Professional Documents
Culture Documents
Handbook of Fluidization and Fluid Particle Systems
Handbook of Fluidization and Fluid Particle Systems
FLUIDIZfiTION and
FLUID-PARTICLE SYSTEMS
edited by
Wen-Ching Yang
Siemens Westinghouse Power Corporation
Pittsburgh, Pennsylvania, U.S.A.
ISBN: 0-8247-0259-X
Headquarters
Marcel Dekker, Inc.
270 Madison Avenue, New York, NY 10016
tel: 212-696-9000; fax: 212-685-4540
The publisher offers discounts on this book when ordered in bulk quantities. For more information, write to Special Sales/
Professional Marketing at the headquarters address above.
Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopying, microfilming, and recording, or by any information storage and retrieval system, without permission
in writing from the publisher.
Consulting Editor
HEINZ HEINEMANN
Berkeley, California
1. Fluid Catalytic Cracking with Zeolite Catalysts, Paul B Venuto and E. Thomas Habib, Jr.
2. Ethylene: Keystone to the Petrochemical Industry, Ludwig Kniel, Olaf Winter, and Karl Stork
3. The Chemistry and Technology of Petroleum, James G Speight
4. The Desulfurization of Heavy Oils and Residua, James G. Speight
5. Catalysis of Organic Reactions, edited by William R. Moser
6 Acetylene-Based Chemicals from Coal and Other Natural Resources, Robert J. Tedeschi
7 Chemically Resistant Masonry, Walter Lee Sheppard, Jr.
8 Compressors and Expanders: Selection and Application for the Process Industry, Heinz P. Bloch, Joseph A.
Cameron, Frank M. Danowski, Jr, Ralph James, Jr, Judson S. Sweanngen, and Marilyn E. Weightman
9. Metering Pumps. Selection and Application, James P. Poynton
10. Hydrocarbons from Methanol, Clarence D Chang
11. Form Flotation: Theory and Applications, Ann N. Clarke and David J. Wilson
12. The Chemistry and Technology of Coal, James G. Speight
13. Pneumatic and Hydraulic Conveying of Solids, O. A Williams
14. Catalyst Manufacture: Laboratory and Commercial Preparations, Alvm B. Stiles
15. Characterization of Heterogeneous Catalysts, edited by Francis Delannay
16. BASIC Programs for Chemical Engineering Design, James H. Weber
17. Catalyst Poisoning, L. Louis Hegedus and Robert W McCabe
18. Catalysis of Organic Reactions, edited by John R Kosak
19 Adsorption Technology: A Step-by-Step Approach to Process Evaluation and Application, edited by Frank L.
Slejko
20 Deactivation and Poisoning of Catalysts, edited by Jacques Oudar and Henry Wise
21 Catalysis and Surface Science: Developments in Chemicals from Methanol, Hydrotreating of Hydrocarbons,
Catalyst Preparation, Monomers and Polymers, Photocatalysis and Photovoltaics, edited by Heinz Heinemann
and Gabor A Somorjai
22 Catalysis of Organic Reactions, edited by Robert L. Augustine
23 Modern Control Techniques for the Processing Industries, T. H. Tsai, J W. Lane, and C. S. Lin
24 Temperature-Programmed Reduction for Solid Materials Characterization, Alan Jones and Brian McNichol
25 Catalytic Cracking: Catalysts, Chemistry, and Kinetics, Bohdan W Wojciechowski and Avelino Corma
26. Chemical Reaction and Reactor Engineering, edited by J. J. Carberry and A. Varma
27 Filtration: Principles and Practices: Second Edition, edited by Michael J Matteson and Clyde Orr
28 Corrosion Mechanisms, edited by Florian Mansfeld
29 Catalysis and Surface Properties of Liquid Metals and Alloys, Yoshisada Ogino
30 Catalyst Deactivation, edited by Eugene E Petersen and Alexis T. Bell
31. Hydrogen Effects in Catalysis: Fundamentals and Practical Applications, edited by Zoltan Paal and P G. Menon
Re-Engineering the Chemical Processing Plant Process Intensification, edited by Andrzej Stankiewicz and
Jacob A Mouhjn
Every chemical engineer, whether a student or practicing, has looked up technical information in Perry’s Chemical
Engineering Handbook. Its compilation was one of the most important contributions to the chemical engineering
education and profession. After more than six decades, it remains one of the field’s most useful general-purpose
reference books. It was in this spirit of serving the profession that I undertook the task of compiling the Handbook
of Fluidization and Fluid-Particle Systems. Through future revisions and additions, I sincerely hope that this hand-
book will become an archivable reference volume for every practitioner in this field, spanning the boundary of
various disciplines. Fluidization and fluid-particle system engineering is being applied in industries as diverse as
basic and specialty chemicals, mineral processing, coal and biomass gasification and combustion for power gen-
eration, environmental technologies, resource recovery, FCC petroleum refining, pharmaceuticals, biotechnology,
cement, ceramics, and other solids handling and processing industries. The first focused handbook ever published in
this extended field, it collects all relevant and important information in a single volume. Both fundamentals and
applications are emphasized. Furthermore, all authors are internationally recognized practitioners in the area of
fluidization and fluid-particle systems.
This handbook contains 28 chapters and is authored by 34 internationally recognized experts from seven
countries; half of them are professors. Particle characterization and dynamics—important in all aspects of particle
production, manufacturing, handling, processing, and applications—are discussed in Chapter 1. Chapter 2 presents
the flow through fixed beds and summarizes packing characteristics of spherical and nonspherical particles, pres-
sure-drop correlations for flow through fixed beds, and heat and mass transfer. Bubbling fluidized beds are pre-
sented in detail in Chapter 3, which covers all important aspects including jetting phenomena and particle
segregation, topics not addressed extensively in other books on fluidization. Other important design considerations
are treated in separate chapters: elutriation and entrainment in Chapter 4, effect of temperature and pressure in
Chapter 5, gas distributor and plenum design in Chapter 6, effect of internal tubes and baffles in Chapter 7, attrition
in Chapter 8, and modeling in Chapter 9. Heat transfer (Chapter 10) and mass transfer (Chapter 11) are also
treated. The approaches for designing and scaling up fluidized bed reactors are elucidated in Chapter 12, ‘‘General
Approaches to Reactor Design,’’ and Chapter 13, ‘‘Fluidized Bed Scaleup.’’
Important industrial applications for fluidized bed reactors are also discussed, including fluid catalytic cracking
(Chapter 14), gasifiers and combustors (Chapter 15), chemical production and processing (Chapter 16), coating and
granulation (Chapter 17), and fluidized bed drying (Chapter 18).
The important variation of bubbling fluidized beds—the circulation fluidized beds—are discussed in detail in
Chapter 19. Chapter 20 summarizes other nonconventional fluidized beds, including spouted beds, recirculating
fluidized beds with a draft tube, jetting fluidized beds, and rotating fluidized beds. The solids handling, transport
and circulating devices are described in Chapter 21, ‘‘Standpipe and Nonmechanical Valves,’’ and Chapter 22,
Wen-Ching Yang
Preface
Contributors
Hsiaotao Bi Department of Chemical and Biological Engineering, University of British Columbia, Vancouver,
British Columbia, Canada
John C. Chen Department of Chemical Engineering, Lehigh University, Bethlehem, Pennsylvania, U.S.A.
Ye-Mon Chen Shell Global Solutions US, Houston, Texas, U.S.A.
Shiao-Hung Chiang Department of Chemical and Petroleum Engineering, University of Pittsburgh, Pittsburgh,
Pennsylvania, U.S.A.
Sakamon Devahastin Department of Food Engineering, King Mongkut’s University of Technology Thonburi,
Bangkok, Thailand
Norman Epstein Department of Chemical and Biological Engineering, University of British Columbia, Vancouver,
British Columbia, Canada
Liang-Shih Fan Department of Chemical Engineering, The Ohio State University, Columbus, Ohio, U.S.A.
Yuru Feng Department of Chemical and Petroleum Engineering, University of Pittsburgh, Pittsburgh,
Pennsylvania, U.S.A.
Leon R. Glicksman Departments of Architecture and Mechanical Engineering, Massachusetts Institute of
Technology, Cambridge, Massachusetts, U.S.A.
Mohammad Golriz Department of Applied Physics and Electronics, Umeå University, Umeå, Sweden
John R. Grace Department of Chemical and Biological Engineering, University of British Columbia, Vancouver,
British Columbia, Canada
Ernst-Ulrich Hartge Technical University Hamburg-Harburg, Hamburg, Germany
Daxin He Department of Chemical and Petroleum Engineering, University of Pittsburgh, Pittsburgh,
Pennsylvania, U.S.A.
Thomas C. Ho Department of Chemical Engineering, Lamar University, Beaumont, Texas, U.S.A.
Masayuki Horio Department of Chemical Engineering, Tokyo University of Agriculture and Technology, Tokyo,
Japan
Behzad Jazayeri Fluor Daniel, Inc., Aliso Viejo, California, U.S.A.
Current affiliation: Energy Engineering Department, Czestochowa Technical University, Czestochowa, Poland.
y Current affiliation: BMH Claudius Peters, Buxtehude, Germany.
Wen-Ching Yang
Siemens Westinghouse Power Corporation, Pittsburgh, Pennsylvania, U.S.A.
Drag Diameter
The drag diameter, dD , is defined as the diameter of a
sphere having the same resistance to motion as the
particle in a fluid of the same density and viscosity
and moving at the same velocity.
Perimeter Diameter
The perimeter diameter, dc , is the diameter of a circle
having the same perimeter as the projected outline of Figure 1 Illustration for projected area diameter, Feret dia-
the particle. meter, and Martin diameter.
Cylinder, dy 6¼ ly 6dy
dp ¼
4 þ 2dy =ly
Mixed sizes 1
dp ¼ P
ðxi =dpi Þ
i
1.1.2 Definitions of Particle Shape For a true sphere, the sphericity is thus equal to 1. For
nonspherical particles, the sphericity is always less than
Natural and man-made solid particles occur in almost
1. The drawback of the sphericity is that it is difficult to
any imaginable shape, and most particles of practical
obtain the surface area of an irregular particle and thus
interest are irregular in shape. A variety of empirical
it is difficult to determine f directly. Usually the more
factors have been proposed to describe nonspherical
the aspect ratio departs from unity, the lower the
shapes of particles. These empirical descriptions of
sphericity. The sphericity, first introduced as a measure
particle shape are usually provided by identifying two
of particle shape, was subsequently claimed to be use-
characteristic parameters from the following four: (1)
ful for correlating drag coefficient (Wadell, 1934).
volume of the particle, (2) surface area of the particle,
There is some theoretical justification for the use of
(3) projected area of the particle, and (4) projected
sphericity as a correlating parameter for creeping
perimeter of the particle. The projected area and
flow past bodies whose geometric proportions resem-
perimeter must also be determined normal to some
ble a sphere. But for other circumstances its use is
specified axis. For axisymmetric bodies, the reference
purely empirical (Clift et al., 1978). Leva (1959) and
direction is usually taken to be parallel or normal to
Subramanian and Arunachalam (1980) suggested
the axis of symmetry.
experimental methods using the Ergun equation for
All proposed shape factors to date are open to criti-
evaluation of the sphericity. This methodology will
cism, because a range of bodies with different shapes
be discussed when Ergun equation is introduced in
may have the same shape factor. This is really inevit-
Chapter 2, ‘‘Flow through fixed beds.’’ For regularly
able if complex shapes are to be described only by a
shaped solids, the sphericities can be calculated from
single parameter. Thus in selecting a particular shape
Eq. (5), and they are presented in Table 2a. As for
factor for application, care must be taken to assure its
commonly occurring nonspherical particles, their
relevance.
sphericities are summarized in Table 2b.
Sphericity
Circularity
Wadell (1933) proposed the ‘‘degree of true sphericity’’
Wadell (1933) also introduced the ‘‘degree of circular-
be defined as
ity’’, defined as
Surface area of volume equivalent sphere
f¼ Circumference of circle having same cross-
Surface area of particle 6¼ sectional area as the particle
2
d d Actual perimeter of the cross-section
¼ v ¼ sv ð5Þ
ds dv ð6Þ
Unlike the sphericity, the circularity can be determined dependence of flow behavior on particle orientation
more easily experimentally from microscopic or photo- (Cliff et al., 1978).
graphic observation. For an axisymmetric particle
projected parallel to its axis, 6 is equal to unity. Use
Operational Sphericity and Circularity
of 6 is only justified on empirical grounds, but it has
Since the sphericity and circularity are so difficult to
the potential advantage of allowing correlation of the
determine for irregular particles, Wadell (1933) pro-
posed that f and 6 be approximated by ‘‘operational
sphericity and circularity:’’
Table 2b Sphericities of Commonly 0 11=3
Occurring Nonspherical Particles
B Volume of particle C
Material Sphericity fop ¼ @ A
Volume of the smallest circumscribing
sphere
Sand
Round sand 0.86 ð7Þ
Sharp sand 0.66 0 11=2
Crushed sandstone 0.8–0.9 Projected area of particle
Coal 6op ¼ @ A
Area of the smallest circumscribing
Pulverized coal 0.73
circle
Crushed coal 0.63–0.75
Activated carbon 0.70–0.90 ð8Þ
Mica flakes 0.28
Fischer–Tropsch catalyst 0.58 For the ellipsoids, the operational sphericity, fop , can
Common salt 0.84 be expressed as
Crushed glass 0.65
Silica gels 0.70–0.90 fop ¼ ðe1 e2 Þ1=3 ð9Þ
Tungsten powder 0.89
Sillimanite 0.75
For the rounded particles, it can be approximated by
Wheat 0.85
Eq. (9). The e1 and e2 are called the flatness ratio and
Source: Adapted from Geldart (1986). the elongation ratio, respectively, and are defined as
2 PARTICLE CHARACTERIZATION
TECHNIQUES
diameter is usually fairly constant for the same mate- between 300 to 500, needs to be measured to minimize
rial. For Portland cement, it is about 1.2; for ground statistical error.
quartz and ground glass, approximately 1.3 (Herdan,
1960).
2.1.3 Gravity and Centrifugal Sedimentation
The microscopic measurement technique is most
suitable for particles relatively uniform in size and The falling speeds of particles in a viscous fluid under
granular in shape, because a large number of particles, the influence of gravity are used to measure the particle
size in the gravity sedimentation technique. The mea- dp
V1 ¼ Vm 1 þ 2:4 ð17Þ
sured speeds are then converted to Stokes diameters by D
applying the Stokes equation [Eq. (4)], assuming that
the particles are all spherical in shape. Since the irre- According to Eq. (17), even if the column is 50 times
gularly shaped particles fall with different orientations the diameter of the particle, there is still a 5% reduc-
in the vertical direction and thus different settling velo- tion in the falling speed of the particle.
cities, similar irregularly shaped particles can have a Two basic suspension systems, the line start system
range of Stoke diameters. Disturbances caused by the (or the two-layer sedimentation system) and the initi-
presence of other particles, the concentration effect, ally homogeneous system are employed in the gravity
can also be important. Recommendations for the sedimentation technique. In the line start system, a thin
upper limit particle concentration to be used during layer of particles is placed at the top of the sedimenta-
the sedimentation analysis varies from 0.01 to 3.0% tion column, and its settling behavior is analyzed by
(Kaye, 1981). An upper limit of 0.05% was recom- different techniques. In the initially homogeneous sys-
mended by Kaye, and in certain practical cases up to tem, the column is homogenized first and its settling
0.2% by volume is permissible. The probability of pattern is subsequently studied.
forming clusters increases with particle suspension Classical techniques for measuring the sedimenta-
concentration. The clusters tend to fall at a higher tion behavior include taking samples with a pipette,
speed and thus introduce measurement error. measurement of height of sediment layer at the bot-
The hindering effect of the containing wall on the tom, and use of balance pan to measure the weight of
falling speed of the particles cannot be ignored either. settled particles. Modern sedimentometers make use of
For a spherical particle, the effect can be expressed by the diffraction pattern of a light beam, the power loss
the Landenburg equation as of an x-ray, or a Doppler shift of a laser beam. The
Silicon
Aluminum Barium Boron nitride
Sensing technique oxide titanate nitride whisker
DR10
Electrical sensing zone 1.00 1.00 1.00 1.00
Laser diffraction and scattering 0.61 0.77 0.94 0.61
X-ray sedimentation 0.82 1.10 — 0.69
Photosedimentation 0.82 0.95 0.58 0.75
Light attenuation 1.00 1.15 0.96 0.96
DR50
Electrical sensing zone 1.00 1.00 1.00 1.00
Laser diffraction and scattering 0.97 1.11 1.32 1.41
X-ray sedimentation 0.84 1.05 — 0.86
Photosedimentation 0.78 0.96 0.74 1.05
Light attenuation 1.33 1.20 1.11 1.04
DR90
Electrical sensing zone 1.00 1.00 1.00 1.00
Laser diffraction and scattering 1.15 1.33 1.56 2.32
X-ray sedimentation 0.90 1.14 — 1.01
Photosedimentation 1.01 1.17 1.53 2.43
Light attenuation 1.20 1.46 1.78 0.98
ometer, x-ray sedimentation, and so on. Conversion The values of the HGI usually range from 15 to 140.
factors are available to convert particle size distribu- The higher the HGI, the higher is the grindability of
tions obtained by one method to that measured by the material. The HGI has been found to correlate
another technique. Austin (1998) suggested a technique with the attrition characteristics of the particles in
to perform this correction and provided an equation fluidized beds and in pneumatic transport lines
for conversion between laser diffractometer and x-ray (Davuluri and Knowlton, 1998). The HGI does not
sedigraph measurements. directly relate to hardness. For example, some materi-
The applicable particle size ranges for various par- als such as plastics are difficult to grind.
ticle sizing techniques are summarized in Table 6.
Sieving
Dry > 10 Sieve diameter
Wet 2–500
Microscopic examination
Optical 1.0–100 Length, projected area,
Electronic 0.01–500 statistical diameters
Zone sensing
Resistivity 0.6–1200 Volume
Optical 1.0–800
Elutriation
Laminar flow 3–75 Stokes diameter
Cyclone 8–50
Gravity sedimentation
Pipette and hydrometer 1–100
Photoextinction 0.5–100 Stokes diameter
X-ray 0.1–130
Centrifugal sedimentation
Mass accumulation 0.5–25
Photoextinction 0.05–100 Stokes diameter
X-ray 0.1–5
Centrifugal classification 0.5–50
Gas permeability 0.1–40
Gas adsorption 0.005–50
Cascade impaction 0.05–30
stone. For a more recent review on attrition, see under controlled conditions. The weight loss of the
Werther and Reppenhagen (1999). coupon is an indication of the erosiveness of the parti-
cular coal and the potential damage to the processing
2.3.3 Abrasiveness Index and handling equipment, and other boiler components.
In pulverized coal combustion, the abrasiveness of the
particles severely limits the life of the pulverizer grind-
3 FLUID DYNAMICS OF A SINGLE PARTICLE
ing elements. The Abrasiveness Index in this applica-
tion is usually determined by the Yancey–Geer Price
For a particle moving in a fluid, the force acting on the
apparatus (Babcock and Wilcox, 1992). In this test,
surface of a particle depends only on the flow of the
four metal test coupons attached to a rotating shaft
fluid in its immediate vicinity. For the simplest case, let
are rotated at 1440 rpm for a total of 12,000 revolu-
us consider a single particle moving at a velocity Ur
tions in contact with a sample of 6350 mm 0
relative to its immediate fluid around the particle. It is
(0:25 in: 0) coal. The relative Abrasiveness Index is
also assumed that the fluid is newtonian and that the
then calculated from the weight loss of the coupons.
Ur is constant. The fluid dynamic parameters can then
Babcock and Wilcox has estimated the wear of full-
be evaluated as follows.
scale pulverizers on the basis of the Yancey–Geer
Price Index.
3.1 Definition of Particle Drag Coefficient
2.3.4 Erosiveness Index
The drag coefficient is defined as the ratio of the force
Babcock and Wilcox (1992) has also developed a on the particle and the fluid dynamic pressure caused
method of quantifying the erosiveness of coal by sub- by the fluid times the area projected by the particles, as
jecting a steel coupon to a stream of pulverized coal shown in Eq. (21) and Fig. 8.
Figure 6 Schematic of attrition apparatus employing the 3.1.1 The Stokes Regime
solids impaction principle.
24
CD ¼ for ðReÞp < 0:2 ð23Þ
ðReÞp
F
CD ¼ ð21Þ
ð1=2Þrf Ur2 Ap The Stokes law regime is also commonly known as the
creeping flow regime. In this regime, the viscosity of
or
the fluid is dominating.
1
F ¼ CD rf Ur2 Ap ð22Þ
2 3.1.2 The Intermediate Regime
The drag coefficient, CD , is a function of particle’s
Reynolds number, ðReÞp ¼ Ur dp rf =m, only. There are CD ¼ f bðReÞp c for 0:2 < ðReÞp < 500 ð24Þ
Equating the drag force and the gravitational force for 3.6 Empirical Drag Coefficient Expression
a single spherical particle, we have
The preferred correlations to be used in different
p pdp3 ranges of particle Reynolds numbers were recom-
Fs ¼ CD rf Ur2 dp2 ¼ rp rf g ð29Þ
8 6 mended by Clift et al. (1978) and are summarized in
Table 8. In 1986, Turton and Levenspiel proposed a
or
single correlation applicable for the complete range of
4 d p r p rf g Reynolds numbers and considerably simplified the cal-
CD ¼ ð30Þ culation of the single-particle drag coefficient.
3 rf Ur2
The Turton and Levenspiel (1986) equation is
or
24 h i
CD ¼ 1 þ 0:173ðReÞ0:657
dp2 rp r f g ðReÞp p
Ur ¼ Ut ¼ ð31Þ ð34Þ
18m 0:413
þ
Equation (4) for the Stokes diameter can be derived 1 þ 16:300ðReÞp1:09
from Eq. (31).
Haider and Levenspiel (1989) subsequently improved
3.4 Wall Effect the equation to cover the nonspherical particles and
proposed
When the fluid is of finite extent, there are two effects.
The fluid streamlines around the particle impinge on
the walls and reflect back on the particle, causing
Table 7 Summary of Theoretical Expressions for the
increasing drag. Also, since the fluid is stationary at
Drag Force ½ðReÞp 1
a finite distance from the particle, this distorts the
flow pattern and increases drag. The simple correction Stokes (1851)
can be expressed as Fs ¼ 3pdp mUr
Oseen (1910) n o
dp 3
Fs ¼ 3pdp mUr 1 þ kc ð32Þ F ¼ Fs 1 þ ðReÞp þ O ðReÞ2p
Lw 16
Proudman
and Pearson (1957)
ðReÞp n o
where Lw ¼ the distance from the center of the particle 3 9
F ¼ Fs 1 þ ðReÞp þ ðReÞ2p ln þ O ðReÞ2p
to the walls, kc ¼ 0:563 (for a single wall), kc ¼ 1:004 16 160 2
(for two walls), and kc ¼ 2:104 (for a circular cylinder).
24 h i
CD ¼ 1 þ 0:1315ðReÞð0:820:05wÞ
ðReÞp p
20 ðReÞp 260 CD ðReÞp
log10 1 ¼ 0:7133 þ 0:6305w
24
24 h i
CD ¼ 1 þ 0:1935ðReÞ0:6305
ðReÞp p
1:5 103 ðReÞp 1:2 104 log10 CD ¼ 2:4571 þ 2:5558w 0:9295w2 þ 0:1049w3
24 h
i
no fully satisfactory method is available for correlating
CD ¼ 1 þ 8:1716e4:0655f ðReÞp0:0964þ0:5565f
ðReÞp the drag.
