (cm/s)
U = minimum fluidizing velocity (cm/s)
U * particle velocity (cm/s)
Z = distance measured downward from the bed surface (cm)
e = porosity of fluidized-bed dense phase
y = gas viscosity (g/cm-s)
3
p = density of solid (g/cm )
S
pg = density of gas (g/cm )
2
CT = solid fracture energy (erg/cm ) or (dyne/cm)
224
-------
APPENDIX E
3.5-cm FLUIDIZED-BED TEST SYSTEM
225
-------
3.5-cm FLUIDIZED-BED TEST SYSTEM
Purpose: To study attrition and sorption properties of fluidized-
bed solids
- Extent of attrition, or attrition rates of sorbents
at narrow temperatures, pressures, gas flow rates,
and gas compositions; comparison of attrition
tendencies of different sorbents
S atmospheres
- Sorption kinetics of sorbents in SCL or
Test Facility:
Flow diagram Figure 1
Reactor assembly and grid Figure 2
Pressure containment Figure 3
Reactor assembly Figure 4
Design Parameters:
Test cell:
Available grids:
Maximum pressure:
Maximum temperature:
Gases:
Gas flow:
Bed materials:
35 cm id
42 cm total depth
Perforated, sintered
1000 kPa, absolute
1000°C
N2, Air, C02, CO, H2, S02, H2$
Limestone, dolomite, sand, char
226
-------
l,AIK,F ^
•u
4tb.W OK-
c
VAlVf
Tb-tH
.471 --
1
1 1
1 1
"l
| 1
I 1
"I
1 1
"I
1 1
"
"Z
MIX
Ml —
«a.(l
».LL DIMENSIONS ARE CEMTlMtTCM
Figure El - Flow Diagram for the Attrition Reactor System
-------
Dwo. F386A07
Weld
600
Inconel 600
Reactor Shell
Distribution
Plate
Weld
Figure E2 - Reactor Assembly and Grid
228
-------
Figure E3 - Pressure Containment
229
RM-80600
-------
Figure E4 -• Reactor Assembly
230
RM-80601
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
. REPORT NO.
EPA-600/7-80-015a
3. RECIPIENT'S ACCESSION NO.
.T.TLE AND SUBTITLE Experimental/Engineering Support
'or EPA's FBC Program: Final Report
Volume 1. Sulfur Oxide Control
5. REPORT DATE
January 1980
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
N.H.Ulerich,-W. G.Vaux, R.A.Newby, and
D. L.Keairns
8. PERFORMING ORGANIZATION REPORT NO.
_. PERFORMING ORGANIZATION NAME AND ADDRESS
Westinghouse Research and Development Center
1310 Beulah Road
Pittsburgh, Pennsylvania 15235
1O. PROGRAM ELEMENT NO.
INE825
11. CONTRACT/GRANT NO.
68-02-2132
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; 12/75 - 12/78
14. SPONSORING AGENCY CODE
EPA/600/13
is.SUPPLEMENTARY NOTES J.ERL-RTP project officer is D. Bruce Henschel, Mail Drop 61,
919/541-2825. EPA-600/7-78-OT)5, -039, and -163 also relate to this work.
16. ABSTRACT ij-^g report gives results of an investigation of the desulfurization perfor-
mance and attrition behavior of limestone and dolomite sorbents for atmospheric and
pressurized fluidized-bed combustion (FBC) systems used with coal. It gives results
of experimental thermogravimetric analyses (TGAs) of the kinetics of SO2 capture
by sorbents, and discusses the further development and application of a kinetic model
for desulfurization, based on TGA results. It also gives results of a basic assess-
ment of sorbent attrition mechanisms in FBC, including some laboratory experimen-
tal tests. Some conclusions from this work are: (1) pressurized FBC systems can
achieve effective SO2 removal at high temperatures (1000 C) or high excess air (300%
without an increase in sorbent requirements over lower temperature/excess air
cases; (2) the agreement between actual FBC data and the TGA-based desulfurization
model has been further demonstrated, using data from both atmospheric and pres-
surized FBCs; and (3) sorbent attrition screening tests indicate that sorbent type and
FBC operating parameters will affect particle attrition. The report presents an
experimental-data-supported sorbent attrition model for the bubbling bed regime in
an FBC.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. cos AT I Field/Group
Pollution Dolomite
Combustion Sorbents
Fluidized Bed Processing
Coal Sulfur Dioxide
Desulfurization Kinetics
Limestone Mathematical Models
Pollution Control
Stationary Sources
13B
2 IB
13H,07A
21D
07D
08G
11G
07B
20K
12A
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
247
20. SECURITY CLASS (TMspage)
Unclassified
22. PRICE
EPA Form 2220-1 O-73)
231
------- |