Pettyjohn and Christiansen (1948) determined the
73:69 e5:0748f ðReÞp free-settling rates of isometric particles of the following
þ shapes and sphericities: spheres (f ¼ 1), cube octa-
ðReÞp þ 5:378e6:2122f
hedron (f ¼ 0:906Þ, octahedron (f ¼ 0:846Þ, cube
ð35Þ (f ¼ 0:806), and tetrahedron (f ¼ 0:670). Their results
suggest that the correction factor should be
For spherical particles, Eq. (35) reduces to
f
24 0:4607ðReÞp K ¼ 0:8431 log ðReÞt < 0:05; ð37Þ
CD ¼ þ 3:3643ðReÞ0:3471 þ 0:065
ðReÞp p
ðReÞp þ 2682:5 0:67 < f < 1
CD here is based on the cross-sectional area of the 4.1 Equation by Haider and Levenspiel (1989)
volume-equivalent sphere. The terminal velocity, Ut ,
can be calculated by From Eq. (30), the terminal velocity for a single
" #12 spherical particle can be obtained as
1=36 g rp rf dv sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Ut ¼ 0:49l ð40Þ
rf ð1:08 fÞ 4 dp ðrp rf Þg
U r ¼ Ut ¼ ð44Þ
3 rf CD
3.8 Drag Coefficient for Particles with Density By combining Eq. (44) and the recommended drag
Lighter than the Surrounding Fluid coefficient correlations in Table 8 or drag coefficient
equations proposed by Turton and Levenspiel (1986),
All the above discussions are for particles with density Eq. (34), and Haider and Levenspiel (1989), Eq. (35),
larger than the surrounding fluid. For particles with the terminal velocity can be calculated. Haider and
density smaller than the surrounding fluid, it has Levenspiel further suggested an approximate method
long been assumed that the free rising velocity is for direct evaluation of the terminal velocity by defin-
governed by the same equations, the only difference ing a dimensionless particle size, dp , and a dimension-
being that of particle movement direction. However, less particle velocity, U , by
Y X Polynomial coefficients
Calculation of the terminal velocity of a porous sphere Equation (60) shows that a permeable sphere with the
is useful and important in applications in water treat- same diameter and bulk density as the impermeable
ment where settling velocities of a floc or an aggregate sphere will have a higher terminal velocity than that
are estimated. It is also important in estimation of of the impermeable sphere. At higher Reynolds num-
terminal velocities of clusters in fluidized bed applica- ber, the experiments by Masliyah and Polikar (1980)
tions. The terminal velocity of a porous sphere can be suggested the following equations:
quite different from that of an impermeable sphere. For 15 < b < 33
Theoretical studies of settling velocity of porous 24 h i
spheres were conducted by Sutherland and Tan CD ¼ 1 þ 0:1315ðReÞpð0:820:05wÞ
ðReÞp ð64Þ
(1970), Ooms et al. (1970), Neale et al. (1973),
Epstein and Neale (1974), and Matsumoto and 0:1 < ðReÞp 7
Suganuma (1977). The terminal velocity of porous
and
spheres was also experimentally measured by
Masliyah and Polikar (1980). In the limiting case of a 24 h i
CD ¼ 1 þ 0:0853ðReÞpð1:0930:105wÞ
very low Reynolds number, Neale et al. (1973) arrived ðReÞp ð65Þ
at the following equation for the ratio of the resistance
experienced by a porous (or permeable) sphere to an 7 < ðReÞp < 120
equivalent impermeable sphere. An equivalent where w ¼ log10 ðReÞp .
impermeable sphere is defined to be a sphere having This confirms the theoretical study of Neale et al.
the same diameter and bulk density of the permeable (1973) that the drag experienced by a porous sphere
sphere. at low Reynolds numbers is less than that for an
2b2 ½1 ðtanh bÞ=b impermeable sphere of similar diameter and bulky den-
¼ ð58Þ sity. The effect of inertia at high Reynolds numbers is
2b2 þ 3½1 ðtanh bÞ=b
higher for a porous sphere than for an impermeable
where b is the normalized sphere radius expressed by sphere of similar diameter and bulk density.
The harmonic mean is actually related to the average Because of its mathematical properties, the standard
spherical particle corresponding to the particle surface deviation s is almost exclusively used to measure the
per unit weight. Mathematically, the harmonic mean is dispersion of the particle size distribution. When the
similar to the volume–surface mean or Sauter mean. skewed particle size distribution shown in Fig. 9 is
replotted using the logarithm of the particle size, the
5.2 Statistical Characterization of Particles with a skewed curve is transformed into a symmetrical bell-
Size Distribution shaped curve as shown in Fig. 10. This transformation
is of great significance and importance in that a sym-
Statistically, the particle size distribution can be metrical bell-shaped distribution is amenable to all the
characterized by three properties: mode, median, and statistical procedures developed for the normal or
mean. The mode is the value that occurs most gaussian distribution.
frequently. It is a value seldom used for describing In the log-normal particle size distribution, the
particle size distribution. The average or arithmetic mean, median, and mode coincide and have an identical
mean diameter, dav , is affected by all values actually value. This single value is called the geometric median
observed and thus is influenced greatly by extreme particle size, dg , and the measure of dispersion, the geo-
values. The median particle size, d1=2 , is the size that metric standard deviation, sg . Thus the log-normal
divides the frequency distribution into two equal areas. particle size distribution can be described completely
In practical application, the size distribution of a typi- by these two characteristic values. To determine
cal dust is typically skewed to the right, i.e., skewed to whether the particles have a distribution close to log-
the larger particle size. The central tendency of a normal distribution, the particle cumulative frequency
skewed frequency distribution is more adequately data can be plotted on a logarithmic probability graph
represented by the median rather than by the mean paper. If the particle size distribution is log-normal, a
(see Fig. 9). Mathematically, the relationships among straight line will result. The geometric median particle
the mean, median, and mode diameter can be size is the 50% value of the distribution as shown in
expressed as Fig. 11. The geometric standard deviation is equal to
where
P
dc z dc
f ¼ ; z ¼ lnðdpi Þ; z ¼ P
i
ð80Þ
dz dc
i
Two important characteristics of powder rheology are Table 10 Typical Angles of Repose and Angles
the angle of repose and the angle of internal friction. of Internal Friction
Simple devices can be constructed to measure both. Angle of
Figure 12 depicts a simple two-dimensional bed with Angle of repose internal friction
transparent walls and a small orifice at the bottom of Material degree degree
the bed. After filling the bed with the powder to be
examined, the powder is allowed to flow out of the FCC catalyst 32 79
test device to the surface of the test stand. The angle Sand 36 64
Resin 29 82
of the powder-free surface measured from the flat sur-
Wet ashes 50 —
face of the test stand, the angle b in Fig. 12, is called
Wheat — 55
the angle of repose. This angle is an intrinsic charac- Oats 21 —
teristic of the powder and should be close to a constant
Wen-Ching Yang
Siemens Westinghouse Power Corporation, Pittsburgh, Pennsylvania, U.S.A.
Fixed beds, also called packed beds, play a vital role in angle and the triangular or the simple rhombic layer of
chemical processes. Their simplicity induced applica- an angle of 60 (see Fig. 1). The vertical stacking of
tions in many unit operations such as adsorption, dry- these layers yields six possible regular packings. For
ing, filtration, dust collection, and other catalytic and the six possible regular packings of spherical particles,
noncatalytic reactors. The primary operating cost for a the voidage is only a function of packing arrangement.
fixed bed is the pressure drop through the packed bed The voidage is also independent of the particle size.
of solids. Thus to understand the design and operation The six arrangements are cubic, two orientations of
of a packed bed requires study of packing character- orthorhombic, tetragonal-spheroidal, and two orienta-
istics of particles and their effect on pressure drop tions of rhombohedral. They are graphically presented
through the bed. in Fig. 2. Formulae for calculating the voidage of these
six arrangements have been developed and the packing
characteristics of these ordered arrangements are sum-
1 PACKING CHARACTERISTICS OF marized in Table 1 (Herdan, 1960). The closest packing
MONOSIZED SPHERES is the rhombohedral with a voidage of 0.2595, and the
loosest, the cubic with a voidage of 0.4764. The stabi-
The simplest case of a packing system is a bed that lity of the systems increases as the voidage decreases.
consists of only uniform and regular monosized sphe- The systems tend towards the orthorhombic state,
rical particles. For this simple system, the voidage of especially if a mechanical disturbance such as vibration
an ordered arrangement of monosized spheres can be is applied to the systems. The voidage of a bed of
derived satisfactorily by mathematical considerations. monosized particles after prolonged shaking is usually
Random packing of monosized spheres is more com- about 0.395, approaching the characteristic of an
plicated and can be described mathematically through eight-point contact packing of orthorhombic.
a coordination number. However, the voidage deter-
mination of random packing of monosized spheres 1.2 Random Packing of Uniform Monosize Spheres
should be done through experiments.
Random packings of uniform monosize spheres are
1.1 Regular Packing of Uniform-Sized Spheres created by irregular and random arrangements of par-
ticles. According to Scott (1960), two reproducible
For convenience, the regular packing can be consid- states of random packing, dense random packing and
ered to be constructed from regular layers and rows. loose random packing, can be experimentally created.
The two basic layers are the square layer with a 90 The dense random packing can be developed by
Figure 2 Six possible arrangements of monosize spheres—one cubic (Figure 2a), two orthorhombic (Figures 2b and 2c), one
tetragonal–spheroidal (Figure 2d), two rhombohedral (Figure 2e).
pouring the spheres into cylindrical containers and A parameter called the layer spacing, bdp , where dp is
then shaking for several minutes. The loose random the diameter of the sphere, has also been used to char-
packing is created by tipping the container horizontally acterize the packing. For most common packings, b is
and rotating slowly about its axis, and then returning between 0.707 and 1.0, the limits corresponding to the
slowly to its vertical position. The dense random pack- rhombohedral packing and cubic packing, respectively.
ing density obtained by several authors ranged from Though the layer spacing has little physical meaning in
0.61 to 0.63, equivalent to a voidage of 0.39 and 0.37, random packing, the concept is useful for describing
corresponding closely to orthorhombic packing with a the packing. It is actually related to the bed voidage in
voidage of 0.3954. The loose random packing obtained the following way:
has a voidage of 0.40 to 0.42, corresponding to a pack- rffiffiffi 1=3
ing between the cubic packing (voidage ¼ 0:4764) and 2 p
b¼ pffiffiffi ð2Þ
orthorhombic packing. Haughey and Beveridge (1969) 3 3 2ð1 eÞ
classified the packing into four different modes: very
loose random packing, loose random packing, poured
random packing, and close random packing. The loose 1.4 Specific Surface Area of the Bed
random packing and poured random packing corre-
spond to Scott’s loose and dense random packings. The specific surface area of the bed is defined as the
The very loose random packing corresponds to the ratio of total particle surface area to the total bed
state where the bed is first fluidized and then the gas volume. Since the number of spheres per unit volume
is slowly reduced until it is below the minimum fluid- of bed is
ization. The bed so formed usually has a voidage of
about 0.44. The close random packing is formed by 6ð1 eÞ
Np ¼ ð3Þ
vigorously shaking or vibrating the bed, and then the pdp3
voidage is usually approaching 0.359 to 0.375.
and
n
ds
bs ¼ ð15Þ
db
Figure 3 Packed bed of minimum voidage for binary par- n equals 1=3 for crushed stones and 1=2 for round
ticle mixtures. sand.
An approach employing the structural ratio to have also been studied by Ouchiyama and Tanaka
predict the voidage of binary particle mixtures for (1989) and by Hoffmann and Finkers (1995). The
both spherical and nonspherical particles is also pro- packing characteristics of spheres of unequal sizes
posed recently by Finkers and Hoffmann (1998). The have also been investigated (Herdan, 1960). In rhom-
structural ratio based on packing of spherical and non- bohedral packings of different spherical particle sizes,
spherical particles is defined as the voidage may be reduced to less than 0.15.
1 3
eb 1 r p
rstr ¼ ð24Þ 3 CRITICAL RATIO OF ENTRANCE AND
1 es
CRITICAL RATIO OF OCCUPATION IN
The structural ratio is then used to calculate the BINARY SYSTEMS
empirical constant G for use in Eq. (16).
When the monosized spheres are arranged in normal
G ¼ rkstr þ 1 ek
b ð25Þ
loosest square or tightest rhombohedral packings,
A value of 0:63 was suggested for k in Eq. (25). The there are critical smaller spheres, which can pass
proposed approach is good for both spherical and through the openings formed by the larger monosized
nonspherical particles. For binary particles with size spheres, as shown in Figs. 6 and 7. These critical dia-
distribution in each fraction, the approach did not meter ratios are called the ‘‘critical ratio of entrance.’’
fare as well. By changing the value k to 0:345, the
proposed approach gave excellent results for data by
Sohn and Moreland (1968). It was suggested that k
relates to the particle size distribution in each fraction
to provide an even more general equation of particle
packing in a packed bed.
Packings of ternary and quaternary mixtures of
solid particles, considerably more complex systems, Figure 6 Critical ratio of entrance for square packing.
Particle of pellet Normal charge Dense packed Multiplier for small tubes
Voidage
where fc is an equivalent sphericity or aspect ratio Based on the work of Zou and Yu (1996), the
defined as the ratio of the maximum length to the Hausner ratio (HR) and the sphericity have the rela-
diameter of the circle having the same projected area tionship
normal to the maximum length, and fd , a sphericity
defined as the ratio of the shortest length to the dia- HR ¼ 1:478
100:136f ð39Þ
meter of the circle having the same projected area.
4.1 Relationship Between the Hausner Ratio and the 5 FACTORS AFFECTING PACKING DENSITY
Sphericity IN PRACTICE
The Hausner ratio is defined as the ratio of tapped Particles, containers, and filling and handling methods
density to loose density, it is a measurement of the contribute to packing density in practice. The impor-
compressibility and cohesiveness of the powder. tant factors can be summarized as
Particle shape, particle size, particle size distribu- dp dp
ew ¼ 1 þ 0:6 e 0:6 ð40Þ
tion, particle coefficient of restitution, particle D D
surface properties (friction)
dp dp
Container shape, container size, container surface ec ¼ ½e þ 0:3ð1 eÞ 1 þ 0:6 0:6 ð41Þ
properties (friction) D D
Deposition method, deposition intensity, velocity of
where e is the average voidage experimentally mea-
particle deposition
sured or the ratio of void volume to the total volume
Vibratory compaction, pressure compaction
of the packed bed. When the ratio of particle diameter
to the bed diameter is less than 0.02, the correction for
wall effect becomes negligible.
6 CORRELATIONS FOR FIXED BED BULK Figure 9 is a plot obtained by Benenati and
VOIDAGE Brosilow (1962) in a bed of spheres showing the radial
oscillations of voidage away from the containing
The voidage is the parameter most frequently cylindrical wall. The observation was confirmed by
employed to characterize the pressure drop in a fixed Propster and Szekely (1977). Experiments performed
bed. For accurate determination, experimental tech- by Goodling et al. (1983) also indicated that for
nique such as the water displacement is usually used. uniformly sized spherical particles, the oscillations in
In the application of catalytic reactors, the catalysis voidage can be up to 5 particle diameters from the
packing is important not only for pressure drop but wall. The oscillations in voidage are down to 2 to 3
also for heat and mass transfer. Furnas (1929) was diameters from the wall, for a mixture of two spherical
probably the first to study the effect of wall on the sizes, and down to 1 particle diameter from the wall for
packing of the particles. He proposed the following a mixture of three particle sizes.
correlations for voidage close to the wall and at the Propster and Szekely (1977) also found that a
core. marked local minimum in voidage existed in the inter-
facial region when small particles were placed on top of have correlations for fixed bed bulk voidage for com-
a bed of larger particles because of penetration of small monly used catalysts such as spheres, cylinders, and
particles into the interstices of large particles. rings. The following correlations are those proposed
However, when the large particles were placed on top by Dixon (1988). A more comprehensive review was
of the smaller particles, the local minimum in voidage conducted by Haughey and Beveridge (1969).
was not as pronounced. When the particle ratio of the For spherical particles
lower layer to the upper layer is less than 2, relatively 2
little particle penetration was observed. Ready pene- dp dp dp
e ¼ 0:4 þ 0:05 þ 0:412 0:5 ð42Þ
tration of small particles could be observed when the D D D
ratio equals 6, corresponding to a critical ratio of
dp dp
entrance of 6.464 for the tightest rhombohedral pack- e ¼ 0:528 þ 2:464 0:5 0:5 0:536
D D
ing (see Sec. 3). This observation will have practical
application in the iron blast furnace or cupola where ð43Þ
alternate layers of coke and iron ore of different 3 0:5
particle sizes are charged into the reactors. dp dp dp
e ¼ 1 0:667 2 1 0:536
Fixed beds of very low tube-to–particle diameter D D D
ratio have also been proposed and studied. For these ð44Þ
reactors, the effect of wall and particle shape on bulk
voidage becomes important. It will be convenient to For full cylinders
These dimensionless groups have been used ever since At low Reynolds numbers, where the pressure drop
as the basis for the pipe flow analogy in almost all does not depend on the fluid density, n ¼ 1, and
packed bed correlations.
From dimensionless analysis, the pressure drop ð1 eÞ2
through a packed bed of particles can be obtained as f / ð68Þ
e3 ðReÞp
n
P 2n n1 U
/ m rf Dn3 ð59Þ At high Reynolds numbers where the viscosity is not
L e
important, n ¼ 2, and
The interstitial fluid velocity in the average direction of
fluid motion is used here. The D is a length analogous ð1 eÞ
f / ð69Þ
to the hydraulic radius of a conduit and is defined as e3
Mean cross-sectional area of flow Because of the large dependence on the voidage, a 30-
D¼ channels through bed
to-70-fold increase in f is predicted as the voidage
Mean wetted perimeter of flow channels changes from 0.3 to 0.7 in the turbulent and viscous
ð60Þ ranges, respectively. The expression found above for f
at low Reynolds numbers has been experimentally
Multiplying the numerator and the denominator by L,
proved to be correct. At high Reynolds numbers,
the height of the bed, we have
however, the expression seems to predict a somewhat
ðTotal bed volume) ðvoidage) stronger dependence of the friction factor on the
D¼ ð61Þ
ðTotal bed surfaceÞ voidage than is actually found experimentally. For a
single particle, voidage becomes 1 by definition, and
or the Blake’s Reynolds number approaches infinity.
e edp This is not surprising, since the Blake analogy based
D¼ ¼ ð62Þ on a capillary flow analogy breaks down in this
S 6ð1 eÞ
range, because such an analogy has no physical mean-
where S is the total surface of solids per unit bed ing.
volume assuming spherical particles, it can be
expressed as
7.3 The Brownell and Kats Correlation
6ð1 eÞ
S¼ ð63Þ
dp Another correlation incorporating the voidage is that
by Brownell and Katz (1947). They introduced a
Equation (63) is essentially similar to Eq. (4).
Reynolds number defined as
Substituting into Eq. (59), we obtain
P 3n Udp rf
n ð1 eÞ ðReÞpe ¼ ð70Þ
/ m2n rn1
f U dpn3 ð64Þ mem
L e3
For conventional pressure drop through a packed bed where the exponent m depends on the ratio of the
of particles, n ¼ 2. Equation (64) becomes sphericity to the porosity and ranges from 2 to 20.
P 2f rf U 2 The use of dimensionless analysis in correlating the
¼ ð65Þ data is justified as long as the identified dimensionless
L gdp
groups represent true similarity. The Brownell and
The friction factor, f , is proportional to Kats correlation clearly does not represent true simi-
larity, because it predicts that the Reynolds number
ð1 eÞ3n decreases with increasing voidage, contradicting the
f / m2n rn2
f dpn2 U n2 ð66Þ
e3 experimental findings.
Figure 11 Radial gas velocity profiles in a 50.8 mm and 101.6 mm diameter bed. (Adapted from Schwarz and Smith, 1953.)
9.1 Rase Heat Transfer Correlations were TR ¼ temperature at inside radius of the vessel.
For spherical particle packing,
Based on the recommendations by Rase (1990), the
hw dp
heat and mass transfer correlations to be used for ¼ 0:19ðReÞ0:79
p Pr
0:33
where
9.1.4 Effective Radial Thermal Conductivity
rf Cpg dp U
q @T ðPeÞd ¼ ð114Þ
¼ kg ð107Þ kg
Ai @r r¼R
hw dp e dp Urf Equation (113) indicates that the heat transfer between
¼ 0:27 500 < < 6000 the particles and the percolating fluid is very fast.
kg 1 e mð1 eÞ
Assuming e ¼ 0:4 and Lp ¼ ndp and substituting into
dp Eq. (113) we have
and 0:05 < < 0:15 ð108Þ
D
n ¼ 0:14ðPeÞd ð115Þ
9.2 Heat Transfer Correlations Recommended by
For ðPeÞd < 10, Eq. (115) implies that just one particle
Molerus and Wirth (1997)
layer is enough to heat the percolating gas to the sur-
face temperature of the particles.
In a recent book by Molerus and Wirth (1997), the
recommended heat transfer correlations for packed
beds can be summarized as follows. For fully devel- 9.3 Analytical Models for Heat Transfer with
oped laminar flow, an approximation formula for Immersed Surfaces
the mean Nusselt number, derived from the pipe flow
analogy, was proposed as When the heat removal is entirely due to the flowing
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi gas and there is no convective particle movement,
hDp Dp Gabor (1970) proposed the following simple model.
ðNuÞD ¼ ¼ 3:663 þ 1:613 ðPeÞp
3
where U is superficial flow velocity. For heat transfer from a cylindrical heater of length Lh
For heat transfer from a single particle inside a and radius Ry immersed in the packed bed,
particle array,
!
hdp 1e @T k @2 T 1 @T
Nu ¼ ¼ 5:49 ð112Þ ¼ e þ ð120Þ
kg e @z Cpg G @r2 r @r
Wen-Ching Yang
Siemens Westinghouse Power Corporation, Pittsburgh, Pennsylvania, U.S.A.
dominant influence of cohesion forces, and its demar- are good also for gases other than air and for tempera-
cation can be expressed by the equation ture and pressure other than ambient.
Figure 3 Bed collapsing stages during bed collapsing characterization of powders. (Adapted from Kwauk, 1992.)
0 U < Umf Fixed bed Particles are quiescent; gas flows through interstices
Umf U < Umb Particulate regime Bed expands smoothly and homogeneously with small-scale
particle motion; bed surface is well defined
Umb U < Ums Bubbling regime Gas bubbles form above distributor, coalesce and grow; gas
bubbles promote solids mixing during rise to surface and
breakthrough
Ums U < Uc Slug flow regime Bubble size approaches bed cross section; bed surface rises
and falls with regular frequency with corresponding
pressure fluctuation
Uc U < Uk Transition to turbulent Pressure fluctuations decrease gradually until turbulent
fluidization fluidization regime is reached
Uk U < Utr Turbulent regime Small gas voids and particle clusters and streamers dart to
and fro; bed surface is diffused and difficult to distinguish
U > Utr Fast fluidization Particles are transported out of the bed and need to be
replaced and recycled; normally has a dense phase region
at bottom coexisting with a dilute phase region on top; no
bed surface
U
Utr Pneumatic conveying Usually a once-through operation; all particles fed are
transported out in dilute phase with concentration varying
along the column height; no bed surface
fixed bed to the bubbling regime is the minimum fluid- 3.3 Transition Between Bubbling and Turbulent
ization velocity for Group B and Group D powders Regimes
while for Group A powders, it is the minimum bub-
bling velocity. The homogeneous expansion, also When the gas velocity is continuously increased, the
called particulate fluidization, occurs only in Group bubbles grow bigger owing to coalescence, and the
A powders for gas–solids systems. The particulate bubbling bed can transfer into a slugging bed if the
fluidization is especially important for liquid fluidized bed diameter is small and the particle diameter is
beds because where most of the occurrence is observed. large, or into a turbulent bed if the bed diameter is
Liquid–solids fluidization will be discussed in Chapter large and the particle diameter is small. If the standard
26, ‘‘Liquid–Solids Fluidization.’’ deviation of pressure fluctuation is measured and
plotted against the superficial fluidization velocity,
two characteristic velocities, Uc and Uk , first suggested
3.2 Transition Between Bubbling and Slugging by Yerushalmi and Cankurt (1979), can be identified.
Regimes The velocity Uc corresponds to the bed operating con-
ditions where the bubbles or slugs reach their maxi-
A slugging regime occurs only in beds with bed height mum diameter and thus have the largest standard
(H) over bed diameter ratio (D) larger than about 2. deviation of pressure flucuation. Continuing increases
With large H=D ratios, the bed provides enough time beyond this velocity, large bubbles start to break up
for bubbles to coalesce into bigger ones. When the into smaller bubbles with smaller pressure fluctuation,
bubbles grow to approximately 2/3 of the bed dia- and eventually the standard deviation of the pressure
meter, the bed enters the slugging regime with periodic fluctuation reaches a steady state. This velocity is
passing of large bubbles and regular large fluctuation denoted as Uk , also a characterization velocity for
of bed pressure drop corresponding to the bubble transition from the bubbling regime to the turbulent
frequency. There are several correlations available to regime. Subsequent study by Chehbouni et al. (1994),
predict this transition, they are discussed in Sec. 10; employing differential and absolute pressure transdu-
‘‘Slugging Beds.’’ cers and a capacitance probe, concluded that the onset
dp rf Utr
ðReÞtr ¼ ¼ 2:28Ar0:419 ð37Þ
m
Bi and Grace (1995b) correlated the literature data
based on the differential pressure fluctuation and
arrived at the following transition equation:
d p rf U c
ðReÞc ¼ ¼ 1:24Ar0:45
m
for 2 < Ar < 1 108 ð38Þ
For gas fluidization of large particles, Catipovic et al.
(1978) further subdivided the regimes into slow bub-
bles, fast bubbles, and rapidly growing bubble regimes.
The transition equations are summarized here.
Regime Between Fast and Slow Bubbles
pffiffiffiffiffiffiffiffiffi
Umf gDB
¼ UB ¼ ð39Þ
emf 2
Figure 5 Regimes of fluidization for transporting and non-
transporting systems. Regime of Rapidly Growing Bubbles
In this regime, the bubble growth rate is of the same
magnitude as the bubble rising velocity, and for
shallow beds,
Perales et al. (1990) also suggested equations to calcu-
late Uk and Utr as follows. dðDB Þ dðDB Þ dh
¼ ¼ UB ð40Þ
dt dh dt
d p rf U k or
ðReÞk ¼ ¼ 1:95Ar0:453 ð32Þ
m
dðDB Þ
¼1 ð41Þ
dh
dp rf Utr
ðReÞtr ¼ ¼ 1:41Ar0:483 ð33Þ
m 3.4 Transition to Fast Fluidization
Since the Uk and Utr are very similar, they suggest that Continuing increases in operating velocity beyond that
the following equation alone may be used to calculate required at turbulent fluidization, a critical velocity,
both Uk and Utr for simplification: commonly called the transport velocity Utr , will be
reached where a significant particle entrainment
dp rf Uk dp rf Utr occurs. Beyond this point, continuing operation of
ðReÞk;tr ¼ ¼ ¼ 1:45Ar0:484 ð34Þ
m m the bed will not be possible without recycle of the
entrained solids. The bed is now said to be in the fast
Bi and Fan (1992) affirmed the existence of turbulent fluidization regime. The transition velocity has been
regimes in gas–solid fluidization and suggested the fol- correlated by Bi et al. (1995) as
lowing criteria for transition to turbulent fluidization.
Utr ¼ 1:53Ar0:5 for 2 < Ar < 4 100:5 ð42Þ
and assuming
4 THEORETICAL AND EMPIRICAL
PREDICTIONS OF MINIMUM 72
¼ 180; g is usually from 48 to 518 ð44Þ
FLUIDIZATION VELOCITY cos2 ðgÞ
we have
The phenomenon of fluidization can best be character-
ized by a P=L versus U plot such as the one shown in P mU
Fig. 6. Below a characteristic gas velocity known as the ¼ 459 2mf ð45Þ
L dp
minimum fluidization velocity, a packed bed of solid
particles remains fixed, though a pressure drop across At the point of minimum fluidization, the pressure
the bed can be measured. At the minimum fluidization drop is enough to support the weight of the particles
velocity, all the particles are essentially supported by and can be expressed as
the gas stream. The pressure drop through the bed is P
then equal to the bed weight divided by the cross- ¼ ðrp rf Þð1 emf Þ ð46Þ
L
sectional area of the bed, P ¼ W=A. Further
increases in gas velocity will usually not cause further Combining Eqs. (45) and (46) with the voidage at the
increases in pressure drop. In actual practice, however, minimum fluidization emf ¼ 0:476, we find
pressure drop at minimum fluidization velocity is actu-
ally less then W=A because a small percentage of the gdp2 ðrp rf Þ
Umf ¼ 0:00114 ð47Þ
bed particles is supported by the wall owing to the less m
Figure 6 Pressure drop vs. fluidization velocity plot for determination of minimum fluidization velocity.
Grace (1970) reviewed the literature available in this are remarkably similar to that obtained by others using
area in 1970 and concluded that the direct a rotating cylinder viscometer. For more discussion on
measurement techniques commonly used to determine the viscosity of fluidized beds, see that by Schugerl
the viscosity of newtonian liquids tended to alter the (1971).
behavior of the bed, e.g., the regions underneath and The dense phase viscosity of fluidized beds at ele-
above the probe tended to have high and low voidages. vated pressure was studied by King et al. (1981) at
Thus the viscosity obtained by those methods may not pressures up to 20 bar by measuring the velocity of a
be the true viscosity of a fluidized bed. He, in turn, falling sphere. They found that for particles less than
proposed an indirect method based on the behavior about 100 mm, increases in pressure caused substantial
of bubbles in fluidized beds. decreases in viscosity. For larger particles, however,
The spherical-cap bubbles observed in fluidized beds the viscosity is almost independent of pressure.
can best be characterized by the included angle as
shown in Fig. 9. The angle was found to be dependent
only on the bubble Reynolds number, as shown in Eq.
(71)
De UB rf
ðReÞe ¼ ¼ 23e0:004y 2008 < y < 2608
m
ð71Þ
where De is the equivalent bubble diameter and can be
calculated as
6VB 1=3
De ¼ ð72Þ
p
The viscosity of the fluidized bed can then be estimated
by the equation
De UB rp ð1 emf Þ
m¼ ð73Þ
ðReÞe
The estimated range of the viscosity of fluidized beds is Figure 9 Schematic of spherical-cap bubbles for determina-
from 4 to 13 poises by using this method. These values tion of fluidized bed viscosity.
G1:2
Figure 12 Davidson’s bubble model—slow bubbles. VB ¼ 1:138 ð106Þ
g0:6
At high gas flow rates where the bubble sizes are inde-
The stream function for the particles in this case can be pendent of the bed viscosity, the inviscid liquid theory
similarly expressed as in Eq. (102). can predict the bubble sizes satisfactorily. At low gas
In both these cases, i.e., UB > Umf =emf and flow rates where the viscosity effects are quite pro-
UB < Umf =emf , the gas exchange between bubble and nounced, the inviscid liquid theory underestimates
particulate phase was obtained by Davidson through the bubble sizes. In this case, the following equation
an analogy to a fixed bed: by Davidson and Schuler (1960) should be used.
q ¼ 3pR2B Umf for a 3-D bed ð104Þ G1:2
VB ¼ 1:378 ð107Þ
g0:6
The gas exchange rate q expressed in Eq. (104) is the
total volumetric flow rate of gas passing through the Equation (107) fits the data by Harrison and Leung
bubble void. It can either pass through the bubble void (1961) and the data by Bloore and Botterill (1961)
or be recirculated back to the bottom of the bubble, fairly well. There is, however, increasing evidence
depending on the relative magnitude of the bubble showing that the gas leakage from the bubble to the
velocity and the minimum fluidizing velocity. emulsion phase can be substantial, especially for large
There are other models for the bubble motion in particles. Nguyen and Leung (1972) performed experi-
fluidized beds of more complexity, primarily those by ments in a 2-D bed with a fluidizing velocity of 1.2
Jackson (1963) and Murray (1965). They are more times the minimum fluidizing velocity and found that
exact than the Davidson’s model but give essentially the bubble volume could be better approximated by
the same results. the following equation:
G
9.4 Bubble Formation in a Fluidized Bed VB ¼ 0:53 fn ¼ bubble frequency ð108Þ
fn
Bubble formation in a fluidized bed was found experi- From the computer enhanced video images of rising
mentally to be very similar to that in an inviscid liquid. bubbles in fluidized beds, Yates et al. (1994) observed
At a very low gas flow rate, the frequency and size of that the bubbles are surrounded by a region of emul-
the bubbles formed are primarily governed by a bal- sion phase in which the solids concentration is lower
ance between the surface tension of the liquid and the than that in the emulsion phase far from the bubbles.
buoyancy force of the bubble. The inertia of the liquid This region of increasing voidage was called the ‘‘shell’’
moved by the rising bubbles becomes more improtant by Yates et al. The volumetric gas in the bubble and in
than the surface tension at higher gas rates. It is in this the shell can be correlated as
Based on this equation, the volume of void following ½ðU Umf Þ=No 0:4
DB ¼ 1:43
coalescence of two bubbles is about 28% larger than g0:2
the volumes of the two constituent voids, because of þ 0:027HðU Umf Þ0:94 ð113Þ
the incorporation of gas from the shells.
The constant 0.027 is a dimensional constant with
9.5 Coalescence of Bubbles in Fluidized Beds units of (cm/s)0:94 . Equation (113) gives reasonable
agreement with data from fluidized beds using indus-
The bubbles in fluidized beds grow in size due prim- trial types of orifice distributor plates. Porous distribu-
arily to three factors: tor plates, as expressed in Eq. (112), behave as though
they contained approximately 1 hole per 10 cm2 of bed
1. The effective hydrostatic pressure descreases
area. The principal effect of adding fines to a fluidized
toward the top of the fluidized bed.
bed of group B powders is the reduction of the mean
2. Bubbles coalesce in the vertical direction with
particle size. At equal values of excess ðU Umf ), this
the trailing bubble catching up the leading
results in increased bed expansion and solid circulation
bubble, and
rates but produces no decrease in mean bubble size.
3. Bubbles coalesce in the horizontal direction
Mori and Wen (1975) assumed that all gas above
with the neighboring bubbles.
the minimum fluidizing velocity went to form a single
The effect of the hydrostatic pressure is usually small, train of bubbles rising along the center line of the bed
and the bubbles grow in size owing largely to coales- and calculated the diameter of bubble that would exist
cence. There are a number of bubble coalescence as
models in the literature. A few of the more well-
known ones are discussed here. DBM ¼ 0:652½AðU Umf Þ2=5 ð114Þ
Geldart (1972) found that the fluidization behavior The bubble diameter at bed height H can then be esti-
of Group B powders was independent both of the mated as
mean particle size and of particle size distribution. In
particular, the mean bubble size was found to depend DBM DB H
¼ exp 0:3 ð115Þ
only on the type of the distributor, the distance above DBM DBo D
the distributor plate, and the excess gas velocity above
that required at the minimum fluidization condition, For perforated plates, the initial bubble diameter DBo
U Umf . Mathematically, it can be expressed as is expressed as
DB ¼ DBo þ KH n ðU Umf Þm ð110Þ AðU Umf Þ 2=5
DBo ¼ 0:347 ð116Þ
No
According to the theory by Davidson and Harrison
(1963), the size of a bubble issuing from a single orifice where A is the area of the bed and No is the total
in a fluidized bed at the minimum fluidization condi- number of orifices. For porous plates, the following
tion can be calculated from Eq. (106). The same idea expression should be used to estimate the initial bubble
was extended by Geldart to multiorifice distributor sizes.
plates by replacing G with ðU Umf Þ=No and by DBo ¼ 0:00376ðU Umf Þ2 ð117Þ
replacing VB with ð1 fw ÞpD3Bo =6. By assuming the
wake fraction fw to be 0.25, we have The validity of the above equations has been tested
within the ranges of the following parameters:
U Umf 0:4
No 30 < D < 130 cm 0:5 < Umf < 20 cm=s
DB ¼ 1:43 þ KH n ðU Umf Þm ð111Þ
g0:2 0:006 < dp < 0:045 cm U Umf < 48 cm=s
Experimentally, it has been found that for the porous Rowe (1976) suggested the following equation to esti-
plates, the following equation applies. mate the bubble size in a fluidized bed:
DB ¼ 0:915ðU Umf Þ0:4 þ 0:027HðU Umf Þ0:94 ðU Umf Þ1=2 ðH þ ho Þ3=4
DB ¼ ð118Þ
ð112Þ g1=4
Figure 16 Type A and Type B slugging beds, and wall slugs. 10.2 Slugging Bed Expansion
where USA is the absolute slug velocity and eS is the From the material balance of solids and gas, it can be
fraction of the bed occupied by the slugs, which can be calculated that the slug length is
expressed as pffiffiffiffi
pffiffiffiffiffiffiffi ½ðU Umf ÞNT D
LS ¼ pffiffiffi ð141Þ
USA ¼ ðU Umf Þ þ 0:35 gD ð134Þ 0:35 g
jets from a gas distributor plate are generally single region (jet penetration depth) may characterize the per-
phase, i.e., gas phase, and small in diameter, usually formance of the reactor and sometimes dictate the
less than 15 mm. The study of jetting phenomena at reactor design. Because of the sizes of solids to be
the distributor plate is of interest because the intense fed continously into the beds, the feeding nozzles and
gas–solid mixing, and the heat and mass transfer thus the jets are usually larger than 15 mm.
induced by the jets, are important for the overall per- There is another category of jets in fluidized beds
formance of fluidized bed reactors. The jetting zone that tend to be substantially larger in size. For exam-
immediately above a grid is characterized by a time- ple, a large central jet is applied in a jetting fluidized
averaged, sharp, vertical variation of bed density. A bed, similar to a spouted bed, to induce solids mixing
large fluctuation of bed density occurs in this zone, and circulation, to facilitate heat and mass transfer
indicating extensive mixing and contacting of solids from one region of a fluidized bed to an adjacent
and gas. For chemical reactors where the reaction region, and to promote granulation and agglomera-
rate is fast, much of the conversion may occur in tion. This type of jetting fluidized beds were used for
this jetting region. To achieve uniform gas distribu- solids mixing and drying, coal gasification, and pow-
tion, a minimum pressure drop equivalent to at least der granulation. The study of large jets, substantially
30% of the fluidized bed pressure drop is required. larger than 15 mm in diameter, is not as common
This necessitates a small orifice diameter in practice, (Yang, 1998b). The jetting fluidized beds are dis-
generally less than 15 mm. Thus most of the existing cussed in Chapter 20, ‘‘Other Nonconventional
jet correlations were developed on the basis of data Fluidized Beds.’’
generated from small jets. An ambiguity in studying the jetting phenomenon is
When solids are fed pneumatically into operating the lack of consensus on what constitutes a jet (Rowe
fluidized beds, they generate jets as well. In this appli- et al., 1979; Yang and Keairns, 1979). The gas issuing
cation, they are gas-solids two-phase jets. The gas and from an orifice might be in the form of bubbles, a
solids entrainment into the jets and the extent of the jet pulsating jet (a periodic jet), or a permanent flamelike
!0:67 0:24
LB rf Uj U Hirsan et al. (1980)
¼ 26:6 pffiffiffiffiffiffiffi
do rp gdp Ucf
!0:83 0:54
Lmax rf Uj U
¼ 19:3 pffiffiffiffiffiffiffi
do rp gdp Ucf
" ! !#0:472
Lmax ðUcf Þatm rf Uj2 Yang (1981); Yang (1998a)
¼ 7:65 for U ¼ Ucf
do ðUcf Þp rp rf gdo
" ! !#0:835
Lmax ðUcf Þatm rf Uj2 Yang (1981); Yang (1998a)
¼C for U > Ucf
do ðUcf Þp rp rf gdo
0:35
Lmax 0:919dp Uj Basov et al. (1969)
¼ ; units in cgs system
do 0:007 þ 0:566dp do0:3
!0:3 2 !0:2 3
Lmax rd Uj2 Merry (1975)
¼ 5:2 f o 41:3 15
do rp dp gdo
!0:47
Lj rp dp 0:585 rf do Uj 0:654 Uj2 Wen et al. (1977)
¼ 814:2
do r f do m gdo
" ! !#0:187
Lj rf Uj2 Yang and Keairns (1979)
¼ 15:0
do rp rf gdo
0:5 q 1:5; for a single jet q ¼ 0:5; for multiple jets q ¼ 1:5
!
Lj rf Uj Shakhova (1968)
¼ 7:8 pffiffiffiffiffiffiffi
do rp gdp
This fundamental relationship has been clouded over depth on the design and operating parameters for the
the years by different experimental arrangements, dif- existing correlations. The wide range of dependence is
ferent definitions of jet penetration depth, and subjec- not very comforting for practical applications in multi-
tive observations of fluctuating jet penetration depth, million dollar projects.
which result in many jet penetration correlations that " #k8
are inconsistent and of limited applicability. Equation ðUmf Þatm
(162) summarized the dependence of jet penetration Lj / rkf 1 rkp2 dok3 dpk4 Ujk5 mk6 gk7 ð162Þ
ðUmf Þp
G ¼ 0:53VB fn ð168Þ flow rate. The solid particle velocity, Us , can be calcu-
lated assuming the gas/solid slip velocity to be the
and terminal velocity of a single particle of the average size.
G ¼ 0:37VB fn ð169Þ For a concentric jet with the inner jet carrying solid
particles, the inlet jet velocity can be calculated based
To describe the jet adequately, especially a large jet, the on the equation
bubble size generated by the jet needs to be studied. A ðMg Þi ðUg Þi þ ðMs Þi ðUs Þi þ ðMg Þo ðUg Þo
substantial amount of gas leaks from the bubble to the rf ðUj Þ2c ¼
emulsion phase during the buble formation stage, par- At
ticularly when the bed is less than minimally fluidized. ð171Þ
A model developed on the basis of this mechanism
predicted the experimental bubble diameter well 11.5 Gas and Solids Entrainment
when the experimental bubble frequency was used as
an input (Yang et al., 1984b). The model is described in The interchanges of gas between the jetting region and
Yang (1999) and in Chapter 20, ‘‘Other Nonconven- the outside emulsion phase can be studied by integra-
tional Fluidized Beds.’’ tion of the measured velocity profiles in the jet or by
tracer gas analysis. The velocity profiles in a jet in a
11.4 Two-Phase Jets and Concentric Jets fluidized bed can be taken to be similar and can be
approximated by either Schlitchting or Tollmien simi-
The inlet jet velocity for a single-phase (gas) jet is cal- larity. The turbulent jet equations already developed in
culated based on the cross-sectional area of the jet the homogeneous medium can then be integrated to
nozzle and the total volumetric flow rate of the jet. If give the entrainment of gas into the jet region in a
the jet also carries solids, as in the case of feeding solids fluidized bed. It is generally agreed that gas is
pneumatically into a fluidized bed, the inlet jet velocity entrained from the emulsion region into the jet at
should be calculated from the equation close to the jet nozzle. This exchange reverses direction
M g Ug þ Ms Us at larger distance from the jet nozzle (Donadono and
rf ðUj Þ2s ¼ ð170Þ Massimilla, 1978; Yang and Keairns, 1980a; Filla et
At
al., 1981; Filla et al., 1983b; Gbordzoe et al., 1988,
Since the voidage in the inlet jet is close to 1, the gas Yang, 1998b).
velocity, Ug , can be calculated based on the cross- The rate of solids entrainment was studied by
sectional area of the jet nozzle and the volumetric jet Donadono et al. (1980), Massimilla et al. (1981),
11.5.1 Model for Solid Entrainment into a 11.6 Interacting Jets in Fluidized Beds
Permanent Flamelike Jet
When there is more than one jet in a fluidized bed, jet
A mathematical model for solid entrainment into a interaction is to be expected in most cases. Thus jet
permanent flamelike jet in a fluidized bed was pro- interaction is an important phenomenon to study for
posed by Yang and Keairns (1982a). The model was operation and design. Unfortunately, not too much
supplemented by particle velocity data obtained by information is available in the literature. Wu and
following movies frame by frame in a motion analyzer. Whiting (1988) studied the interaction of two adjacent
The particle entrainment velocity into the jet was
found to increase with increases in distance from the
jet nozzle, to increase with increases in jet velocity, and
to decrease with increases in solid loading in the gas–
solid two-phase jet. High-speed movies indicated that
the entrained particles tended to bounce back to the jet
boundary more readily under high solid loading con-
ditions. This may explain why the entrainment rate
decreases with increases in solid loading in a two-
phase jet. A ready analogy is the relative difficulty in
merging into a rush-hour traffic as compared to mer-
ging into a light traffic.
The simple model by Yang and Keairns (1982a) for
solid entrainment into a permanent flamelike jet shown
in Fig. 21 resulted in the equation
Wj ¼ 2prp ð1 ez Þ
C1 tany 3 1 C1 do C 2 do
Lj þ þ C2 tany Lj þ
2
Lj
3 2 2 2
Figure 21 A model for solids entrainment into a jet in a
ð172Þ fluidized bed.
Figure 22 Schematic of jet interaction of multiple jets in fluidized beds. (Adapted from Wu and Whiting, 1988.)
particle segregation. It is not unusual to have one part ties (Nienow et al., 1978a). In a separate study by
of the fluidized bed reactor operated in mixing mode Burgess et al. (1977), the initial condition of the bed
while the other part of the same reactor is in segrega- was found to be important. The well-mixed initial con-
tion mode. It is important to point out that the segre- dition (as compared to the unmixed initial condition of
gation pattern or the particle distribution profile in the two separate layers of flotsam and jetsam) led to less
bed is set up by the dynamic equilibrium between the segregation at all gas flow rates. Unfortunately, only
two competing mechanisms of solids mixing and par- limited experimental data are available with a well-
ticle segregation. Particle segregation can usually be mixed starting mixture (Chen and Keairns, 1975,
prevented by operating a fluidized bed at a sufficiently 1978).
high fluidizing velocity. On the other hand, a bed with There are two primary objectives for investigating
particle size ratio between the largest and the smallest the particle segregation phenomenon in gas fluidized
as small as 3 can be made to segregate by operating at beds. In one respect, the fluidized beds are studied to
a small enough fluidizing velocity (Wen and Yu, determine the operating conditions required to pro-
1966b). mote bed mixing and eliminate or minimize particle
Past studies on particles in a gas fluidized bed have segregation. A mixing index can generally be defined
concentrated primarily on the mixing aspect of the in this case to measure the closeness to perfect mixing.
phenomenon, notably those by Rowe and Nienow On the contrary, the other objective is to study the
(1976) using two separate layers of flotsam and jetsam optimum conditions under which clean separation
as a starting mixture. The flotsam is the lighter or can be accomplished between different materials (or
smaller components; which tend to remain at the top components) in the bed. For this case, both the degree
layer of the bed upon fluidization, while the jetsam is and the rate of particle separation are important
those heavier or larger components, which tend to stay aspects of investigation.
close to the bottom part of the fluidized bed. These It is generally recognized that both particle mixing
words were coined originally by Rowe et al. (1972) and segregation in a gas fluidized bed require the pre-
and now have become widely accepted terminology. sence of gas bubbles. No mixing or segregation of par-
The works performed by Professor Rowe and his ticles can occur in a fixed bed, because relative motion
associates have been reviewed in Rowe and Nienow between the particles is not possible. The only way the
(1976). A quantitative analysis was proposed for the segregation can happen in fixed beds is for one of the
mixing of two segregating powders of different densi- components to be sufficiently small, usually less than
where eB is the volume fraction of bubble in the where Y is a correction for deviation from the two-
fluidized bed. phase theory, and the factor 0.38 for drift flux is
The lateral particle dispersion coefficient was also from the experimental evidence that the particles car-
studied by Shi and Fan (1984) and Subbarao et al. ried up in the drift travel on average at about 38% of
(1985). A one-dimensional diffusion model was used the bubble velocity. The value of wake and drift frac-
by Shi and Fan (1984) to characterize lateral mixing tion, bw and bd , can be obtained from Fig. 25, while
of solids. Through dimensional analysis and nonlinear the correction for deviation from the two-phase theory,
regression analysis of the literature data, they arrived Y, can be found from Fig. 26, also due to Geldart
at an equation for the lateral dispersion coefficient for (1986). Values of bw , bd , and Y for common materials
general application. are summarized in Table 4.
Figure 25 Dependence of wake and drift fraction on particle characteristics. (Adapted from Geldart, 1986.)
Case I dB =dS 10
In a gas fluidized bed of a mixture of wide size and as was observed experimentally by Knowlton (1977)
density distribution, the distinction between a flotsam and by Chiba et al. (1979).
and a jetsam becomes less clear because the individual A study by Chiba et al. (1979) on the minimum
components are distributed axially into an equilibrium fluidization velocity of binary particle mixtures indi-
distribution governed primarily by hydrodynamics. cated that the fluidization curve shown in Fig. 27 was
Uchida et al. (1983) modified the equation by rp ¼ rpi arithmetic mean ð206Þ
i¼1
Cheung et al. (1974), substituting the volume fraction
for the weight fraction in the original equation and and
introducing an additional empirical constant, m, and Xn
xi
suggested the following equations. d p ¼ 1= harmonic mean ð207Þ
d
m i¼1 pi
ðUmf ÞJ ð1fF Þ
ðUmf ÞM ¼ ðUmf ÞF ð201Þ Most of the literature correlations for calculating the
ðUmf ÞM
minimum fluidization velocity of a multicomponent
where mixture are derived based on the assumption that the
M could be correlated with the gas velocity by the significantly larger than those measured experimen-
following equation: tally. Equations (211) and (213) are only recommended
1 for binary mixtures with volumetric jetsam concentra-
M¼ ð211Þ tion less than about 50%, and Eq. (213) is also only
ð1 þ eZ Þ
good for systems with a particle size ratio less than 3.
where Equation (213) cannot be applied to a bed with a high
aspect ratio where slugging occurs.
U UTo
Z¼
eU=UTo ð212Þ The equations of Nienow et al. were modified later
U ðUmf ÞF by Rice and Brainnovich (1986) and Peeler and Huang
The takeover velocity UTo , where dM=dðU ðUmf ÞF ) is (1989). A more recent study on segregation by size
a maximum and M ¼ 0:5, could be estimated from the difference was conducted by Wu and Baeyens (1998).
following empirical equations: They found that the excess gas flow rate required to
prevent segregation in a fluidized bed with a wide size
UTo ðUmf ÞJ 1:2 distribution of powder can be calculated from the mix-
¼
ðUmf ÞF ðUmf ÞF ing index expression shown in the equation
þ 0:9ðrR 1Þ1:1 dER
0:7
2:2ðxJ Þ0:5 ðH Þ1:4 0:75
G
ð213Þ M ¼ 1 0:0067dR 1:33 B ð217Þ
A
The rR in Eq. (213) is the density ratio; the dER , the
For dR ffi 2, good mixing ðM 0:9Þ can be achieved
size ratio; and the H , the reduced bed height. They are
when the visible bubble flow ðGB =AÞ is larger than
defined as
about 0.094 m/s. Since temperature has limited effect
r on the visible bubble flow rate, it is expected that tem-
rR ¼ J ð214Þ
rF perature has very little effect on the particle separation
due to difference in size alone.
f J dJ Simple quantitative equation can only be formu-
dER ¼ ð215Þ
fF dF lated for binary systems of near spherical particles.
For binary systems of granular particles, Rowe et al.
H (1972) suggested the following dependence for segrega-
H ¼ 1 exp ð216Þ
D tion tendency.
Equation (213) gave reasonable prediction of UTo for Segregation tendency
binary systems of different particle densities 2:5 0:2
ðrJ =rF 6¼ 1Þ. For binary systems with particles differing r dB
/ ½U ðUmf ÞF J ð218Þ
only in size, i.e., rJ =rF ¼ 1, Eq. (213) gave UTo values rF dS
V1 ð1 ðCF ÞW Þ VJ1 where
ln ¼
V1 ð1 ðCF ÞW Þ VJ1
o
ðU Umf2 Þ
A
fw
ð1 "w Þ
ð219Þ m¼ ð221Þ
ðU Umf2 Þ
A
fw
ð1 "w Þ V1
t
V1
There is certainly some experimental evidence indicat-
If we assume further that ðCF Þw ¼ 1:0, i.e., the bubble ing that the assumption of ðCF ÞW ¼ 1:0 is reasonable
wake contains pure flotsam (acrylic particles), Eq. at least for highly segregating systems (Yang and
(219) can be written as Keairns, 1982b).
Equation (220) was used to fit the experimental data
V
ln J1o ¼ mt ð220Þ with excellent results. The success of the correlation
VJ1 does suggest possible physical significance of the para-
meter, m, defined in Eq. (221). A constant m means a pletely segregated conditions. The data were analyzed
constant volumetric exchange rate between the two based on a mixed bed of jetsam and flotsam residing on
fluid beds during the transient period. A key assump- top of a pure jetsam bed. The interchange velocities of
tion in Eq. (220) is that ðCF ÞW ¼ 1:0. This assumption jetsam between the two beds were found to depend
is reasonable for a highly segregating system like the strongly on the excess fluidization velocity.
acrylic–dolomite system. The rate of particle separa-
tion depends critically on ðCF ÞW as shown in Eq.
Continuous Operation and Industrial Applications
(219). Unfortunately, experimental values of ðCF ÞW
for different systems are generally not available. A good example of continuous particle separation is
The mixing and segregation kinetics of mixtures of the industrial application in the ash-agglomerating flui-
iron and glass particles, a strongly segregating bed, dized bed gasifier. The ash remained from gasification
were also studied by Beeckmans and Stahl (1987) of coal is agglomerated into bigger and denser agglom-
employing both initially fully mixed and initially com- erates and is continuously removed from the fluidized
Figure 4 Elutriation rate constant Ki1 versus the particle size for fines and superfines for (a) limestone and (b) a-alumina. (Data
from Smolders and Baeyens, 1997.)
Taking into account the role of bubbles for the trans- Figure 8 Effect of the bed height on the entrainment flux (a)
port of particles into the freeboard, an influence of bed above the TDH and (b) just above the bed surface.
height and also of internals in the fluidized bed could (Dt ¼ 0:61 m, silica sand, dp ¼ 61 mm, data measured by
be expected. Baron et al. (1990) studied the influence of Baron et al., 1990.)
the bed height, using a fluidized bed column of 0.61 m
diameter and silica sand with a surface mean diameter
dp of 65 mm. Their results for the entrainment flux E1 Tweedle et al. (1970) showed by inserting screen
above the TDH and the entrainment flux E0 just above packings into the fluidized bed that the entrainment
the bed surface are shown in Figs. 8a and 8b, respec- was reduced due to the reduced bubble size.
tively. The results show a slight increase of the entrain-
ment fluxes with increasing bed height, the influence
5.5 Influence of Internals in the Freeboard
being more pronounced for the entrainment flux E0
at the bed surface and higher gas velocities. The reason
The picture of the influence of internals in the free-
for the increase of the entrainment flux is the increase
board is not very clear. On the one hand, they may
of bubble sizes with height, which obviously dominates
act as a kind of gas–solid separators, thus helping to
the decrease in bubble frequency and therefore in the
reduce the TDH (Martini et al., 1976) and the entrain-
probability of bubbles to coalesce near the bed surface.
ment flux (Baron et al., 1988). On the other hand, they
This indicates, that at least for the system studied by
may increase the entrainment due to the increased gas
Baron et al., the ‘‘bubble nose’’ mechanism is the dom-
velocity in the freeboard (Tweedle et al., 1970), or they
inating mechanism for particle ejection into the free-
may not affect the entrainment at all (George and
board. Choi et al. (1989), however, found no influence
Grace, 1981). Specially designed baffles and inserts in
of bed height on entrainment during their study of
the freeboard may help to reduce the entrainment flow
entrainment from fluidized bed combustors. In this
significantly (Harrison et al., 1974).
study they used two fluidized bed combustors with
cross-sectional areas of 0:3 0:3 m2 and 1:01
0:83 m2 , respectively. The smaller one was equipped 5.6 Influence of Temperature and Pressure
with an extended freeboard of 0:45 0:45 m2 . In
fact, the bed height effect in Fig. 8 is small compared, Studies of the influence of pressure have been carried
e.g., to the effect of the gas velocity. out by Chan and Knowlton (1984a, 1984b). They
dp;bed
U, m/s D, m mm Reference Comments
1 Ki1
g
dpi
2
0.3–1.0 0.07–1.0 0.1–1.6 Yagi and Aochi Cited by different authors
¼ 0:0015
Re0:6 1:2
t þ 0:01
Ret (1955) as cited in different ways,
mðU uti Þ 2
by Wen and original publication
Chen (1982) inaccessible
8 !1:88
2 >
> U2 U2 0.3–0.7 0:05 0:53 0.04–0.2 Zenz and Weil Correlation aiming at FCC
>
>
7
>
< 1:26 10 for < 3:10 (1958) fluidized beds
Ki1 g dpi r2p g dpi r2p
¼ !1:18
rg
U >> U2 U2
>
>
>
: 1:31
10
g d r2
4
for > 3:10
pi p g dpi r2p
!0:5
3 0.6–1.0 0.102 0.7 Wen and
Ki1 ðU Uti Þ2
¼ 1:52
105
Re0:725
i Hashinger (1960)
rg ðU Uti Þ g dpi
!0:5 !0:15
4 0.9–2.8 0.031–0.067 0.7–1.9 Tanaka et al.
Ki1 ðU Uti Þ2 rs rg
¼ 4:6
102
Re0:3
t
(1972)
rg ðU Uti Þ g dpi rg
" 0:25 #
5
Ki1
u 0:5 U 0.6–2.4 0:91 0:91 0.06–1.0 Merrick and Correlation derived for
ti mf
¼ A þ 130
exp 10:4 Highley (1974) bubbling fluidized bed
rg
U U U Umf
combustors
with A ¼ 103 . . . 104
Ki1
u
6 0.6–3.0 0.076 0.06–0.35 Geldart et al.
¼ 23:7
exp 5:4 ti
rg
U U 0.3 1.5 (1979)
!
7 K 2 1:65 0.1–0.3 0:61 0:61 0–0.125 Lin et al. (1980)
i1 4 U
¼ 9:43
10
rg
U g dp
!
U2
for 58 1000; 0:1m=s U 0:3 m=s;
g dpi
0 < dp 74 mm
10 Ki1 U u Sciazko et al
¼ 1:6 1 ti
rg
U uti U (1991)
3:4
12 kg 0.1–0.6 0.071 0.03–0.2 Nakagawa et al.
Ki1 ¼ 0:35rs Uð1 eÞH
m2 s 0:08 0:08 (1994)
0:15 0:15
with
ð1 eÞH ¼ 7:41
103 R1:87 A0:55
t
0:64
HFB
X U Uti
and R ¼ xi for uti < U
Uti
8
13 < 23:7
r
U 2:5 exp 5:4 uti for Re < 3000 0.2–0.8 0.076, 0.152 0.017–0.077 Tasirin and
kg g
U
Ki1 ¼
u Geldart (1998c)
2
m s : 14:5
r
U 2:5 exp 5:4 ti for Re > 3000
g
U
Dt
U
with Re ¼
n
14 Ki1 dp 13:1 0.3–7.0 0.06–1.0 0.05–1.0 Choi et al. (1999) Correlation based on a
¼ Ar exp 6:92 2:11Fg
0:5 0:303
0:902
m Fd wide range of different
units, materials, and
with Fg ¼ g dp ðrs rg Þ ðgravity force per projection area) operating conditions,
e.g., temperature and
rg U 2 pressure
Fd ¼ Cd (drag force per projection area)
2
Equation Reference
1:2
1 TDH ¼ 0:85U ð7:33 1:2 log UÞ Chan and Knowlton (1984b)
U2
2 TDH ¼ 1000 Fournol et al. (1973)
g
3 TDH ¼ 18:2db George and Grace (1978)
4 TDH ¼ 13:8db Fung and Hamdullahpur (1993)
1500Hb Rep
5 TDH ¼ Sciazko et al. (1991)
Ar
All parameters to be inserted in SI units.
Table 3 Correlations for the Calculation of the Entrainment Flux E0 at the Bed Surface
Equation Reference
E0 ¼ 0:1
rs ð1 "mf ÞðU Umf Þ bubble wake model
A
db ðU Umf Þ2:5 r3:5
g
g
0:5
2 E0 ¼ 3:07
109 2:5
Wen and Chen (1982)
m
3:5
298
3 E0 ¼ 9:6
A
ðU Umf Þ A
db
2:5
ðT in kelvinsÞ Choi et al. (1989)
T
J. G. Yates
University College London, London, United Kingdom
Range Correlation
24 h i
i.e., CD ¼ 1 þ 0:1315 Reð0:820:05wÞ
Re
(C) 20 < Re 260 CD Re
log10 1 ¼ 0:7133 þ 0:6305w
24
24
i.e., CD ¼ 1 þ 0:1935 Re0:6305
Re
(E) 1:5 103 Re 1:2 104 log10 CD ¼ 2:4571 þ 2:5558w 0:9295w2 þ 0:1049w2
(F) 1:2 104 < Re < 4:4 104 log10 CD ¼ 1:9181 þ 0:6370w 0:0636w2
[Group A in the Geldart (1973) classification], pressure sharply with pressure up to about 20 bar and more
is predicted to have little if any effect on umf . The gradually thereafter. Similar conclusions had been
reason for this is that the flow of gas around these reached by King and Harrison (1982) who showed
small particles is laminar and so the fluid–particle that on the basis of Eq. (3) umf is independent of pres-
interaction force is dominated by the gas viscosity sure for laminar flow (Remf < 0:5) while for turbulent
(which is essentially independent of pressure in the flow (Remf > 500) umf is inversely proportional to the
range considered); the Galileo number, Ga, is a linear square root of gas density and hence pressure; pressure
function of pressure. As particle size increases, how- dependence would be expected to be weaker in the
ever, inertial forces become more important and at dp intermediate regime.
> 500 mm (Geldart Group B) they begin to dominate Olowson and Almstedt (1991) measured the umf of a
over the viscous forces; this causes umf to decrease range of particles in Geldart Groups B and D at pres-
sures from 0.1 to 0.6 MPa and found a general
decrease with increasing pressure. They found the
effect of pressure to be well described by the Ergun
equation and a number of simplified correlations
derived from it, although the accuracy of these correla-
tions differed significantly.
The effect of temperature on umf may also be seen
by examination of Eq. (15). Since the density of a gas is
inversely proportional to its absolute temperature, gas
density will decrease with increasing temperature; the
viscosity of a gas on the other hand increases with
increasing temperature mg being proportional to T n
where n is usually between 0.5 and 1.0. The combined
effect of changes in density and viscosity results in the
Galileo number decreasing steadily with increasing
temperature, but it is not immediately obvious from
Figure 1 Effect of pressure on umf , based on Eq. (15). (From Eq. (15) how this will affect umf . Figure 2 shows Eq.
Rowe, 1984.) (15) plotted with a square root dependence of gas
Figure 3 Effect of temperature and pressure on terminal fall velocity of Group A particles in air.
by Davidson and Harrison (1963). The theory is based used x-rays to observe both slugs and bubbles and
on the assumption that as a bubble rises in a fluid bed found that in all cases of Group A materials, breakup
the shear force exerted by the particles moving down- occurred by fingers of particles falling in from the roof
wards relative to the bubble sets up a circulation of gas and that the effect was more pronounced the higher the
within the bubble and that the velocity of this circula- pressure.
tion, uc , approximates to the bubble rise velocity, ub . In a separate study King et al. (1981) determined the
When through coalescence the bubble diameter viscosity of the emulsion phase of fluidized beds of
increases to the point where ub > ut , the terminal fall glass ballotini of a range of sizes by measuring the
velocity of the bed particles, the solids in the wake will rate of fall under gravity through the bed of a small
be drawn up into the bubble, causing it to break up metal sphere. The beds were fluidized by both carbon
into smaller bubbles with lower rise velocities. The dioxide and nitrogen at pressures of up to 20 bar. They
bubble would thus not be expected to exceed a certain found that an increase in gas pressure led to a substan-
size determined by the value of ut , and beds of small, tial decrease in the viscosity of the finest powder but
light particles should show ‘‘smoother’’ fluidization that the viscosities of beds of powders larger than
characteristics than those of coarser materials. Now about 100 mm were almost independent of pressure.
ut decreases as gas pressure increases (see above), and
so if the foregoing theory is correct, Group A powders 4.1.2 Group B Powders
should show smoother behavior at elevated pressures;
In the case of these materials King and Harrison (1980)
this effect has been widely observed (Hoffmann and
reported no effect of pressure on bubble size up to 25
Yates, 1986).
bar, but Hoffmann and Yates (1986) found that mean
An alternative theory developed by Upson and Pyle
bubble diameters increased slightly up to 16 bar and
(1973) and Clift et al. (1974) suggests that bubble
decreased thereafter up to 60 bar; the extent of these
breakup is caused by an instability in the bubble roof
effects was dependent on the vertical position in the
allowing particles to rain down through the void and
bed at which the observations were made.
so divide it into two. On the basis of hydrodynamic
Hoffmann and Yates also measured changes in the
theory due to Taylor (1950), Clift et al. (1974) showed
value of the bubble velocity coefficient, K, with increas-
that the only factor determining the stability of the
ing pressure:
bubble roof is the apparent kinematic viscosity of the
emulsion phase and that bubbles will become less 1=2
gdb
stable as this viscosity decreases. A decrease in dense- ub ¼ K ð23Þ
2
phase viscosity would result from an increase in the
voidage of that phase, and as we have already seen The work gave values of K for a freely bubbling bed of
this is an increasing feature of beds of Group A pow- alumina powder at a point 40 cm above the distributor,
ders as operating pressure is increased. King and decreasing from 0.9 at 1 bar to 0.7 at 20 bar; this was
Harrison (1980) reported results of a study of particles followed by a gradual increase to 1.2 at 60 bar. The
of Groups A and B at pressures of up to 25 bar. They observed decrease up to 20 bar is consistent with the
Particle Max.
Bed cross properties pressure, Observations with increasing
Reference Section, cm material rp ; gcm3 dp ; mm Gas kPa pressure (P, kPa)
Godard and Richardson diam: ¼ 10:0 (i) diakon 1.18 125 air 1414 Umb -Umf increases.
(1968) (ii) phenolic 0.24 186
resin
Botterill and Desai diam: ¼ 11:4 (i) copper — 150, 625 air CO2 1000 Quality of fluidization increases for
(1972) (ii) shot — 160, 800, large particles; heat transfer
(iii) sand — 2740 coefficients double from 1 to 10
coal 1430 bar for larger particles but not
smaller ones.
Knowlton diam: ¼ 29:2 (i) coals 1.12 230–780 N2 6900 (i) Fluidization becomes smoother
(1977) cokes 1.57 at P > 1035.
(ii) and chars (ii) Bed entrainment increases
siderite 3.91 290 significantly at P > 1035.
(FeCO3 ) (iii) No change in bed expansion or
density.
Saxena and Vogel diam: ¼ 15:2 (i) dolomite 2.46 765 air 834 Umf decreases and follows Ergun
(1977) (ii) sulphated 3.19 717 correlation.
dolomite
Canada and MacLaughlin 30.5 square glass 2.48 650 air 9000 Slugging less pronounced at high
(1978) beads 2.92 2600 R-12 pressure; heat transfer coefficients
increase.
Crowther and Whitehead diam: ¼ 2:7 (i) synclyst 0.9 63 Ar 6900 Umb -Umf increases; fully particulate
(1978) (ii) coals 1.3 19–63 CF4 fluidization occurs with finest
particles.
Denloye and Botterill diam: ¼ 11:4 (i) copper 8750 160, 340, air, 1000 Bed-to-surface heat transfer
(1978) shot 620 argon coefficient increases.
(ii) sand 2600 160, 570, Co2 and
1020, 2370 freon
(iii) soda 2450 415
glass
Guedes de Carvalho et al. diam: ¼ 10 (i) ballotini 2.9 64 N2 2800 No significant effects with 74 mm
(1978) 211 sand and 21 mm ballotini; bubbles
(ii) sand 2.69 74 CO2 become smaller and break up
from rear with 64 mm ballotini.
Particle Max.
Bed cross properties pressure, Observations with increasing
Reference Section, cm material rp ; gcm3 dp ; mm Gas kPa pressure (P, kPa)
Subzwari et al. 2-D bed FCC 0.95 60 air 700 ei increases; number and size of
(1978) 46 15 Powder bubbles decrease: bubbles split by
division from roof: maximum
stable bubble size decreases.
Varadi and Grace 2-D bed sand 2.65 250–295 air 2200 Umb -Umf increases; no increase in
(1978) 31 1:6 bubble splitting; bubbles split by
division from roof.
Borodulya et al. diam. 10 cm sand 2.48 126–1220 air 8000 Bed-to-surface heat transfer
(1980) 2.60 coefficient increases.
King and Harrison diam. 10 cm ballotini 2.9 64 air 2500 For 64 mm particles, bubbles become
(1980) 101 N2 smaller and less stable; no effect
475 CO2 of pressure on larger particles.
Knowlton and Hirsan semicircle (i) sand 2.63 438 Nz 5171 Jet penetration length increases;
(1980) diam. (ii) char 1.16 419 existing correlations underestimate
30.5 cm (iii) siderite 3.99 421 jet lengths.
(FeCO3 )
Rowe and MacGillivray diam.= silicon 3.19 58 air 400 Bubbles become smaller; bubble
(1980) 20 cm carbide velocity increases; visible bubble
flow diminishes.
Xavier et al. diam. = (i) ballotini 2.82 61 CO2 2500 Bed-to-surface heat transfer for fine
10 cm (ii) ballotini 2.97 475 N2 powders only slightly affected; for
(iii) ballotini 2.91 615 N2 larger particles, heat transfer
(iv) polymer 0.55 688 N2 coefficient increases with pressure.
Kawabata et al. 2-D bed sand 2.63 300 air 800 Umf decreases; bubble sizes
(1981) 30 1 2.58 430 unchanged, but they become
2.59 600 flatter; bubble velocity decreases.
Guedes de Carvallo et al. diam.=5 ballotini 2.9 64 N2 500 Mass transfer between slug and the
(1982) 101 dense phase does not decrease
with pressure as much as expected
if gas diffusivity alone is the
dominant process.
Particle Max.
Bed cross properties pressure, Observations with increasing
Reference Section, cm material rp ; gcm3 dp ; mm Gas kPa pressure (P, kPa)
Chitester et al. diam.=10.2 (i) coals 1.25 88–361 N2 6485 ei increases; for coarse particles, bed
(1984) (ii) chars 1.12 157–376 height increases at a given gas
(iii) ballotini 2.47 96–374 flow rate; for fine particles it does
not always do so. No evidence for
10:2 1:9 (i) coal 1.25 195 maximum stable bubble size up to
(ii) char 1.12 203 810 kPa. Bubble splitting occurs
from roof. Bed appears
homogeneous at highest pressures.
Weimer and Quaderer 13 carbon 0.850 66 CO þ H2 8500 For 66 mm powder, ei increased
(1985) carbon 0.850 108 from 0.53 to 0.74, ui increases
carbon 0.850 171 sevenfold, dbmax decreases sixfold;
pressure effects were similar but
of lower magnitude.
Hoffmann and Yates 17:8 12:7 alumina 1.417 450 N2 8000 dmax decreases but bubble velocity
(1986) alumina 1.488 695 increases; bubble stability
silicon 3.186 184 decreases; bubble flow in center of
carbide bed increases.
Chan et al. 38.1 sand 2.565 100 N2 3200 Bubble size decreases, bubble
(1987) sand 2.565 200 frequency increases.
sand 2.565 400
coke 1.507 400
breeze
char 1.251 400
Jacob and Weimer 9.7 carbon 0.85 44 CO þ H2 12,420 Particulate bed expansion is
(1987) 122 adequately described by Foscolo–
Gibilaro theory.
Cai et al. 28.4 silica gel 0.834 476 air 800 Higher pressure causes smaller
(1989) silica gel 0.706 280 bubbles and lower transition
silica gel 0.711 165 velocity from bubbling to
silica gel 0.844 1057 turbulent regimes.
FCC 1.172 65
FCC 1.667 53
resin 1.330 566
sand 2.580 98
tional to gas velocity to the power 4.1 but increased system pressure constant, so that the dominant effect
sharply to the 8.4 power of velocity at a pressure of 31 was an increase in gas viscosity. This would also have
bar. Chan and Knowlton also observed that particles the effect of decreasing terminal fall velocity and so
were entrained at fluidizing gas velocities lower than increasing entrainment at a given fluidizing velocity;
their terminal fall velocity and attributed this to this is just what was observed, as may be seen from
momentum transfer from smaller particles to bigger Fig. 7.
ones, causing them to be carried over at velocities
lower than the corresponding ut . A similar observation
had been made previously at ambient pressure by 7 HEAT TRANSFER
Geldart et al. (1979). Pemberton and Davidson
(1983) also studied the effect of increased pressure on 7.1 Bubbling Beds
solids entrainment from bubbling beds and found a
similar trend in the entrainment rate coefficient to Their excellent heat transfer properties constitute one
that reported by Chan and Knowlton (1984). They of the more attractive features of fluidized beds from a
attributed the increase in entrainment to the fact that design point of view. Transfer between gas and parti-
at high pressures gas bubbles within a bed are generally
smaller than for the same volumetric flowrate at ambi-
ent pressure; and that the entrainment flux produced
by particle ejection from the roof of a bursting bubble
is given by
S. B. Reddy Karri
Particulate Solid Research, Inc., Chicago, Illinois, U.S.A.
Joachim Werther
Technical University Hamburg-Harburg, Hamburg, Germany
1 INTRODUCTION
Main Advantages
Can minimize weeping; good turndown ratio; low pres-
sure drop; can support internals; can undergo thermal
expansion without damage; ports are easily shrouded;
Possible Disadvantages well suited to multilevel fluid injection; solids can flow
Expensive; difficult to avoid stagnant regions; more from above the grid to below.
subject to immediate bubble merger; difficult to
clean; difficult to modify; not advisable for sticky
solids; requires peripheral seal; ports not easily
shrouded.
Details of some nozzles that are currently used in
circulating fluidized beds (CFB) combustors are shown
in Fig. 1 (VGB, 1994). There are significant differences
between bubble caps (No. 7 in Fig. 1) and nozzles (No.
1 in Fig. 1) with respect to the prevention of solids
back flow: in the case of nozzles, the high velocity of
the gas jet prevents the solids from flowing back into
the wind box. On the other hand, in the case of the
bubble cap design, the gas flowing out of the bubble
Possible Disadvantages
cap into the bed has a rather low velocity. In this case,
the backflow of solids is avoided by letting the gas flow Defluidized solids beneath the grid; can be a less
downward from the holes in the inner tube to the lower forgiving mechanical design.
Figure 1 Distributors and nozzles used in large circulating fluidized bed combustors. (After VGB, 1994.)
Possible Disadvantages
3 GRID DESIGN CRITERIA
Difficult to construct; requires careful design to ensure
good gas distribution; requires high pressure drop for 3.1 Jet Penetration
good gas distribution.
Gas flowing from the grid holes can take the form of
2.5 Pierced Sheet Grids (Laterally Directed Flow) either a series of bubbles or a permanent jet, depending
on system parameters and operating conditions.
Produced by punching holes in a relatively thin plate. However, a permanent jet prevails for most industrial
Holes are of a semielliptical shape with slanting, conditions. Jet penetration is one of the most impor-
strongly conical openings in the direction of entry. It tant design parameters since it helps in
is primarily used in fluid bed drying applications.
Holes can be oriented in such a way to promote certain 1. Determining how far to keep the bed internals,
mixing patterns or drive the solids toward discharge such as feed nozzles, heat exchanger tubes, etc.,
nozzle. away from the grid to minimize erosion of
internals.
2. Deciding on grid design parameters such as
Main Advantages hole size and the gas jet velocity required to
Promotes solid mixing; prevents stagnant solids achieve a certain jetting region.
buildup. Facilitates discharge of most of the solids. 3. Minimizing or maximizing particle attrition at
The holes are angled so that the grids can be non- grids.
weeping for coarse solids.
Knowlton and Hirsan (1980) reported that the jet
penetration for upwardly directed jets fluctuated
greatly. Karri (1990) noted that jet penetration can
vary as much as 30% for upwardly directed jets.
However, the jet emanating from a downwardly direc-
ted grid hole is stable, and its penetration length does
not significantly fluctuate with time. Figure 2 indicates
jet penetration configurations for jets oriented
upwardly, horizontally, and downwardly. According
to Karri, the jet penetrations for various orientations in the jetting zone near the grid, otherwise the internals
can be approximately related by: could be severely eroded.
There are numerous jet penetration correlations (Zenz, For a grid, achieving equal distribution of gas flow
1969; Shakhova, 1968; Merry, 1971; Yang and through many parallel paths requires equal resistances
Keairns, 1979; Knowlton and Hirsan, 1980; Yates et and sufficient resistance to equal or exceed the maxi-
al., 1986; Blake et al., 1990; Roach, 1993) in the litera- mum value of any unsteady state pressure fluctuation.
ture. Massimilla (1985) and Karri (1990) have shown It has been determined experimentally that the ‘‘head’’
that the jet penetrations predicted by these correlations of solids in some fluidized beds above an upwardly
can vary by a factor of 100 or more. Among them, directed grid port can vary momentarily by as much
Merry’s correlation for horizontal jets was shown as 30%. This is due to large fluctuations in the jet
(Karri, 1990; Chen and Weinstein, 1993; Roach, penetration for an upwardly directed jet, as discussed
1993) to give reliable predictions, although this corre- in the previous section. The equivalent variation down-
lation was derived for horizontal jets issuing into an stream of a downwardly directed port is less than 10%.
incipiently fluidized bed, which is not exactly the same Thus as a rule of thumb, the criteria for good gas
situation as for a grid jet. Merry’s correlation to calcu- distribution based on the direction of gas entry are
late the penetration of horizontal jets is (Karri, 1990):
!0:4 !0:2
Lhor rg;h Uh2 rg;b dp 0:2 1. For upwardly and laterally directed flow:
¼ 5:25
dh rp ð1 emf Þgdp rp dh Pgrid 0:3Pbed ð3Þ
ð2Þ 2. For downwardly directed flow
The jet penetration lengths for upwardly and down- Pgrid 0:1Pbed ð4Þ
wardly directed jets can be calculated from Eq. (1).
These equations take into account the effects of pres-
and
sure and temperature on jet penetration. Knowlton
and Hirsan (1980) and Yates et al. (1986) found that 3. Under no circumstances should the pressure
the jet penetration increases significantly with system drop across a large-scale commercial grid be
pressure. In addition, Findlay and Knowlton (1985) less than 2500 Pa, i.e.,
found that the jet penetration decreases with increasing
system temperature. Bed internals should not be placed Pgrid 2; 500 Pa ð5Þ
Figure 5 The relationship between hole density and grid hole pitch for both triangular and square pitch.
5 EROSION
2. Nozzles should not be located any closer than designing a grid with the proper pressure drop criteria,
half the jet penetration height from the bed as presented in Sec. 3.2. Poorly designed bubble caps
wall. tend to have erosion problems due to the secondary
circulation of solids. Therefore bubble caps should be
The basic equation for erosion rate is of the form
designed to minimize secondary circulation of solids.
(Karri, 1990)
Erosion has often been experienced at the nozzles
Ke r2g;h Uh3 dh2 dp3 r2p used in CFB combustors (Fig. 1). A dominant mechan-
Erosion / ð15Þ ism leading to erosion is the pressure-induced gas flow
’
reversal that will be discussed below in Sec. 6. Solids
which have entered into the nozzle during a period of
5.2 Erosion at Distributor Nozzles flow reversal are entrained out once the gas flows at
high velocity in the outward direction again. The
Erosion in the nozzle or orifices is often associated with entrained solids in the high-velocity flow in the nozzle
weepage of solids. This can be avoided by carefully hole may cause severe erosion of the wall of the hole. A
6 WEEPAGE OF SOLIDS
preferable to introduce the suspension at the lowest ideal shaft work to compress each kilogram of gas is
point of the plenum (Fig. 14a,d,e) to minimize the given by
accumulation of solids or liquids in the regions in- ð P2
accessible to reentrainment For two-phase systems, it dP
Ws;ideal ¼ ð16Þ
is preferable to have some sort of deflection device P 1 rg
(Fig. 14d,e,f) between the outlet of the supply pipe
If an ideal gas behavior is assumed, then Eq. (16) trans-
and the grid to prevent the solids from preferentially
forms into
passing through the middle of the grid due to their high
" #
momentum. This preferential bypassing of solids ðg1Þ=g
g P2
causes maldistribution of gas. In addition, the config- Ws;ideal ¼ PQ 1 ð17Þ
g 1 1 1 P1
uration of Figs. 14e and 14f are preferable over the
configurations of Figs. 14a,d. or
" ðg1Þ=g #
g P1
9 POWER CONSUMPTION Ws;ideal ¼ P2 Q 2 1 ð18Þ
g1 P2
Since the grid contributed a considerable fraction of
Due to heat of compression, the raise in temperature
total pressure drop across a given fluid bed system, it
can be calculated from
is always important to estimate the power consump-
tion of the blower that drives the gas through this ðg1Þ=g
P
system. Suppose a stream of gas is to be compressed T2 ¼ T1 2 ð19Þ
P1
from an initial pressure of P1 to a higher pressure of P2
to pump it through the entire fluid-bed system. Using where g ¼ ratio of specific heats of gas ffi 1:67, 1.4, and
thermodynamics for adiabatic reversible compression 1.33 for monatomic, diatomic, and triatomic gases,
with negligible kinetic and potential energy effects, the respectively.
Ring no. Radius of each ring Length of each ring Number of holes on each ring
(i) (ri ), m (Li ) 2ri , m % of total length (Ni )
Thus adding a shroud to the grid reduces the attri- Determine the gas velocity through the grid hole
tion rate to 67% of the rate without a shroud. (trial and error). Assume Cd ¼ 0:8:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2Pgrid 2 13008
10.2 Polyethylene Reactor Grid Design U h ¼ Cd ¼ 0:8 ¼ 29:5 m=s
rg;h 19:2
Example 4. Design a flat perforated-plate grid Determine the volumetric flow rate of gas
for the polyethylene reactor schematically shown in
Fig. 15 and calculate the gas jet penetration depth. pD2 pð4:6Þ2
Q ¼ Usup ¼ 0:5 ¼ 8:3 m3 =s
Use a triangular pitch. System parameters are 4 4
Determine the number of grid holes required:
Usup ¼ 0:5 m=s rg;h ¼ 19:2 kg=m3
Q 1 8:3 1
rg;b ¼ 17 kg=m3 rp ¼ 641 kg=m3 N¼ ¼ ¼ 3582
Uh pdh2 =4 29:5 ðp=4Þð0:01Þ2
rB ¼ 272 kg=m3 Pgrid ¼ 0:4Pbed Hole density:
dh ¼ 0:01 m dp ¼ 508 mm emf ¼ 0:45 3582
Nd ¼ ¼ 215 holes/m2
t ¼ 0:019 m ðp=4Þð4:6Þ2
In the past sixty years, worldwide research has led to uniform and active gas–solid flow to enhance heat and
an essential understanding of the most important mass transfer so as to improve the overall performance
properties of fluidized beds. In particular, it has been of fluidized bed reactors, especially to make the scale-
found that the flow structures of G–S, L–S, and G–L– up easier. Therefore the study and improvement of
S fluidization mainly vary with increasing fluid velo- internals are very important for raising the productiv-
city. For example, several flow patterns or regimes ity of fluidized bed reactors and extending the useful-
have been identified in gas–solid fluidized beds ness of fluidization processes.
(Grace, 1986), such as particulate fluidization, bub- In this chapter the mechanism of the effect of inter-
bling/slugging fluidization, turbulent fluidization, fast nals to improve the performance of fluidized beds is
fluidization, and pneumatic conveying regimes (Fig. 1). discussed, and highly efficient internal tubes and baffles
The complexity of gas–solid flow structure of the used in gas–solid fluidized beds are presented. First,
above regimes is as follows: the inherent structures of two-phase flow without
internals are introduced in Sec. 1 so that the impor-
1. There are quite different flow behaviors on an
tance and necessity of internals for gas–solid fluidized
equipment scale, such as the intensity of gas–
beds can be explored. Second, various internal tubes
solid overall movement and backmixing, gas–
and baffles are collected from the literature; they are
solid dispersion, and the residence time distri-
classified and discussed in Sec. 2. Recent progress and
bution (RTD) of both gas and solid in beds.
some simple criteria for designing in this area are also
2. The bed parameters on a micro scale vary with
covered. Section 3 focuses on explaining the effects of
the radial/axial bed locations and time.
internals on two-phase flow.
3. There are many substructures in beds, such as
bubbles, clouds, wakes, channels, and clusters;
their size distribution in time and space, as well
1 INHERENT FLOW STRUCTURES OF GAS–
as the frequency of their presence and disap-
SOLID FLOW
pearance, are normally quite complicated.
4. The effect of particle size and size distribution
Bubbles, clusters and nonuniform flow structures are
on flow regime is complex.
inherent characteristics of the gas–solid fluidization
The influence of complex flow structures on reactor systems, which greatly influence the performance of
performance is complicated, even in scaled up reactors. fluidized beds in many applications. Thus an under-
Internals are usually introduced to modify the above standing of the inherent flow structures is essential
gas–solid flow structures, in an effort to form a more for introducing internal tubes and baffles with clear
purpose, and judging the improvement effect of inter- which the bubble phase is dispersed and the emulsion
nals objectively. phase is continuous.
For a gas–solid fluidized bed, in which most gas
passes through the bed as bubbles, the operational
1.1 Flow Structure of Bubbling Fluidization region of gas velocity is greatly extended by the pre-
sence of bubbles. This is one of the main reasons that
When the gas velocity approaches and goes over the G–S fluidized beds are more widely used in industries
minimum fluidization velocity, Umf , all fine bed parti- than L–S fluidized beds.
cles that belong to Group A in the Geldart particle In addition, the rising bubbles cause the motion of
classification scheme are suspended by the gas. The particles, which obviously intensifies the solids mixing
whole bed expands uniformly, and the gas passes the on the macro scale and leads to the temperature uni-
bed through the particles. This is called the particulate formity and high bed/surface heat transfer character-
fluidized regime. As the gas velocity increases to bub- istic. This is another attractive advantage of G–S
bling velocity, Ub , bubbles begin to appear in the bed. fluidization.
But for the Group B particles, the Umf and Ub are the Consider the case of gas forming a bubble within a
same value. mass of particles. The only way in which the solids can
One of the most characteristic phenomena of gas– be moved out of the way to form a bubble is that there
solid fluidized beds is the formation of gas bubbles, must be a force being exerted by the gas flowing past
which dominate the behavior of fluidized beds in rela- the solids. Most of the gas then finds its way back into
tively low-velocity regimes. In analyzing the behavior the bubble to complete its passage through the bed.
of bubbling fluidized beds, it is essential to distinguish The resulting flow into and out of the emulsion
between the bubble phase and the emulsion phase, the phase enhances the mass transfer. In the case where
latter consisting of particles fluidized by interstitial gas. the bubble eruption and collapse occur at the bed sur-
A bubbling bed can conveniently be defined as a bed in face, there is also intensely enhanced mixing due to
2 CLASSIFICATION OF VARIOUS
Figure 4 Axial distribution of solids concentration in a INTERNALS
CFB.
Internals include tubes, baffles, and other obstacles
inside a fluidized bed. In some cases, for example,
1.4 Flow Structure of Pneumatic Conveying with a strongly exothermic or endothermic reaction
in which substantial amounts of heat need to be
When a continuous particle feeding of a CFB is cut off, removed or supplied from the bed, immersed tubes
particles become dilute and well distributed in the bed. may be a necessity as surfaces for heat exchange. In
Because of the low solids holdup and short residence other cases, baffles or other obstacles can act as inter-
time, this bed can only be used in pneumatic nals to improve the quality of fluidization or to divide
conveying, or in particular chemical reactors, where a bed into a number of stages in parallel or in series to
the reaction rate is very high or extra active catalyst promote smooth fluidized bed operation. It should be
is used. All internals used in risers can be applied in recognized that cyclone diplegs, downflow standpipes,
this situation, so it will not be discussed here. feeding nozzles, and detective probes only come into
In general, nonuniform structures, in both time and contact with a very small part of the bed, and their
space, is widespread in bubbling, turbulent, and fast effects on overall bed behavior are likely to be negligi-
fluidization regimes. On the one hand, such nonunifor- ble. Although the end configurations of the inlet and
mity can enhance the mass and heat transfer of a bed. the outlet of a fluidized bed (i.e., of a CFB) can also
On the other hand, it decreases the contact efficiency of affect the two-phase flow structure, they will not be
gas and solids and makes the scale-up rather difficult. discussed here. Therefore, in this section, only baffles,
Internals are usually introduced not to eliminate the tubes, and some novel geometric structures, which are
nonuniform flow structure completely but to control introduced to significantly strengthen the contact effi-
its effect on chemical reactions. The function of inter- ciency of gas–solids and to improve the fluidization
nals varies in different fluidization regimes, as do the performance, are treated. Internals have different
types and parameters of internals. Taking these pur- structural shapes and arrangements in a bed, hence
poses into consideration, internals may be successfully each has its own performance characteristics.
applied to catalytic reactors with high conversion and Internals reported in the literature are classified into
selectivity, and some other physical processes. five groups as baffles, tubes, packings, inserted bodies,
If you have one of the following four requirements, and other configurations, which are listed in Table 1 in
please do not hesitate to choose a suitable type of detail.
internals:
1. To modify the nonuniform flow structure of 2.1 Baffles
gas–solids, that is, to avoid the overall non-
uniformity on equipment-scale, and to make Baffles can prevent bubbles from growing continu-
the nonuniformity in time much smaller than ously, redistribute bubbles across the cross section of
the characteristic time of a chemical reaction; the bed, strengthen the heat and mass exchange
Category Type Configuration Main features References Aspect studied Application field
Wire mesh 1. Increase bubble splitting Dutta and Suciu Effectiveness of baffles Chemical reaction process
2. Low erosion (1992) (synthetic reaction of ethyl
acetate)
Perforated plate 1. Improve radial Gelperin (1971) Classification of a binary Turbulent fluidized bed
solids distribution Kono (1984) mixture of coal particles Gas–solid fluidized bed
2. Increase bubble Zheng (1990) Bubble frequency, dense bed Circulating fluidized bed
frequency Dutta and Suciu height
(1992) Radial distribution of solids
Zhao (1992) Effectiveness of baffles
Hartholt (1997) Solids backmixing
Mixing and segregation of
solids
Louver plate 1. Enhance gas and solids Jin (1982) Bubble behavior Pilot chemical reactors
exchange
2. Decrease elutriation
3. Improve radial bubble
distribution
Ring 1. Improve radial voidage Zhu et al. (1997) Radial and axial Fast fluidized bed
distribution Zheng et al. voidage distributions Catalytic circulating
2. Enhance radial gas and (1991b,1992) fluidized bed area
solids mixing Jiang et al. (1991) Chemical reaction
3. Improve gas-solids (ozone decomposition)
contact efficiency
4. Suppress axial solids
mixing
5. Increase conversion of
ozone decomposition
reaction
Vertical banks 1. High heat-exchange Volk (1962) Scale up of chemical reactor Turbulent fluidized bed
coefficient
2. Low erosion
phenomenon
3. Low scale-up effect
Station, irregular 1. Keep bubbles small and Huang (1997) Pressure drop Gas–solid–solid fluidized bed
uniformly distributed Solids holdup reactor
2. Reduce carryover of
particles
3. Increase bed expansion
4. Increase chemical
conversion
5. Impede solids motion
Packings
Station, regular 1. Increase pressure drop van der Ham et al. Hydrodynamics and Circulating fluidized bed
2. Increase solids holdup (1991, 1993, mass transfer (CFB)
3. Improve gas–solid 1994) Small scale CFB
contact efficiency
Pilot-plant scale CFB
Inserted Pagoda-shaped (See Fig. 8) 1. Break up bubbles Jin et al. (1982) Bubble behavior Turbulent fluidized bed
bodies bodies 2. Enhance gas–solid Fluid mechanics reactor
contact Pilot-plant trials
Category Type Configuration Main features References Aspects studied Application field
Ridge-shaped (See Fig. 9) 1. Control bubble size Jin et al. (1986) Bed expansion Bubble fluidized bed
bodies 2. Improve the quality of Bed homogeneity Turbulent fluidized bed
fluidization Change of flow region
3. Extend the range of Emulsion phase
turbulent fluidization behavior
Inverse cone 1. Improve radial solids Zheng et al. (1990) Radial solids Circulating fluidized bed
distribution distribution
2. Decrease solids holdup
Bluff bodies 1. Increase pressure drop Gan et al. (1990) Concentration profiles Fast fluidized bed
2. Decrease solids holdup
Inserted
3. Increase gas and particle
bodies
velocities
(cont.)
4. Improve radial voidage
distribution
5. Enhance gas–solid
contact efficiency
6. Reverse radial flow
pattern
Spiral flow 1. Enhance gas–solid Li (1997) Vortex strength Fast fluidized bed
pates contact Cui (1996) Cocurrent and countercurrent
2. Lengthen solids residence flow patterns
time Bed pressure drop
Average density
Swages 1. Decrease pressure drop Davies and Pressure drop Vertical pneumatic
Graham conveying tubes
(1988)
Other
configur-
ations Center (See Fig. 12) 1. Very high solid Milne et al. Stable operation Circulating fluidized bed
circulating circulation rate (1992, 1994) Temperature
tube 2. Uniform residence time Liu et al. (1993) distribution
distribution Wang et al. (1993) Axial solid
3. Simple structure Fusey et al. (1986) distribution
2.3 Packings
the efficiency of gas–solid contact. But they prevent the try, and they have effectively improved the synthesis
particle movement so seriously that the bed would yield of acrylonitrile.
easily be defluidized. In the 1990s, a large research effort was put into the
Up to the 1960s, packings in fluidized beds received second group and some equipment has been devel-
considerable attention in North America, and beds oped. For example, the gas–solid–solid circulating
filled with packings have fallen into two groups fluidized bed reactor (GSSCFR) is such a novel reactor
(Grace and Jarrison 1970): that can carry out, in addition to a primary reaction, a
heat transfer, a mass transfer, or a momentum transfer
1. Screen-packed beds in which the packings con-
operation or even another reaction simultaneously. In
sist of small open-ended wire mesh screen
a GSSCFR, a packed bed and a fluidized one are inter-
cylinders
connected as shown in Fig. 7. Large catalyst particles
2. Packed fluidized beds in which small particles
are fixed in the packed bed. Another amount of fine
are fluidized in the interstices of much larger
particles is fluidized, which is often used as an adsor-
unfluidized particles
bent or a heat carrier, circulating between the packed
Compared with other types of packings, screen cylin- bed and the fluidized bed. For reversible reactions, the
der packings occupy only 5% of the bed volume and powder adsorbent can selectively adsorb and remove
thus attract special attention. There were many inves- the desired product while it flows through the packed
tigations of it on gas–solid contact, bed expansion, bed, so the reactor can shift the equilibrium in favor of
bubble size, solids entrainment, fluidization quality, the desired product. For strongly exothermic or
and effect on chemical reactions. It was found that endothermic reactions, the powder can remove or
the screen cylinder type tends to keep bubbles small supply the heat of reaction and make a uniform tem-
and uniformly distributed so that slugging is perature distribution in the reaction zone.
suppressed, carryover of particles is reduced, bed In a GSSCFR, the behavior of a gas–solid two-
expansion is increased, and chemical conversions are phase concurrent flow through the packed bed greatly
increased. Although metal screens have many advan- affects the mass and heat transfer. The gas and powder
tages, they are easy to transfigure in fluidization beds flow concurrently through a packed bed to intensify
for poor strength, and cannot be operated for a long the gas–solid contact (Yamaoka, 1986; Shibata et al.,
period of time. Moreover, solids motion is impeded so 1991; Song et al., 1995; Van der Ham et al., 1996).
that heat transfer rates are less favorable than in Huang et al. (1996a,b) extended the research started
corresponding unpacked systems, and segregation of by Song et al. (1995), and examined the axial distribu-
particles may be significant. Dead spots and channel- tion of powder holdup in the packed bed; they devel-
ing also tend to occur, and it may be difficult to dump oped a mathematical model to predict the powder mass
the fluidized particles from a bed containing fixed flux and the mean dynamic powder holdup. They
packings. But till now, chemical reactors packed with found that the powder mass flux is influenced by the
metal screens are still applied in the acrylonitrile indus- powder inventory, the gas velocities, and the void of
the packed bed. More powder inventory, less gas velo- heat effects. The favorable bed/surface heat transfer
cities, and larger voids of the packed beds are of benefit rates and solids mixing characteristics of gas fluidized
for increasing the powder mass flux and hence for beds result from the motion of particles, motion caused
enhancing the heat and mass transfer. by rising bubbles; therefore it is important not to
A GSSCFR exhibits good operating stability and impede the particle movement greatly. However,
simple construction and can be operated at a wider most packings in a fluidized bed seriously prevent the
range of superficial gas velocity than a gas–solid– movement of particles, and they are harmful to heat
solid trickle flow reactor (Yamaoka, 1986; Shibata et transfer too. Though the effective thermal conductivity
al., 1991; Song et al., 1995). The experimental results of of a packed fluidized bed can be about 75 times that of
oxidation–dehydrogenation of butene to butadiene a normal packed bed (Ziegler, 1963), it is still lower
show that the C4 H8 conversion and C4 H6 selectivity than that of a fluidized bed. It seems that packed flui-
can be significantly increased in a GSSCFR (Huang et dized beds are only useful when it is not practicable to
al., 1998). divide the solid material into particles fine enough to
Some early Soviet work (Goikhman et al., 1968) permit fluidization at reasonable fluid flow rates, or
suggests that packings (which are much larger than when a highly exothermic reaction takes place on the
the fluidized particles) may be more effective if they surface of the large fixed particles.
are light enough to float freely in the bed instead of
being essentially fixed in position as described above. 2.4 Insert Bodies
This appears to be a peculiar approach, but more work
is necessary before floating packings can be evaluated It has been suggested that a combination of horizontal
with confidence. and vertical internals has an important effect on
Despite the considerable amount of research work improving fluidization quality. Thus various kinds of
that has been carried out in the laboratory on packings insert bodies come into being, such as: pagoda-shaped
in fluidized beds, the technique has not yet found a bodies by Jin et al. (1982), ridge-shaped bodies by Jin
place in large-scale industrial applications because of et al. (1986b), inverse cones by Zheng et al. (1990), half
its fatal weakness of preventing solids movement. oval bluff-bodies by Gan et al. (1990), and spiral flow
It is well known that, compared with a packed bed plates by Li (1997).
reactor, the most outstanding advantage of a fluidized On the basis of an empirical investigation on the
bed reactor is its strong ability for heat exchange. That effects of the behavior of the bubbles in fluidized
is why it is especially suitable for processes with strong beds and different types of internals on the motions
cones and bluff bodies. According to Gan et al. (1990), 2.5 Other Configurations
by forcing gas and solids to flow upward in the narrow
ring section near the wall, half oval bluff bodies can There are also other elements in fluidized beds, not
reverse the radial flow distribution in a circulating flui- installed as internals, but necessary for some purposes.
dized bed and cause higher solids concentration and One example is that by Davies and Graham (1988).
lower gas and solids velocities in the core, and lower They swaged indentations 3.5 mm deep and 16 mm
solids concentration and higher gas and solids veloci- wide onto the wall of a riser 0.152 m in diameter to
ties in the wall region. Gas and solids are forced to flow test their effect on pressure drop in the riser. They
upward in the narrow annular region near the wall. At found that swages evenly spaced along the circular
the same time, the gas–solid mixing along the radial riser wall would scrape the dense downflow solids
direction could be also greatly intensified. away from the wall and cause higher bed voidage
Besides the single- or multiple-turn plate, the spiral and lower pressure drop in the fully developed region
flow plate is another kind of internal that may lead to than those in a bare tube.
gas and solid swirling. The main functions of a spiral Another example is the strange ‘‘nose’’ in a circulat-
flow plate are lengthening the residence time of solids ing fluidized boiler. As we know, the circulating coal
and enhancing the contact between gas and solids. ash usually feeds into a boiler from one side of the bed
There are many kinds of spiral flow plates, with differ- wall, which easily results in annular flow and decreases
ent structures and opening area ratios, and even with the burning efficiency. Thus a protruding object like a
different arrangements, which are applied in drying ‘‘nose’’ is set at the other side in order to prevent the
processes (Fig. 11). The vortex strength is a common nonuniform flow.
index to evaluate them. Li (1997) found that a reversed The third example, fast fluidization, can also be
conic plate is better than a conic spiral flow panel in achieved in a concentric circulating fluidized bed.
PVC drying processes. One option is to have solids flow up in the center
Generally, the introduction of solid obstacles into a and down in the annular shell. This has been called
bed further complicates the system, and there is as yet an internally circulating fluidized bed. Alternatively,
no relevant theoretical analysis available. one can have solids flow up in the annular shell and
flow down in the center, leading to what has been et al. (1992a,b) and by earlier workers (Fusey et al.,
called the integral circulating fluidized bed. These 1986) show stable operation of this system. One advan-
two options are shown in Fig. 12. Since both have tage is that a very high solids circulation rate (near
been abbreviated to ICFB, the former will be identified 600 kg=m2 :s) can be realized in this system (Milne et
as i-ICFB and the latter as o-ICFB with the i and o al., 1994). Milne et al. (1994) have shown that a stable,
indicating the inside and outside location of the riser. isothermal operation at high temperature can also be
The i-ICFB originated from the spouted bed with a achieved in an i-ICFB unit and have suggested that it
draft tube. Preliminary hydrodynamic studies by Milne can be ideal for short contact time reactions where
uniform residence time distribution and stable high uniform solids feeding at the riser bottom and the
temperature operation are needed, such as in pyrolysis. energy requirements of a large amount of external
o-ICFB has been developed as an alternative to the solids separation and recirculation. There are several
regular CFB systems to address the concerns of non- advantages: (1) simple structure; (2) high solids circu-
Figure 12 Structure of integral circulating fluidized bed. mount importance or where an existing system does
not perform as well as expected, vertical tubes warrant
serious consideration. Clearly, combinations of various
lation rate; (3) uniform solids feeding. As for i-ICFB,
kinds of internals are likely to be useful in certain
the solids circulation rate is not independently con-
applications.
trolled but varies with riser and auxiliary gas velocities.
Liu et al. (1993a,b) and Wang et al. (1993) found that
axial solids distributions in the riser of o-ICFB are very 3 EFFECT OF INTERNAL TUBES AND
similar to those in regular CFB risers. The cross- BAFFLES
sectional average solids holdup is somewhat higher
than that in regular CFB risers operated under similar Many references in the literature mention that the
conditions (Wang et al., 1993). Typical radial solids improved fluidization results from the use of tubes
concentration profiles measured by Wang et al. and baffles as internals. Harrison and Grace (1971)
(1993) show that a ‘‘core-annulus’’ structure still exists and Sitnai and Whitehead (1985) have reviewed the
inside the annular riser, with the dense layer at the influence of internals on the behavior of fluidized
outer wall slightly thicker than that on the inner wall beds for reactor design purposes. For example, on
(Fig. 13). Measuring at two vertical planes in the annu- the one hand, most transverse baffles in a fluidized
lar riser on the opposite sides of the central standpipe, bed effectively narrow the gas–solid residence time dis-
Wang et al. (1993) found that the radial voidage tribution, prevent rushing at the surface of the bed,
distributions are symmetrical. and reduce the entrainment of solid particles. On the
It has already been stressed that it is impossible to other hand, a disadvantage of the use of baffles in a
specify a single type of internal that is best for all bed is the presence, underneath each baffle, of a region
applications. This is because each successful applica- of lower density where the solid particles are more
tion of fluidization depends upon different properties dilute. Another disadvantage is the axial size-classifica-
of fluidized beds to varying degrees, and a type of tion of particles. Furthermore, increasing the diameter
internal that helps one process may hinder another. of the fluidized bed magnifies these problems.
For example, particle segregation is an adverse effect In this section, the effect of internals on a fluidized
in many applications, but not, of course, if a fluidized bed is analyzed from the following four aspects.
bed is used as a particle classifier; then an internal
member that promotes segregation (i.e., fixed pack- 3.1 Bubble Behavior
ings, horizontal screens) will be appropriate. In many
continuous operations, horizontal baffles may be of Bubble size is a very important parameter in the design
considerable benefit. Where heat transfer is of para- and simulation of fluidization. We can accept that
most of the effects that internals produce in a fluidized past each layer of baffles, only to collapse when they
bed are derived from the same primary effect: control reach the surface of the bed. With the increase of gas
of the bubble size and its distribution. The movement velocity, a small, nonstationary, particle-free section is
of bubbles in a bed containing no internals has already observed underneath a baffle with bubbles splitting
been discussed in Sec. 1.1. above it. At much higher gas velocities, the effects of
Jin et al. (1982) have investigated the ability of inter- baffles become apparent—breakup, regeneration, and
nals to break up bubbles in a two-dimensional fluidized uniform distribution of bubbles. When the superficial
bed by the photographical method (Fig. 14). For com- gas velocity further increases, stationary particle-free
parison, Fig. 14a shows the bubbles in a free fluidized sections are observed under each baffle and smaller
bed. When an internal module is introduced, such as a bubbles are formed above them. The particles, which
baffle, a tube or a pagoda-shaped body, the situation is follow the bubbles upwards, mainly circulate through
modified to different degrees according to the module the bed in one of two ways: through the spaces around
type, as shown in Figs. 14b,c,d. the baffles, or directly past the louvers in the baffles.
Figure 15 shows the effect of transverse baffles Figure 15b illustrates the rain of particles, which is
(louver plates) on the behavior of bubbles at different observed coming off the louver plates. Under these
gas velocities. The baffles do not visibly influence the conditions, contact between the gas and solids is
bubble behavior (coalescence or rise velocity) at low excellent.
gas velocities (Fig. 15a). Small bubbles push upward
Figure 15 Effect of transverse baffles (louver plates) on the behavior of bubbles at different gas velocities.
and intensifies turbulence. Also, there is an increas- Besides splitting bubbles, internals alter the eruption
ingly rapid exchange between the bubble and emulsion mechanism for single bubbles at the bed surface,
phases throughout the bed. causing a significant fraction of the single bubbles to
In general, by catching and breaking up the bubbles, generate large wake eruptions. Levy et al. (1986) have
the internals have the effect of reducing their size. On found that without any internals in the bed, when sin-
one hand, when a bubble is split by an internal, the gle bubbles erupt, the bulge material of a bubble
sum of the volumes of the daughter bubbles is less than always reached a greater height above the bed free sur-
the volume of the parent one, indicating that some of face than the wake material. With horizontal tubes
the original bubble gas leaks into the emulsion phase present, up to 60 percent of the single bubble eruptions
of the bed during the splitting process. Furthermore, are characterized by the upward projection of wake
the amount of leakage increases as the number of material, reaching a height above the free surface
daughter bubbles increases. These facts could be of which far exceeds those attained by the bulge material.
importance in the design of fluidized bed reactors. The wake-dominated eruption of the single bubble
With the reduction of the mean bubble size, the appears to be due to the formation of elongated ver-
expansion ratio of a bed changes. Figure 18 shows tical cavities in the region between the internals and the
quite clearly the effects that different types of internals free surface. A quantitative understanding of the
have on the expansion ratio in a two-dimensional flui- effects of internals on bubble eruptions may lead to
dized bed. On the other hand, small bubbles rise rela-
tively slowly in a fluidized bed, and so the result is to
increase the total concentration of bubbles in the bed.
A thick mist of small bubbles has a beneficial effect on
the conversion of a chemical reaction.
Two different methods have been developed for pre-
diction of the bubble size in a bed with internals
formed by horizontal rows of tubes (Jodra and
Aragon, 1983). One, based on a new correlation,
makes it possible to find the size of a bubble above
the tube as a function of the operating conditions
and characteristics of the tube arrangement. The
other is based on the general theory of bubble forma-
tion in submerged orifices. Both methods can be used
to estimate bubble sizes, which are acceptably close to
the experimental values, and can be applied to other
types of internals, such as perforated plates, meshes
and arrangements of horizontal bars of noncircular Figure 18 Effects of internals on the expansion ratio in a
cross sections. two-dimensional fluidized bed.
Figure 1 Attrition modes and their effects on the particle size distribution (q3 ¼ mass density distribution of particle sizes dp ).
Figure 2 SEM images of two catalyst materials. The surface mean diameter was 112 microns in the case of the ‘‘97-R(fresh)’’
and 103 microns in the case of the ‘‘97-G(fresh).’’ (From Reppenhagen and Werther, 1999a.)
and Granatstein, 2001). In any case, these are gradu- A batch of catalyst particles was passed in a first
ally acting phenomena, that are strongly related to the phase through a cyclone under conditions of pure
influence of the solids residence time, the effect of abrasion, then in a second phase under conditions
which is discussed in Sec. 2.2. But there may be also where additional surface fragmentation takes place,
singular phenomena that affect the materials attritabil- and finally in a third phase again under conditions of
ity. A broken steam pipe in the stripping unit of an the first phase of pure abrasion. It is quite obvious that
FCC plant may for example cause a steam jet that the surface fragmentation leads to an increased sus-
generates extra stress on the catalyst particles. Such a ceptibility to abrasion, i.e., at the end of the experiment
singular event would not only cause a higher extent of the abrasion rate is distinctly higher than at the begin-
attrition during its occurrence but also it would most ning, even though the operating conditions are identi-
probably have a long-term effect. The reason for this is cal. This can be explained by comparing SEM images
a serious damage of the morphology of the particles. taken from the catalyst particles before and after the
Reppenhagen and Werther (1999a) observed such a subjection to the excessive stress (cf. Fig. 5). It can be
phenomenon during their cyclone attrition tests as seen that parts of the particles’ surfaces have been
they exceeded, for several passes through the cyclone, chipped off, leading to a new surface that is distinctly
a certain threshold energy. The result of this experi- rougher than the old one and therefore more prone to
ment is shown in Fig. 4. abrasion. However, with an increasing number of
Figure 4 Influence of surface fragmentation on the cyclone loss rate at conditions of pure abrasion (uc;in ¼ 20 m=s, c ¼ 0:1).
(From Reppenhagen and Werther, 1999a.)
passes under conditions of pure abrasion, the rough between particles can be extremely high and result in
surface, which is responsible for the increased loss significant particle breakage.
rate, would become smoother again and the cyclone
loss rate would then decrease to its initial value. 2.2.2 Wall-Hardness
One can assume that the particle degradation increases
2.2 Process Conditions with the hardness of the vessel wall. This effect will
increase with increasing ratio of particle-to-vessel dia-
The process conditions, which are resulting from the meter and may thus be only relevant inside a cyclone,
system design and the operating conditions, will pri- or in the vicinity of bed inserts.
marily influence the attrition of the bed material by
generating the stress on the individual particles. In 2.2.3 Solids Residence Time
general, the stress leading to attrition of a given bulk The relationship between the solids residence time and
material may be a mechanical one due to compression, the amount of material that is produced by attrition is
impact, or shear, a thermal one owing to evaporation generally nonlinear. As an example, Fig. 6 shows the
of moisture or temperature shock, or a chemical one by typical time dependence of the attrition of a fresh cat-
molecular volume change or partial conversion of the alyst that is subjected to attrition in one of the test
solid into the gas phase. devices described in Sec. 4.
The elutriated mass is defined to be the attrition
2.2.1 Gas and Solids Velocity product, and consequently the attrition rate is defined
as elutriated mass per unit time. It is clearly seen that
The gas velocity is usually directly related to the par- the rate of attrition is decreasing with time. The reason
ticle velocity, which is the most important factor in is that at the beginning the fresh catalyst particles have
generating the mechanical stress by interparticle colli- very irregular shapes and contain many faults. This
sions or by particle–wall impacts (British Materials results in a high rate of initial particle degradation,
Handling Board, 1987). The forces involved in the during which the particles break and their edges and
degradation process may be generated by high-speed asperities are knocked off. With progressing time, the
collisions resulting preferably in breakage. particles become smaller, rounder, and smoother, and
Alternatively, the energy may be transmitted through the number of their weak points decreases. The elutria-
a matrix of comparatively slowly moving particles tion rate therefore decreases continuously with time
resulting mostly in abrasion. Particularly in the distri- and tends to a more or less constant value, which
butor region of a fluidized bed, where grid jets are can be interpreted as a kind of steady-state level
issuing into the bed of particles, the impact velocities where only abrasion takes place.
The shape of the curve in Fig. 6 is typical for all or an indirect way, i.e., thermal shock, changes in
particle degradation processes. However, one has to particle properties, changes in the gas density.
distinguish between batch and continuous processes. The heating of fresh cold particles fed into a hot
In batch processes the whole bed material is always process can cause various phenomena that may lead
at the same state of attrition, which changes during to particle degradation. These are thermal stress,
the operating time according to Fig. 6. On the other decrepitation, evaporation of moisture, hydrate
hand, there is a particle residence time distribution in decomposition, and impurity transformation. On the
continuous processes. The attrition rate of the entire other hand, particle properties such as strength,
material is therefore constant, although the state of hardness, and elasticity may also be affected by the
attrition of each particle again changes according to temperature. With respect to the resistance to degrada-
Fig. 6. tion, there is an optimum temperature range for any
Many attempts have been made to describe the time specific type of material. At lower temperatures, parti-
dependence of the attrition rate in batch fluidized bed cles become brittle, while at higher temperatures they
processes. Gwyn (1969) studied the degradation of cat- may soften, agglomerate, or melt and lose discrete par-
alysts in a small-scale test apparatus and defined the ticulate properties. Consequently, it is important that
elutriated particles as the only attrition product. He the particle friability, which is the major factor for the
described the increase of the elutriated mass mloss attrition propensity of a given material, is assessed
with time t based on the initial solid bed mass mb;0 under conditions that are similar to those found in
by the now widely known Gwyn equation, the process where the attrition is occurring.
Moreover, the temperature can have a strong effect
mloss
¼ Ka
tb ð1Þ on the gas density, which affects the fluidization state
mb;0 and with it the particle motion and the stress the par-
ticles are subjected to.
For several batches of the same catalyst material with
quite different mass median diameters, Gwyn found 2.2.5 Pressure
the exponent b to be constant, whereas the attrition
constant Ka was found to decrease with mean particle The absolute pressure is unlikely to have a direct effect
size. Equation (1) is therefore valid for a particular size on attrition unless it affects the amount of adsorbed
distribution only. Other empirical correlations for the surface layers. But there is again an effect on the gas
time dependence of the mass ratio mloss =mb;0 have been density that is similar to the effect of temperature men-
suggested by e.g. Pis et al. (1991) and Dessalces et al. tioned above. Moreover, the rate of pressure change
(1994). may have more influence.
Figure 7 Changes in the particle size distribution of a FCC catalyst material that was subjected to cyclone attrition. (Zenz and
Kelleher, 1980.) The respective experimental procedure is described in Sec. 5.3.
velocity are used, then no relative attritability informa- larger particles tend to slug in the small cylinder.
tion can be obtained. However, in the catalyst development, where at first
In addition to its use as a pure friability test, this only a little batch of catalyst is available, this appara-
type of test has also been used in the above-mentioned tus is an important friability test, because it requires
grid-jet modeling approach of Ghadiri and coworkers only a small amount of material (approximately 5 to
(Ghadiri et al., 1992a, 1994, 1995; Boerefijn et al., 10 g).
1998, 2000). In this approach the particles were fed
in a single array to the eductor.
4.3.3 Fluidized Bed Tests
Fluidized bed tests may be used for both purposes, the
4.3.2 Grace–Davison Jet-Cup Attrition Test
determination of the catalysts’ friability (Forsythe and
The Grace–Davison jet-cup attrition test is often used Hertwig, 1949; Gwyn, 1969) and the investigation of
to test the friability of catalysts (Weeks and Dumbill, attrition mechanisms (Werther and Xi, 1993). Most
1990; Dessalces et al., 1994). The jet-cup apparatus is fluidized bed tests are currently carried out as so-called
sketched in Fig. 9. submerged-jet tests, where high-velocity gas jets sub-
The catalyst sample is confined to a small cup, into merged in a fluidized bed produce high attrition rates
which gas is tangentially added at a high velocity in a well-defined short period of time. The majority of
(about 150 m/s). After a test run over a period of an these tests are based on the device suggested by
hour, the so-called Davison Index (DI) is determined Forsythe and Hertwig (1949) (Fig. 10a).
by measuring the increase in the weight fraction of The setup consists of a 0.0254 m ID and 1.52 m long
particles below 20 microns. The increase is determined glass pipe, which bears a canvas filter at its upper end,
from both the fraction of elutriated fine material and a and which is sealed by an orifice plate at the bottom.
particle size analysis of the remaining fraction in the This latter plate contains a single 0.4 mm ID orifice in
cup. its center. The apparatus is operated in such a way that
Some authors (Dessalces et al., 1994) assume that the jet gas velocity approaches the speed of sound in
the stress in the jet-cup is similar to that prevailing in the orifice. The filter keeps all material inside the sys-
gas cyclones. With respect to fine catalysts, this type of tem. To assess the degradation extent, one should
test works as good as the impact test described above, screen the material by wet sieving through 325 mesh
but its applicability is limited to smaller sizes because (44 mm). The attrition rate is defined as the ratio of the
increase in weight percent of 325 mesh material and 50 m ID and 1000 mm height, and one with
the weight percent age of þ325 mesh particles in the column C having 100 mm ID and 1110 mm height.
initial material. Using this test procedure, one has to Although in all columns the same stress was applied
take into account that the attrition-produced fines that to the material and the same gas velocity in the
are kept in the system may affect the material’s attrit- enlarged top sections was adjusted, column A led to
ability. Moreover, there are difficulties in using this test significantly higher results in the measured elutriated
as a friability test, because the results are time inte- mass. From a comparison of the particle size distribu-
grated and have to be assessed with the help of the tion of the elutriated material, it could afterwards be
particle size distribution. The materials that are to be concluded that the height of column A had not
compared should therefore have a similar initial parti- exceeded the transport disengaging height (TDH),
cle size distribution and an identical pretreatment. and that coarser particles were allowed to reach the
These difficulties are avoided in the design suggested gas outlet. From this observation, Werther and Xi
by Gwyn (1969), shown in Fig. 10b. Here, the attrition (1993) suggested that as a design rule for this type of
products are not kept inside the system; it is rather attrition test, the height between the bed surface and
assumed that they are elutriated. In the enlarged the gas outlet should at least exceed the TDH.
diameter top section gravity separation defines the Moreover, it should be noted that, when one com-
limiting diameter of the elutriable particles. The attri- pares different materials’ test results, the density of the
tion rate is assumed to be given by the elutriation rate. particulate materials must be taken into account. If the
The steady-state elutriation rate can therefore be used gas mass flow and the temperature are kept constant,
as a friability index. then a variation in the solids density will result in a
It should be noted here that the quantitative results shift of the cut size and thus in the amount of material
obtained in a Gwyn-type attrition apparatus will in collected as ‘‘attrition product.’’ Another point con-
general depend not only on the cut size of the gravity cerns the particle size distribution of the bed material.
separator but also on the entrainment and elutriation If the original solid sample is prepared by sieving e.g.
conditions in the main column. Werther and Xi (1993) or sifting so that the smallest size is significantly larger
compared, for example, attrition test results of the than the cut size of the gravity separator (Werther and
same catalyst obtained from three differently sized Xi, 1993), one can be fairly sure that the elutriated
Gwyn-type units, one with column A having 50 mm material is indeed due to attrition. On the other
ID and 500 mm height, one with column B having hand, if the cut size of the gravity separator is located
Figure 11 Experimental setup for the investigation of cyclone attrition. (After Reppenhagen and Werther, 1999a.)
Hence it can be concluded that for each operating must be taken into account. As for the latter para-
condition the fines concentration in the material tends meter, Reppenhagen and Werther (1999a) have given
to a characteristic value of which the accumulation of a simple example. They compared the test results of a
fines is balanced by the release of fines. When this previously screened fraction with those of an
characteristic concentration is reached, the loss rate is unscreened fraction of the same material. By taking
at steady state, i.e., it is equal to the production rate of the differences in the mean particle size into account
fines. Reppenhagen and Werther (1999a) suggested (cf. Sec. 5.3), they observed a distinctly higher loss rate
that we take this steady-state value as a characteristic (approximately 25% higher) in the case of the
value for both the assessment of a material’s attritabil- unscreened fraction owing to the additional loss of
ity and the study of cyclone attrition mechanisms in sufficiently shrunk mother particles.
dependence on the various influencing parameters. For In general, it must be recognized that the greater
this purpose, they defined the cyclone attrition rate as effort, and the large amount of material required, for
the currently available cyclone attrition test methods
m_ loss;c;steady-state m_ a;c;fines
rc ¼ ¼ ð6Þ will generally prevent them from being used as a pure
m_ c;in m_ c;in friability test, i.e., a test for a sole ranking of various
materials without any reference to a particular process.
where m_ loss;c;steady-state is the cyclone loss flow at
However, the tests are useful for both the study of
steady state, m_ a;cfines is the production rate of fines
general attrition mechanisms in cyclones and the
due to attrition inside the cyclone, and m_ c;in is the
evaluation of a given material’s susceptibility to the
mass flow into the cyclone.
particular attrition stress inside cyclones, which may
It should be noted that Reppenhagen and Werther
be needed as an input parameter in a cyclone attrition
(1999a) have exclusively concentrated on the attrition
model.
of FCC and fluidized bed catalysts, the abrasion-
produced fines of which were throughout all the
experiments entirely elutriable. In other words, inde-
pendently of the cyclone, the operating conditions, and 5 SOURCES OF ATTRITION IN A FLUIDIZED
the catalyst’s material density, the produced fines were BED SYSTEM
sufficiently small to be in a particle size range in which
the grade efficiency of the cyclone is smaller than unity. A first approach to finding attrition in fluid beds was
However, for a transfer of the described test concept to made by Zenz (1971). He pointed out that there are
other investigations, the same attention must be paid various regions in a fluidized bed reactor system in
as in the case of the Gwyn-type tests, i.e., when com- which the stress acting on the bed particles and the
paring results with different materials the density of the corresponding attrition mechanisms are quite different.
particulate material and its particle size distribution In the subsequent works (Zenz, 1974; Vaux and
m_ abr:;fines;j ¼ nor
Cj
dpb
rf
dor
2
u3or ð9Þ
Figure 15 Variation of the steady-state attrition-induced Figure 16 Variation of the steady-state attrition-induced
loss rate of coal ash (0.2 to 0.315 mm) with the excess gas loss flow of limestone fractions with (u-umf ). (Ray et al.,
velocity (u-umf ). (Pis et al., 1991.) 1987a.)
two groups (Zenz, 1974; Zenz and Kelleher, 1980; m_ loss;c;steady-state m_ abr:;fines;c
Reppenhagen and Werther, 1998, 1999a; Werther rc ¼ ¼ ¼ Kc
mnc
u2c;in
m_ c;in m_ c;in
and Reppenhagen, 1999). The experimental procedures
ð19Þ
of both groups have already been explained in Sec.
4.3.4. where Kc is the cyclone attrition rate constant, which
As a result of their experiments with FCC catalyst summarizes all particle properties that are relevant to
material in isolated cyclones, Zenz (1974) and Zenz the abrasion process, and n is an exponent assuming a
and Kelleher (1980) observed even after a few passes power law dependence of the attrition rate on the
a significant change in the particle size distribution solids-to-gas loading ratio mc at the entrance of the
(cf. Fig. 7). The content of coarser particles was cyclone. According to the procedure described in Sec.
decreased in comparison to their initial one, whereas 4.3.4, the authors carried out comprehensive attrition
the content of smaller particles was increased. From tests with a spent FCC catalyst and a 90 mm ID
many experiments of this kind, Zenz and Kelleher cyclone in order to scrutinize this model approach. In
these experiments the gas velocity uc;in at the entrance
of the cyclone was varied between 8 and 24 m/s with
the solids loading mc ranging from 0.05 to 1. In Fig. 19
the measured cyclone attrition rates rc obtained under
steady-state conditions are plotted against the cyclone
inlet velocity on a double logarithmic grid.
The relationship between rc and uc;in predicted by
Eq. (19) is seen to be confirmed over a wide range,
which is indicated by the solid lines. However, as
indicated by the dashed lines, there are some distinct
deviations when the gas velocity exceeds a certain
threshold. The threshold on its part seems to be depen-
dent on the solids loading, i.e., the smaller the solids
loading the smaller the gas velocity from which the
deviation occurs. Reppenhagen and Werther (1999a)
explain this deviation by the occurrence of another
Figure 18 Bubbling bed attrition test results obtained by attrition mechanism in addition to abrasion, namely
Werther and Reppenhagen (1999) in a 200 mm diameter the so-called surface fragmentation of the catalyst par-
Gwyn-type test plant under pure abrasion conditions with ticles, which results from the combination of increasing
fresh FCC catalyst. kinetic energy at increasing velocities and increasing
u2
rc ¼ Kc
pc;in
ffiffiffiffiffi ð20Þ
mc
Figure 21 Experimental setup for the investigation of the Figure 22 Measured loss flows from the fluidized bed sys-
attrition-induced loss flow of fresh FCC catalyst from a flui- tem shown in Fig. 21 operated with previously screened fresh
dized bed system. (After Werther and Reppenhagen, 1999; FCC catalyst. (After Werther and Reppenhagen, 1999;
Reppenhagen and Werther, 1999b.) Reppenhagen and Werther, 1999b.)
In a first step, they identified three regions as main with Dt being the diameter of the fluidized bed column,
attrition sources, namely the grid jets, the bubbling bed
m_ c;in
itself, and the cyclone section. Accordingly, they mc ¼ ð25Þ
described the quasi-stationary loss rate of the overall rf
uc;in
Ac;in
system as the sum of their individual contributions: and
Since the stress conditions throughout all the experi-
At
ments could be regarded to be under pure abrasion uc;in ¼ u
ð26Þ
conditions, the model equations Eqs. (9), (18), and Ac;in
(22) could be applied to describe the generation of which leads to the following completed model equation
elutriable fines; the following reasoning was made: for the overall process:
Since the abrasion-produced fines are rather D6t
small, i.e., typically smaller than about 1 to 3 m_ loss;tot ¼ Cj
dpb
rf
u3
4
dor
n2or
microns, the respective grade efficiency of the
cyclones is distinctly smaller than unity. þ Kb
mb
ðu umf Þ3
Therefore it can be assumed that the fines are
pffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffi A2:5
immediately lost after their production, even if þ Cc
dpc
m_ c;in
rf
2t
u2:5 ð27Þ
they are produced in the jetting region or in the Ac;in
bubbling bed. In contrast to this, it can be
From this model equation (27), it is obvious that the
assumed that after an initial phase where the ori-
attrition effects in the overall system depend not only
ginal elutriable material is sifted off, the mother
on the prevailing attrition mechanisms but also on the
particles are entirely kept inside the system. Due
solids transport in the freeboard which determines the
to the mode of pure abrasion the change in their
solids flow into the cyclone and thus the amount and
particle size is negligible, and so is their contribu-
the particle size distribution of the material that is sub-
tion to the loss rate. Hence some kind of steady
jected to cyclone attrition. Hence in a second step of
state can be assumed where the production of
the overall modeling, these transport effects must be
fines, which originate from the almost unchanged
taken into account, which has in a first approach
mother particles, is balanced by the loss rate of
been simply done by measuring the necessary data.
fines.
Figure 23 shows a comparison of the experimental
Hence we can write: data depicted in Fig. 22 with the calculation from the
model equation (27). The required attrition rate
m_ loss;tot;steady-state ¼ m_ loss;j;steady-state þ m_ loss;b;steady-state constants Cj , Kb , and Cc that describe the materials
susceptibility to attrit in the respective regions have
þ m_ loss;c;steady-state
been determined by the corresponding attrition tests
¼ m_ abr:;fines;j þ m_ abr:;fines;b as described in Sec. 4.3. Cj has been determined from
exactly that Gwyn-type test facility that is shown in
þ m_ abr:;fines;c
Fig. 14 and was set to zero in the case of the porous
2
¼ nor
Cj
dpb
rf
dor
u3or plate distributor; Kb has been measured in a 200 mm
ID Gwyn-type test apparatus, and Cc has been
þ Kb
mb
ðu umf Þ3 determined from exactly that cyclone attrition-test
u2 procedure that is described in Sec. 4.3.4 using the
þ m_ c;in
Cc
dpc
pc;in
ffiffiffiffiffi equipment sketched in Fig. 11. The parameters m_ c;in
mc
and dpc were measured in the apparatus sketched in
ð23Þ Fig. 21 under the assumption that m_ c;in may be
þ Kb
mb
ðu umf Þ3
A3t ð31Þ
þ 3:81
Cc
dpc ðuÞ
rf
A2c;in
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
X u
wi
exp 5:4 ti
u3:75
i
u
Figure 23 Comparison of measured and calculated loss On the basis of Eq. (31), the total loss rate and the
flows for the two different system designs. The input data contributions of the individual sources of a given sys-
for the calculation that are affected by entrainment had tem can be a priori calculated in dependence on the
been measured. (After Werther and Reppenhagen, 1999; superficial gas velocity. Using the same values for the
Reppenhagen and Werther, 1999b.) material’s attrition rate constants and the system
design as in the above considerations, this is demon-
strated in Fig. 24.
approximated by the solids circulation rate measured Obviously, there is a strong sensitivity of cyclone
in the return line. Figure 23 shows a very good agree- attrition inside a fluidized bed system against the
ment between the measured and the calculated overall superficial gas velocity. It is significantly higher than
loss flows even though two different process designs are the sensitivities of jet- and bubble-induced attrition,
considered. respectively. Furthermore, it is obvious that owing to
However, as mentioned above, Eq. (27) requires these different dependencies, the role of the main attri-
information on the solids transport effects. Hence for tion source changes with the gas velocity. In the lower
an a priori modeling, the implementation of an velocity range—in this particular example below
entrainment/elutriation model is required. Werther 0.55 m/s—the gas distributor is the main attrition
and Reppenhagen (1999) have given an example of source, whereas at higher velocities cyclone attrition
such a modeling approach: the bed particle size is dominant.
distribution and thus dpb were simply assumed to be
constant, and mc;in and dpc were substituted by
X
m_ c;in ¼ At
Gs ¼ At
Gsi
i
X
¼ At
wi
Ki ð28Þ
i
G G
dpc ðuÞ ¼ P s ¼ P s ð29Þ
Gsi w i
Ki
i xi i xi
the interval. Even if there were a further particle degra- ences of geometry and operating parameters. With the
dation, the attrition products would remain within the definition of the surface mean diameter, Eq. (35) can
size interval K1 . On the other hand, the particles within be written as
the size interval Kn undergo abrasion. This results in X X
both the loss of fine material and the loss of shrunk m_ loss ¼ m_ loss j ¼ C
x i
Q2;i ð36Þ
Figure 25 Abrasion-induced mass transfer between particle size intervals. (From Reppenhagen and Werther, 2001.)
Thomas C. Ho
Lamar University, Beaumont, Texas, U.S.A.
The conversion in gas–solids fluidized bed reactors has The first attempt at modeling gas–solids fluidized bed
been observed to vary from plug flow, to well below reactors employed ideal or simple one-parameter
mixed flow, mainly depending on reaction and fluidi- models, i.e., plug flow, complete-mixed, dispersion,
zation properties (Levenspiel, 1972). Historically, two and tank-in-series models. However, the observed
classes of models have been proposed to describe the conversion of sometimes well below mixed flow in
performance of fluidized bed reactors; one is based the reactors could not be accounted for by these mod-
on a pseudohomogeneous approach and the other on els, and the approach was dropped by most research-
a two-phase approach. The pseudohomogeneous ers. The next attempt considered residence time
approach, where the existence of more than one distribution (RTD) model in which all the gas in the
phase is not taken into account, proposes that we use bed is considered equal in terms of gas residence time.
the conventional multiphase flow models for the flui- However, since an operating fluidized bed consists of a
dized bed reactors. These conventional models may bubble phase and an emulsion phase of completely
include ideal flow models, dispersion models, residence different gas contacting hydrodynamics, the approach
time distribution models, and contact time distribution is inadequate and was also dropped.
models. The two-phase approach, however, considers Gilliland and Kundsen (1970) then modified the
the fluidized bed reactors to consist of at least two RTD models and assumed that the faster gas stayed
phases, a bubble and an emulsion, and proposes a mainly in the bubble phase and the slower in the emul-
separate governing equation for each phase with a sion. Their approach was to distinguish the effect of
term in each equation describing mass interchange the two classes of gas with different effective rate
between the two phases. Among the two-phase models, constant. The overall conversion equation in their
the bubbling bed model proposed by Kunii and proposed model is
Levenspiel (1969) and the bubble assemblage model ð1
CA
proposed by Kato and Wen (1969) have received the ¼ ekt E dt ð1Þ
CA0 0
most attention. Figure 1 illustrates the development
and evolution of these various flow models for flui- and the following equation was proposed to describe
dized bed reactors. the effective rate constant, i.e.,
3.3.2 Calculation Procedure Based on Bubble The distance from the distributor to the nth compart-
Assemblage Model ment is then
Let the height of the nth compartment be hn , where hn ¼ h1 þ h2 þ h3 þ
þ hn ð56Þ
n ¼ 1; 2; 3, to n (see Fig. 3). Based on an arithmetic The gas interchange coefficient based on unit volume
average of the bubble size, the height of the initial of bubble phase (bubble and cloud) can be shown as
compartment immediately above the distributor
becomes 0 ubr ðUmf ="mf Þ
Fon ¼ Fon ð57Þ
ubr þ 2ðUmf ="mf Þ
d þ ðch1 þ do Þ
h1 ¼ o ð46Þ
2 Hence, the material balance for the gaseous reactant
or around the nth compartment becomes, for the bubble
phase,
h1 ¼
2do
ð47Þ 0
2c ðSUCA;b Þn1 ¼ Fon Vb CA;b CA;e n
ð58Þ
where þ ðrA;c Vc Þn þ ðSUCA;b Þn
U and for the emulsion phase,
c ¼ 0:14rp dp ð48Þ
Umf 0
Fon Vb CA;b CA;e n ¼ rA;e Ve n ð59Þ
which is a proportionality constant relating the bubble
diameter for a given operating condition. The height of where rA;c and rA;e are the reaction rates per unit
the second compartment then becomes volume of the cloud and the emulsion phase, respec-
tively. Note that for the first-order reaction, the expres-
2þc sions for rA;c and rA;e are rA;c ¼ kr CA;b and
h2 ¼ 2do ð49Þ
ð2 cÞ2 rA;e ¼ kr CA;e , respectively.
3.3.3 Computational Procedure for Conversion sion phase for the nth compartment are then calculated.
The nth compartment concentrations, CA;bn and CA;en ,
The computational procedure for conversion and con-
are computed from (CA;b Þn1 and (CA;e Þn1 using Eqs.
centration profile in a fluidized bed reactor is given
(58) and (59). The calculations are repeated from the
below. The following operating conditions are needed,
distributor until the bed height equivalent to Lmf is
i.e., particle size (dp ), particle density (rp ), minimum
reached. For bed height above Lmf , the voidage is
fluidization velocity (Umf ), gas superficial velocity (U),
adjusted by Eq. (41), and Vcn , Vbn , and Ven are obtained
distributor arrangement (no ), column diameter (dt ),
using the same procedure as that shown for the height
incipient bed height (Lmf ), reaction rate constant (kr ),
smaller than Lmf . The calculation is repeated until the
and order of reaction. It should be noted that the
bed height reaches Lmf þ 2ðL Lmf ).
model requires no adjustable parameters.
First, Eq. (38) is used to calculate the expanded bed
3.3.4 Discussion of the Bubble Assemblage Model
height, L. Next, Eq. (50) is used to compute the size of
the nth compartment. Using Eqs. (51) through (55), the The computation using the bubble assemblage model
volumes of the cloud, the bubble phase, and the emul- indicates that for most of the experimental conditions
John C. Chen
Lehigh University, Bethlehem, Pennsylvania, U.S.A.
Figure 2 Nusselt numbers for particle–gas heat transfer in dense bubbling beds for Prg ¼ 0:7.
and horizontal tubes, since these are the geometries of Wender and Cooper’s correlation (1958) for vertical
most practical interest. tubes (SI units) is
The most common approach assigns thermal resis-
0:43
tance to a gaseous boundary layer at the heat transfer hc dp 4 0:23 cpg rg
surface. The enhancement of heat transfer is then ¼ 3:51 10 CR ð1 eÞRep
kg kg
attributed to the scouring action of solid particles on !0:66
0:8
the gas film, decreasing the effective film thickness. The cps rs
works of Leva et al. (1949, 1952), Dow and Jakob cpg rg
(1951), Levenspiel and Walton (1954), Vreedenberg
(1958), and Andeen and Glicksman (1976) utilized for 102 Rep 102
this approach. Correlations following this approach ð18Þ
generally present a heat transfer Nusselt number in
terms of the fluid Prandtl number and a modified where CR is a correction factor for tubes located away
Reynolds number, with either the particle diameter from the bed axis.
or the tube diameter as the characteristic length An empirical fit to the data of Vreedenberg (1952)
scale. Some useful examples of such correlations are gives the following function for CR :
given here.
Leva’s correlation (1952) for vertical surfaces, for 2
r r
larger particles, is CR ¼ 1:07 þ 3:04 3:29 ð19Þ
Rb Rb
hc dp
Nuc ¼ 0:525ðRep Þ0:75 ð17Þ where r is the radial position of the heat transfer tube
kg and Rb is the radius of the bed.
Figure 4 Normalized heat transfer coefficients for horizontal tube in freeboard of bubbling beds. H measured from level of
collapsed (packed) bed. (Data of Biyikli et al., 1983.)
! !
4.1.1 Fluidization Regime mg 1:85 105
Umf ¼ Remf ¼ 0:139
At the intended operating condition, would the bed r g dp 1:18ð125 106 Þ
fluidize in the bubbling regime?
¼ 0:0175 m=s
The answer is found by spotting the operating point
on the map of fluidization regimes plotted in Fig. 1. The terminal velocity is obtained from Eq. (6b),
Calculating the necessary parameters, we find the " #5=7
Archimedes number from Eq. (3): dp8=5 ðrs rg Þ
Ut ¼ 0:072g
grg ðrs rg Þ 3 r2=5
g mg
3=5
Ar ¼ dp " #5=7
m2g
ð125 106 Þ8=5 ð2480 1:18Þ
¼ 0:072 9:81
ð9:81Þð1:18Þð2480 1:18Þð125 106 Þ3 ð1:18Þ2=5 ð1:85 105 Þ3=5
¼ ¼ 164
ð1:85 105 Þ2
¼ 0:730 m=s
The Reynolds number at minimum fluidization is
and the dimensionless velocity is, by definition, from
found by solving Eq. (1):
Eq. (7),
ðRemf Þ2 ð1 emf ÞðRemf Þ Ug Umf 0:160 0:0175
Ar ¼ 1:75 þ 150 U ¼ ¼ 0:20
fs ðemf Þ3
ðfs Þ2 ðemf Þ3 Ut Umf 0:730 0:0175
ðRemf Þ2 ð1 0:42ÞðRemf Þ The dimensionless particle diameter is, from Eq. (4),
164 ¼ 1:75 þ 150
ð1:0Þð0:42Þ3 ð1:0Þ2 ð0:42Þ3
dp Ar1=3 ¼ 1641=3 ¼ 5:46
Remf ¼ 0:139
This locus is within the bubbling regime on the map of
from which we obtain the minimum fluidization Fig. 1, so the proposed operation will be a dense
velocity by Eq. (2), bubbling fluidized bed.
where where
ap = surface area of single particle ¼ pdp2 Dt rg Ug 0:029 1:18 0:16
np = number of particles per unit bed volume ¼ ReD ¼ ¼ 296
6ð1 eÞ=ðpdp3 Þ mg 1:85 105
ap np = 6ð1 eÞ=dp ¼ 6ð1 0:585Þ=ð125 106 Þ
¼ 1:99 104 particles/m3 . and the heat transfer coefficient is obtained as
which shows that particle convection (first term) which gives the value for the heat transfer coefficient
dominates over gas convection (second term).
The resulting heat transfer coefficient is then kg 0:0262
hc ¼ 0:0328 ¼ 0:0328 ¼ 486 W=m2 K
calculated as l 1:77 106
kg 0:0262
hc ¼ 2:27 ¼ 2:27 ¼ 476 W=m2 K 4.1.4 Heat Transfer Coefficient from Packet Model
dp 125 106
Use the packet model to estimate the convective heat
The correlation of Molerus et al. (1995) calls for transfer coefficient for the above case of a horizontal
Eq. (23). The necessary parameters are given by Eqs. tube in a bubbling bed. Additionally, the bed has a
(24)–(28): porous plate distributor, and the center line of the
heat transfer tube is located of Lt ¼ 0:19 m above the
Ue Ug Umf ¼ 0:160 0:0175 ¼ 0:143 m=s distributor.
!2=3 !2=3 To apply the packet model, we first need to calculate
1:85 105
1=3
mg 1
l¼ ¼ some bubble characteristics, utilizing hydrodynamic
rs rg g 2480 1:18 theory. The necessary equations can be obtained
1=3 from Chapter 3, or from the book of Kunii and
1
¼ 1:77 106 m Levenspiel (1991). First, to estimate the bubble dia-
9:81 meter (db ) at the elevation of the tube,
" 1=3 #1
Ue rs cps dbm dbo
B1 1 þ 33:3 Ue ¼ 1 þ 33:3 db ¼ dbm
Umf gkg e0:3Lt =Db
" 1=3 #1 where
0:143 2480 753
0:143 ¼ 1:60
0:0175 9:81 0:0262 dbo ¼ 2:78ðUg Umf Þ2
!1=2 2=3
rg rs cps g ¼ 2:78ð0:16 0:0175Þ2
B2 1 þ 0:28e2s;mf Ue Umf
rs rg gkg 9:81 ¼ 0:00575 m ¼ 0:575 cm
" #0:4
¼ 1 þ 0:28 0:582 0:143 0:0175 pD2b ðUg Umf Þ
1=2 dbm ¼ 0:65
1:18 2480 753 2=3 4
¼ 1:194
2480 1:18 9:81 0:0262 " #0:4
p 20:32 ð16 1:75Þ
U 0:0157 ¼ 0:65 ¼ 19:0 cm
B3 1 þ 0:05 mf ¼ 1 þ 0:05 ¼ 1:005 4
Ue 0:143
so that
and the Nusselt number is obtained from Eq. (23),
dbm dbo 19:0 0:575
hc l 0:125es;mf db ¼ dbm ¼ 19:0 0:3ð0:19=0:203Þ
¼
þ 0:165Pr1=3
g e 0:3Lt =Db
e
kg B1 1 þ B2 kg =2cps mg
¼ 5:09 cm ¼ 0:0509 m
!1=3
rg 1
The rise velocity of the bubble, relative to the gas/
rs rg B3
particle emulsion, is
rb 24 ¼ 122 K=m2 K
es;m ffi ¼ ¼ :0110
rs 2200 which is within 2% of experimental value.
Thomas C. Ho
Lamar University, Beaumont, Texas, U.S.A.
Sphericity
Type of particles s Source
(a) From geometric considerations. (b) From Leva (1959). (c) From
Uchida and Fujita (1934). (d) From Carmån (1937). (e) From Leva et
al. (1948, 1949). (f) From Brown et al. (1949).
Source: Kunii and Levenspiel, 1991.
r X Sa S dp Sd
Mesh size (g=cm3 ) (—) (cm2 =cm3 ) (cm2 =cm3 ) (cm) (—)
dp;m dt Lm
Investigators Gas Process Particles (mm) (cm) (cm)
by Krell et al. (1990). Typical results from these past An attempt was made to correlate the average mass
studies are summarized in the following subsections. transfer coefficient reported by Resnick and White
(1949) for small size fluidized particles, and the follow-
ing equations are obtained (for air system, Sc ¼ 2:35):
2.3.1 Average Mass Transfer Coefficient, kg;bed
1. For particles with size between mesh 14 and 20
The reported average mass transfer coefficients for
(dp ¼ 1000 mm),
fluidized bed particles are summarized in Fig. 2. As
indicated, the average mass transfer coefficient for flui- Shbed ¼ 0:200 Re0:937
p for 30 < Rep < 90
dized bed particles, again, can be higher or lower than ð21Þ
that estimated from Eq. (2) for well-dispersed single
spheres. The Sherwood number for fluidized bed par- 2. For particles with size between mesh 20 and 28
ticles is generally lower than that for single spheres at (dp ¼ 711 mm),
lower Reynolds numbers; however, it can be higher at
Shbed ¼ 0:274 Re0:709
p for 15 < Rep < 80
higher Reynold numbers (Rep > 80). Similar to that
for fixed bed particles, the Sherwood number for ð22Þ
fluidized particles can be well below the theoretical 3. For particles with size between mesh 28 and 35
minimum of the Sherwood number for single spheres, (dp ¼ 570 mm),
i.e., Shsingle ¼ 2:0. This is due to the potential overesti-
mation of active surfaces involved in the mass transfer Shbed ¼ 0:773 Re1:107
p for 8 < Rep < 60
operation. In a bubbling fluidized bed, most of the ð23Þ
particles are expected to stay in the emulsion phase
with CA i
at equilibrium with CA in the emulsion gas. 4. For particles with size between mesh 35 and 48
Many of these particles may thus be considered as (dp ¼ 410 mmÞ;
inerts from the mass-transfer point of view because Shbed ¼ 0:071 Re0:926
p for 6 < Rep < 40
they do not contribute to significant amount of mass
ð24Þ
transfer to the bubbling gas. The fact that all the
particle surfaces are included in the mass transfer cal- 5. For particles with size between mesh 48 and 65
culation as described by Eq. (7), therefore, overesti- (dp ¼ 275 mm),
mates Sex;particles , which in turn decreases the average
Shbed ¼ 0:041 Re1:036
p for 4 < Rep < 15
mass transfer coefficient to a value even below the
theoretical minimum for single spheres. ð25Þ
2.3.2 Total Particle Exterior Surfaces, Sex;particles 2.4 Transfer Between Immersed Isolated Spheres and
a Fluidized Bed
The procedure for determining the total particle exter-
ior surfaces in a bed segment of a fluidized bed is Another type of mass transfer encountered in a flui-
essentially identical to that in a fixed bed, as described dized bed involves the transfer between an immersed
previously, i.e., using Eqs. (10) through (15). The ebed active particle and a fluidized bed. Fluidized bed com-
appearing in Eqs. (10) and (12), however, represents bustors are a typical example where active coal parti-
bed voidage in a fluidized bed (or ef ) in this applica- cles burn in the matrix of chemically inert particles of
tion. sand and ash which constitute 97 to 99% of the whole
bed mass. The mass transfer between the immersed
active particles and the fluidizing gas is expected to
2.3.3 Comparison of Kg;bed between Fixed Bed and
be different from that in a bed consisting of only active
Fluidized Bed Particles
particles, because the inert particles may serve as an
The experimental study of Resnick and White (1949) impermeable barrier for mass flow.
included the interesting observation that, although the The subject of mass transfer involving isolated
bed hydrodynamics is significantly different, the active particles and the fluidizing gas has been studied
measured mass transfer coefficients for the fixed and by several groups of researchers. Hsiung and Thodos
fluidized beds can be correlated with Rep by the same (1977) measured the mass transfer coefficients of a
set of constants as shown in Fig. 3. In the figure, the batch of naphthalene particles fluidized in a matrix
dashed lines represent fixed bed operations and the of inert particles similar in size and density.
solid lines represent fluidized bed operations. As Vandershuren and Delvossalle (1980) experimentally
indicated, the Sherwood number (or mass transfer studied the drying process of moist particles in a flui-
coefficient) for the same group of particles continues dized bed of similar dry particles. Palchonok and
to increase with the Rep , even during the transition of Tamarin (1983, 1985) measured the mass transfer coef-
the operation from a fixed bed to a fluidized bed. It ficients for an isolated active particle of various sizes
appears that Eqs. (21) through (25), although derived (dp;a ¼ 3:5–30 mm) freely moving within a fluidized
for fluidized bed operations, can be extrapolated to bed of large inert particles (dp;i ¼ 1:2–6.3 mm). Kok
fixed bed operations as well. et al. (1986) measured the particle–gas mass transfer
dNA i
¼ kg;bed Sex;particles ðCA CA Þ ð7Þ
dt
Figure 7 Nomenclature used in developing the model of Kunii and Levenspiel. (From Kunii and Levenspiel, 1991.)
A comparison between Eqs. (44) and (30) yields Inserting Eq. (48) into Eq. (47) yields
yfs dp2 d
kg;bed ð1 ef Þa 0 Shbed ¼ K ð53Þ
KGB ¼ ð45Þ 6Dð1 ef Þ bc
d
The above equation provides a correlation for estimat-
Note that Eq. (45) implies that ing Shbed (or kg;bed ) based on particle and fluidization
properties, which then can be used to determine the
d
kg;bed ¼ K ð46Þ rate of mass transfer employing Eq. (7) based on the
ð1 ef Þa 0 GB
homogeneous bed approach. In the application, the
Combining with Eq. (11) the equation yields following equations provided by Kunii and
Levenspiel (1991) are needed for calculating fluidiza-
kg;bed dp y d= ð1 ef Þa 0 KGB dp y tion properties appearing in Eq. (53)
Shbed ¼ ¼
D D U Umf
ð47Þ d¼ ð54Þ
yfs dp2 d ub
¼ KGB
6Dð1 ef Þ ub ¼ ðU Umf Þ þ ubr ð55Þ
0:5
It is worth mentioning that, according to Kunii and ubr ¼ 0:711ðg db Þ ð56Þ
Levenspiel (1991), KGB represents the transfer of A 1 ef ¼ ð1 emf Þð1 dÞ ð57Þ
from the particle phase to the bubble phase via two
sources, i.e., transfer from particles dispersed in the The above equations indicate that, besides gas and
bubble phase and transfer of gas across the bubble– particle properties, the fluidization properties are a
cloud boundary. These are discussed below. function of bubble size only.
yfs dp2 d 6 Shsingle D where CAs is the concentration of tracer A within the
Shbed ¼ g þ Kbc particle in equilibrium with the concentration CA i
of
6Dð1 ef Þ b fs dp2 y
( ) ð61Þ tracer gas at the gas–particle interphase. Note that
d fs dp2 y for nonporous and nonadsorbing particles, m ¼ 0
¼ g Sh þ K and Zd ¼ 0, and Eq. (63) reduces to Eq. (53). For
ð1 ef Þ b single 6D bc
highly adsorbing particles, m is on the order of thou-
Like Eq. (53) this equation provides a correlation for sands, in which case Zd ! 1:0 and Eq. (63) becomes
estimating Shbed (or kg;bed ) based on particle and Eq. (61). For porous but nonadsorbing particles,
fluidization properties, which then can be used to m ¼ ep , and Eq. (63) can be used to estimate Shbed.
determine the rate of mass transfer employing Eq. In demonstrating the effect of the particle adsorbing
(7). Eqs. (54) through (57) are also needed for calculat- property on the mass interchange coefficient,
ing fluidization properties appearing in Eq. (61). Wakabayashi and Kunii (1971) reported that the
Again, all these equations indicate that, besides gas KGB can be greatly enhanced even with a small value
and particle properties, the fluidization properties are of (gb Zd ). Their reported values are shown in Table 4.
a function of bubble size only. Note that, for a given Rietema and Hoebink (1975) also reported that their
bed of solids and constant bubble size, Eq. (61) reduces measured gas interchange coefficients increase in bub-
to the form (Kunii and Levenspiel, 1991): ble size and commented that this was just the reverse of
the prediction of Eq. (49) for Kbc . According to Kunii
Shbed ¼ aRep þ b ð62Þ
and Levenspiel (1991), this is because the first term of
The equation indicates that Shbed is linearly related to Eq. (63) dominates for large bubbles, caused by more
the particle Reynolds number with a slope a and an vigorous splitting and coalescing of bubbles.
intercept b. This conclusion is similar to but not in
complete agreement with that generated empirically 3.2 Model of Partridge and Rowe (1966)
based on the curve fitting of experimental results,
where an exponential relationship was derived as In this class of models, the cloud surrounding the bub-
expressed in Eqs. (21) through (25). It may be that ble is considered the primary mass transfer boundary,
the bubbles do not have a constant size over a wide and the bubble and the cloud phases are considered a
range of Rep . However, for a short range of Rep , this perfectly mixed single phase, as indicated in Fig. 6(b).
conclusion seems reasonable. The mass transfer equation then has the form
kgt Calculated
Model Reference Coefficient equivalent kgt (m/s)
5. Kunnii and Levenspiel Kunii and Levenspiel kgb ¼ 0:6D1=2 ðg=db Þ1=4 kgcQ 0.0051
(1969) kgc Sb þ Q
q ¼ 2Umf db w
Q ¼ q þ kgb Sb
4Demf ub
kgc ¼
pdc
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
rffiffiffiffiffiffiffiffiffiffiffi
6. Toei et al. Toei et al. (1969) kgc þ kgs 0.0056
1:02emf ub Dð 1Þ a 1
kgc ¼
1 þ 2emf =ða 1Þ db a aþ1
7. Murray through-flow, Murray (1965) kgb ¼ 0 q=Sb 0.0160
no diffusion q ¼ Umf db w
Empirical value (from Charvarie and Grace, 1976a) 0.0160
and for a spherical three-dimensional bubble, it has the dNA dCA;b
¼ Vbubble ¼ KGT Vbubble
form dt dt ð77Þ
ðCA;e CA;b Þ
Umf 4Demf ub 0:5
kgt ¼ þ ð75Þ
4 pdb Note that this equation is essentially in the same form
as that of Eq. (30) proposed by Kunii and Levenspiel
The above equation can be converted to KGT , the mass
(1991), i.e., KGT reported by Sit and Grace (1978) is
interchange coefficient per unit volume of bubble,
directly comparable to KGB described previously by
based on the appropriate surface-to-volume ratio, i.e.,
Kunii and Levenspiel (1991).
6 In a similar attempt, Hatano and Ishida (1986) stu-
KGT ¼ k died the particle–gas mass transfer coefficient in a
db gt
ð76Þ three-dimensional fluidized bed of nonadsorbing glass
1:5Umf 12 Demf ub 0:5 beads with dp ¼ 0:18 mm using optical fiber probes.
¼ þ 1:5
db db pdb Tracer gas concentrations in and around rising single
bubbles was measured continuously by the penetrative
The corresponding mass transfer equation for KGT probes, while the bubble boundary and the zone with
appearing in Eq. (76) is the prominent particle movement were detected by
As pointed out by Kunii and Levenspiel (1991), the 4 RELATION BETWEEN MASS AND HEAT
mass transfer rate in a fluidized bed can be substan- TRANSFER COEFFICIENTS
tially enhanced for highly adsorbing particles even
though the amount of particles dispersed in the bubble The correlation equations for particle–gas mass and
phase may be small. Similar observations have also heat transfer coefficients are closely related when the
been reported in the literature, see, e.g., Yoshida and Sherwood number (Sh) is equivalent to the Nusselt
Kunii (1968), Drinkenburg and Rietema (1973), Yates number (Nu), and the Schmidt number (Sc) is equiva-
and Constans (1973), Chiba and Kobayashi (1970), lent to the Prandtl number (Pr). All the particle–gas
Nguyen and Porter (1976), and Gupalo et al. (1978). correlation equations for the Sherwood number and
Among these investigators, Chiba and Kobayashi the Nusselt number are therefore interchangable. For
(1970) derived a theoretical expression for the ratio example, the correlation equation for estimating the
of interphase mass transfer coefficients in the presence Nusselt number for particles in a fixed bed is expressed
(Kbe0 ) and absence (Kbe ) of adsorption effects: as
0:5
Kbe0 1 emf a0 hdp
¼ 1 þ 0:67m 1:5 þ 0 ð85Þ Nubed ¼ ¼ 2 þ 1:8ðRep Þ0:5 ðPrÞ0:333
Kbe emf a þ1 kg ð87Þ
for Rep > 80
where m is the adsorption equilibrium constant defined
in Eq. (66) and a 0 is defined as
whereas the Prandtl number (Pr) is defined as
ub emf
a0 ¼ ð86Þ Cp m
Umf Pr ¼ ð88Þ
kg
Equation (85) was later confirmed experimentally by
Bohle and van Swaaij (1978), who measured mass Note that the corresponding equation for the
transfer coefficients for a number of adsorbing (e.g., Sherwood number has the form
propane) and nonadsorbing (e.g., helium) gases in a kg;bed dp y
fluidized bed of silica-alumina. A typical comparison Shbed ¼ ¼ 2 þ 1:8ðRep Þ0:5 ðScÞ0:333
D ð89Þ
is shown in Fig. 8, where the effect of adsorption is
obvious with the enhancement being as high as 100%. for Rep > 80
Determine. The relation between Shbed and Rep Combining (A) through (D) yields
based on Eq. (61) with db ¼ 0:1, 0.2, 0.37, 0.5, and
d fs dp2 y
1 cm. Shbed ¼ gb Shsingle þ K
1 ef 6D bc
Solution. Equation (61) has the form
! U 1:21
¼ ½ð0:005Þð4:84Þ þ 0:000804ð43:63Þ
d f dp2 y 6:77
Shbed ¼ gb Shsingle þ s K
1 ef 6D bc ¼ 0:0088U 0:011 ðEÞ
rg Udp ð0:028ÞUð0:00118Þ
Rep ¼ ¼ ¼ 0:184U ðFÞ
m 0:00018
NOMENCLATURE
Metallurgical and
Physical operations Chemical syntheses mineral processes Other
the fluidized bed with vertical baffles that reduce the prolonging particle on-stream time. Reactors in series
bubble size, yielding more uniform flow structures in and parallel are common arrangements for fluidized
the fluidized bed. For drying processes involving sticky bed operations (Figs. 2e and f). These systems can be
particles or particles with high moisture contents, shal- used in conjunction with multistep reaction processes.
low bed operation is essential to smooth fluidization. The crosscurrent multicell type of fluidized bed (Fig.
Shallow beds are also operated with small bubble sizes, 2f) allows recovery of rapidly released gas products
providing efficient gas–solid contact. Figures 2b and c from solid particles where the particle residence time
show the configurations of multistage shallow bed is to be long. A fluidized bed can be also operated in
operations. For multistage drying operation, particles the presence of an external force field other than the
are introduced from the top stage and dried through gravitational field, such as centrifugal and magnetic
multistage fluidized beds while hot gas enters at the fields. The configuration shown in Fig. 2h is a typical
bottom stage. This countercurrent gas–solid contact centrifugal fluidized bed. For fast reactions and pro-
improves the efficiency of the fluidized bed operation. cesses requiring on-line catalyst regeneration, the cir-
Particle exchange between fluidized beds or between culating fluidized bed is usually used as shown in Figs.
regions in a fluidized bed can be accomplished through 2g and i. The downer reactor offers advantages for
external or internal solids circulation, as shown in Figs. ultrafast reactions and processes with good product
2d and e. The external or internal solids circulation selectivity. Through gas spouting, spouted beds (Fig.
allows the coupled reaction and regeneration scheme 2j) offer means to fluidize large particles (Group D)
to be implemented for the solid particles, thereby with well-structured internal particle circulation.
Figure 3 Schematic diagram of fluidized bed reactor developments. (From Krishna, 1994; Jazayeri, 1995.)
These approaches are discussed in conjunction with the Contact schemes of gas–solid systems in fluidized beds
flow regime behavior, which is given in the following are classified by the state of gas and solids motion. For
section. noncirculating systems, the gas at low velocity merely
The particle size distribution (PSD) significantly percolates through the voids between packed particles,
impacts the reactant conversion in a fluidized bed reac- while the particles remain motionless in the fixed bed
tor. Sun and Grace (1990) examined the three different state. With an increase in gas velocity, particles move
particle size distributions, wide, narrow, and bimodal, apart and become suspended; the bed has entered the
on the performance of a catalytic fluidized bed reactor fluidization state. Further increase of gas velocity sub-
using the ozone decomposition reaction. They found jects the flow to a series of transitions from a bubbling
that a fluidized bed with particles of wide size distribu- fluidization regime at low velocities to a dilute trans-
tion yields the highest reactant conversion. Of further port regime at high velocities accompanied by signifi-
interest, the property of particle entrainment and elu- cant variations in gas–solid contact behavior. Some
triation is a function of particle size, density, and gas–solid flow structures associated with given flow
shape. Both entrainment and elutriation rates increase regimes are presented in Fig. 5.
Figure 5 Flow regime diagram of fluidized bed reactors. (a) Fixed bed, (b) bubbling fluidized bed, (c) turbulent fluidized bed, (d)
circulating fluidized bed, and (e) dilute transport bed.
Figure 7 Generalized map of fluidized regimes. (From Kunii and Levenspiel, 1997.)
1=2
Umf 4Dm emf Ub accumulation in the wall region can be analyzed
kq ¼ þ ð5Þ using the core–annular model, and the mass balance
3 pdb
of solid particles can be expressed as
The effectiveness of gas–solid contact is reflected in the p 2 p p
reactant conversions. Figure 10 shows the extents of D Gs ¼ ðD 2dw Þ2 Vpc ð1 e Þc rp þ
4 4 4 ð6Þ
reactant conversions in a bubbling bed reactor for
various reactions as a function of dimensionless rate ð4Ddw 4d2w ÞVpw ð1 e w Þrp
coefficients along with model predictions based on
The cross-sectional averaged voidage is
both the plug flow and the CSTR flow patterns for
bubbling bed reactors (Kunii and Levenspiel, 1991). p 2 p p
D ð1 e Þ ¼ ðD 2dw Þ2 ð1 e c Þ þ
The conversions are typically lower than those pre- 4 4 4 ð7Þ
dicted by either model, and this is an indication of ð4Ddw 4d2w Þð1 e w Þ
inefficient gas–solid contact.
In the turbulent regime, interactions between gas Upon rearrangement of Eqs. (6) and (7), the ratio of
voids and the dense phase are strong, yielding extensive solid particles in the wall region to the overall particle
global and local solids mixing, thereby enhancing gas– content can be expressed as
solid contact. In high-velocity operations of a fluidized
ð1 e w Þ4hð1 hÞ ð1 2hÞ2 ð1 e c Þ
bed reactor, contact resistance of the bulk gas with ¼1
particles in clusters or wall layers is present. Particle ð1 e Þ ð1 e Þ
accumulation in the wall region also imposes a hin- Vpc Gs =½rp ð1 e Þ
¼ ð8Þ
drance on gas–solid contact owing to low gas flow Vpc Vpw
near the wall. Since significant amounts of particles
dw
are in the wall layer, this layer plays an important h¼
role in mass and heat transfer. The extent of solid D
Using the correlation of radial voidage profiles devel-
oped by Zhang et al. (1991),
2:5 11 r
eðrÞ ¼ e ð0:191þf þ3f Þ f¼ ð9Þ
R
A ratio of 0.51 is obtained for typical conditions
(e ¼ 0:88 and h ¼ 0:05).
Like Fig. 10 for the bubbling bed reactor, the effects
of gas–solid contact in a riser reactor can be described
in terms of the extent of reactant conversion for var-
Figure 9 Gas transfer pathway in the bubbling regime. ious reactions along with predictions as given in Fig 11
transfer is generally higher in a bubbling bed than that been suggested as a means for improving the gas–
in a CFB. solid contact. The flow in a dilute transport reactor is
close to plug flow. At high operating gas velocities the
solids holdup is very low. This flow regime is applicable
2.2.6 Flow Regime Selection
for fast reactions with short residence time require-
Each flow regime has its unique flow characteristics, ments. As noted, fluidized bed reactors are often used
which are illustrated in Table 2 using Group A particles due to nearly isothermal operation and intensive mix-
as an example. Knowing these flow characteristics can ing. Clearly, this is the case for coal combustion and
aid in the selection of a flow regime for a given reaction. catalyst regeneration in FCC systems. Cracking cata-
The solids holdup in a packed state is typically in the lysts deactivate rapidly by coking, and the regeneration
range of 0.45–0.65 and decreases with increasing gas process is dominated by highly exothermic coke com-
velocity. Typical axial profiles of solids holdup for bustion. The residence time for catalyst regeneration
Groups A and B particles are given in Fig. 15, while determines the regenerator size. Operation in the tur-
radial profiles for the fast fluidization and dilute trans- bulent regime ensures efficient contact between oxygen
port regimes are given in Fig. 16. The turbulent flui- and catalysts while avoiding the formation of hot spots
dized bed regime provides improved gas–solid contact and preventing catalysts deactivation by sintering and
efficiency with heat transfer and gas backmixing com- steaming in catalyst regeneration.
parable to or slightly higher than those in the bubbling For selective reactions requiring isothermal opera-
bed regime while maintaining a relatively high solids tion with desired conversion sensitive to backmixing,
concentration. This unique regime has been applied to such as butane oxidation, partial oxidation of natural
many process operations such as the MTG process gas, and acrylonitrile production, a compromise
developed by Mobil, butane oxidation, and acryloni- between isothermal operation and the plug–flow
trile production. When the gas velocity to the fast flui- requirement must be made in regime selection. The
dization regime is increased, the core–annular flow turbulent regime can accommodate this compromise.
structure occurs. The relative dimension of the annular However, regime selection also requires consideration
region to the core region is not proportional to the size of catalyst characteristics and system configurations.
of the riser, which makes riser scale-up a challenging Oxidation of hydrocarbons is routinely carried out in
issue. To minimize the segregated flow between the core fixed-bed bundle reactors and fluidized bed reactors
and wall regions, baffles and other internal structures with simultaneous presence of oxygen and hydrocar-
(Jiang et al., 1991; Ran et al., 1999) can be used. bon at the active sites of the catalyst. Sze and Gelbein
Manipulations of particle size distribution have also (1976) proposed a reactor–regenerator process for oxi-
Flow nature Bubbles and Dispersed dilute Core-annular flow, Dispersed dilute flow
emulsion phase phase and dense particle accumulation with a thin particle
phase near wall and particle layer on the wall
clustering
Solids holdup 0.5, decrease with 0.3–0.5, decrease Bottom dense region: < 0:05, decrease with
gas velocity with gas velocity 0.05–0.4, gas velocity and increase
Upper dilute region: with solids flow rate
< 0:05, decrease with
gas velocity and increase
with solids flow rate
Axial profiles of Fig. 16 Fig. 16 Fig. 16 Fig. 16
solids holdup
Radial profiles Uniform Slightly dense in Large variations Uniform except in the
of solids fraction wall region thin wall layer
Gas backmixing High Pea ¼ 0:1–0.4 Pea ¼ 5–8, close to plug Plug flow
(dispersion) flow in central zone;
strong backmixing near
the wall
Radial gas mixing Per ¼ 2–20 Per ¼ 2–20 Per ¼ 100–1000 Per ¼ 100–200
Axial solids mixing Bubble agitation Very well Poor, Pea ¼ 5–10 Poor
Radial solids mixing Well Very well Core–annular structure Poor
poor, Pea ¼ 100–1000
Gas–solid contact Poor Well Poor in wall region Well
Temperature profile Uniform May have slight Slight axial gradient Large axial gradient
axial gradient
Heat transfer Very high Very high Low Poor
Solids flow Entrainment High entrainment Solids circulation Solids transport
Gas velocity Low Higher Highest Highest
(throughput)
Residence time Long/wide Long/wide Short/relatively wide Short/narrow
distribution
dative synthesis of aromatic nitriles. The same type of 17. The averaged activity of a fluidized bed can be
process with a circulating fluidized bed reactor is used maintained even though each catalyst particle under-
for selective oxidation of n-butane to maleic anhydride goes rapid deactivation–regeneration. However, this
(Contractor and Sleight, 1987). In this process, n- process does not apply to situations in which the pro-
butane is oxidized in the riser by the oxygen-loaded duct and selectivity are sensitive to catalyst activity, as
catalyst, which in turn is reoxidized in a separate deactivated catalysts may lead to undesirable products.
regenerator operated in the bubbling bed regime. The Advances in catalyst development have led to a shift
separated reaction–regeneration configuration allows in industrial focus from the case of low-throughput
for optimum flow regime to be employed for each of bubbling bed reactors with low catalyst activities to
the reactors associated with different types of reaction. high-throughput turbulent bed reactors or risers with
Catalyst regeneration also provides a means of high catalyst activities. Catalytic cracking of petroleum
maintaining overall catalyst activity within systems feedstock to light oils, gasoline, and solvents can be
with rapid catalyst deactivation. Deactivated catalyst carried out in bubbling beds or riser reactors. In bub-
particles can be continuously removed from a reaction bling beds, catalyst particles are well mixed and hot
zone while regenerated catalyst from a regeneration spot formation is avoided. With the development of
zone is continuously recycled back to the reaction new zeolite catalysts with high activities in the 1960s,
zone in a flow scheme similar to that shown in Fig. riser reactors were subsequently implemented and pro-
merization to form a broad size distribution of poly- reactor wall. Often, the production rate of the gas-
mer particles. In polymerization, reactant diffuses from phase polymerization in a fluidized bed reactor is lim-
the gas phase, through the boundary layer surrounding ited by the maximum heat removal rate. The preferred
the particles, and into the catalyst pores to the active methods for removing heat from a fluidized bed reactor
sites where polymerization through a coordination- involving gas-phase polymerization are
insertion mechanism occurs to form a solid polymer. 1. Recycling unreacted gas through a heat exchan-
The catalyst particles are composed of small metal ger before returning the gas to the fluidized bed aids in
fragments and explode into a large number of smaller cooling the reactor. The heat removal rate depends on
fragments during the reaction. In the early stage of the external heat exchanger capacity and gas flow rate.
polymerization, these smaller particles are quickly In order to maintain this cooling sink, monomer con-
encapsulated by polymer molecules and grow to large sumed in the polymerization reaction is replaced by
polymer particles. For polyethylene, the reaction adding excess monomer gas to the recycle stream.
begins with small particles (20–200 mm in diameter) 2. Liquid, with a boiling point lower than the reac-
and finishes, 3–5 hours later, with a mean particle dia- tion temperature, is intermittently injected into the bed
meter of 200–2000 mm. From gas-phase monomer to with a spray pattern and immediately evaporates,
solid-state polymer, ethylene experiences a dramatic thereby removing heat. The injected liquid can be
physicochemical transition. Similarly, the particle monomer to be polymerized or inert liquid hydro-
morphology undergoes a dramatic change, and the carbon.
developmental stages are shown in Fig. 20. 3. Fluidized bed reactors are connected in series as
Implementation of fluidized beds for gas-phase poly- a staging operation. The fluidizing gas from the prior
merization also provides a typical example in which reactor is cooled off before entering the next fluidized
particles grow in size during a short reaction time. bed.
Key factors to fluidized bed polymerization processes 4. Part of the gaseous stream leaving the top of the
are localized/overall temperature control and handling reactor may be condensed by means of an external heat
of sticky particles due to process conditions operated exchanger and then reinjected into the fluidized bed in
near the resin melting temperature. Without proper liquid form. The boiling point of condensable liquid
cooling, the reactor temperature would increase until has to be lower than the operating temperature of
the catalysts become inactive or particles fuse to the polymerization. Thus the reaction heat can be removed
Figure 20 Morphology developments and particle growth of polyethylene particles during gas-phase polymerization. (From Xie
et al., 1994).
Figure 21 Particle flow patterns under various liquid contents in a fluidized bed reactor. (a) Dry particle fluidization, (b) particle
agglomeration, (c) agglomerate segregation, (d) bottom channeling, (e) whole bed channeling, (f) paste or slurry bed.
Figure 23 Effects of moisture content on the bed expansion As indicated in Fig. 4, the hydrodynamics and reaction
in a fluidized bed. kinetic models, both phenomenological modeling
Ud mg rp dp g This chapter
Ud2 þ 85:7eds Ud 0:571ð1 eds Þ3 ¼0
dp rg rg
!0:132
e Uþ1 0:055 rp rg dp 0:02 Jiang et al. (1999)
e ¼ 3:0 Ar expð0:157F45 Þ
Uþ2 rg D
3
eds U This chapter
ð1 eds Þ ¼ emb þ 0:02Ar0:13
Uc
0:3
ebs U This chapter
ebs ¼ 0:07 expð0:25F45 Þ
Ut
efs efs es ¼ ðebf es Þ exp½aðz Hf Þ ebf ¼ ð1 eÞ Kunii and Levenspiel (1997)
ðH
Wtf Wtf 1 exp½aðH Hf Þ This chapter
¼ efs dz ¼ es ðH Hf Þ þ ðebf es Þ
Arp Hf a
the wall and a central dilute region. In the mesoscale, Various core-annular models have been developed
the flow is characterized by the presence of particle to describe the gas–solid flow (Horio et al., 1988; Bai et
clusters moving upward in the core and moving down- al., 1995; Bolton and Davidson, 1998). The main
ward near the wall. Phenomenological models attempt difference among these models lie in the degree of
to capture these main flow features and express the complexity and in the assumptions associated with
flow behavior in simplified mathematical form to simplifications. Core-annular flow structures become
describable levels. Cluster concepts evolved from the dominant in the upper dilute region. Thus, when the
observations of large slip velocity between gas and dilute flow is predominantly present in the riser, mod-
particles. This type of model describes the motion of els based on core-annular structures can be applied to
clusters and can be combined with reaction models to reactor models. Kunii and Levenspiel (1990) extended
predict reactor performance when cluster properties the conventional fluidized bed model (a dense lower
are known. region coupled with a freeboard upper region) to cir-
Figure 27 Model predictions of concentration in a turbulent Noting that eVgz ¼ U, the boundary conditions for
fluidized bed. (From Jiang and Fan, 1999.) Eq. (34) can be written as
4 CONCLUSION
Figure 29 Axial profiles of concentrations for propylene, Fluidized bed reactor systems have been proved useful,
oxygen, and carbon dioxide in the process of ammoxidation reliable, and cost-effective, suitable to many industrial
of propylene. (From Wei et al., 2000a.) applications. The general approaches for designing a
fluidized bed reactor are described in this chapter with
specific attention given to the knowledge required for
from 0.02 to 10, in which the yield of acrylonitrile the selection of flow regimes and to the interplay
increases significantly with an increase in Pea . between transport properties and kinetics. Examples
Figure 31 shows the model analysis of the effects of of several industrial fluidized bed reactor processes
radial gas dispersion coefficient on radial profiles of are given to illustrate the operational principles. No
propylene concentration. The radial mass transfer single universal approach for designing a fluidized
has a significant effect on the conversion and yield. bed reactor system exists. At present, fluidized bed
When the radial Peclet number decreases from 1400 reactor system design relies heavily on correlations,
to 200, the conversion of propylene increases by over engineering models, and plant observations. The com-
10%, and the yield of acrylonitrile increases by about putational fluid dynamic approach may offer viable
7%. Since the reaction is first order with respect to and attractive options over the traditional approaches;
propylene, risers are operated under dilute conditions however, challenges remain in considering reactive
at Per ¼ 200, so the radial concentration distribution
of propylene is uniform and radial mass transfer is not
Leon R. Glicksman
Massachusetts Institute of Technology, Cambridge, Massachusetts, U.S.A.
Figure 1 Apparent reaction-rate constant vs. reactor diameter and bed height. (From Frye et al., 1958.)
Figure 2 Solid diffusivity in axial direction for large units. (From May, 1959.)
The differences in behavior between small labora- height. At the same superficial gas velocity and solid
tory beds and larger demonstration units can in part recirculation rate, the larger diameter column had a
be attributed to a switch from porous plate distributors higher solids fraction. The average slip velocities
in the small bed to discrete hole or bubble caps in the derived from these data are also higher for the large
larger beds. The porous plates give a better quality of diameter riser; see Hartge et al. (1985). Yerushalmi
fluidization, e.g., smaller bubbles, for shallow beds and and Avidan (1985) found a similar trend when com-
beds of moderate depth; see Rowe and Stapleton paring a 15.2 and 7.6 cm column. Noymer et al.
(1961). (1995) also compared two columns of 5.08 and 7.68
Yerushalmi and Avidan (1985) suggest that the cm diameter of the same height, which were used to
axial dispersion coefficient of solids in slugging and simulate larger pressurized fluidized bed combustors.
turbulent flow varies approximately linearly with the They found higher solids loading for the larger dia-
bed diameter, as with Thiel and Potter (1978). meter riser at equal gas velocity and solid recircula-
tion. In addition, the fraction of the wall covered by
3.3 Influence of Bed Diameter on Circulating or Fast clusters was higher for the larger diameter column
Fluidized Beds when the two beds had equal solids flow and when
the two beds had equal cross section averaged solids
Arena et al. (1988) measured the hydrodynamic beha- concentration. The increased wall coverage should
vior of two circulating fluidized beds with riser dia- lead to higher heat transfer rates from the bed to
meters of 0.041 and 0.12 m ID of roughly the same the wall.
Rhodes et al. (1992) compared the solids flux pro- to the turbulent regime. The results are for the center
files across the cross sections of a 0.152 and a 0.305 m of the riser excluding the entrance and exit regions. It
diameter circulating bed riser. They found a region would be interesting to determine if the similar voidage
where the solid profile, given by the ratio of local profiles hold for larger diameter risers.
flux to average flux, had a similar variation over the The thickness of the downflowing layers at the wall
cross section and was insensitive to the level of solid of the CFB is typically defined as the distance from the
flux. The variation of the local solids flux over the wall to the position of zero average vertical solid flux.
radius was a function of the gas velocity and the riser Zhang et al. (1995) made measurements of the layer
diameter. In the larger riser, the profiles were some- thickness on a 12 MWth and a 165 MW CFB boiler.
what flatter and the thickness of the downflowing They found that the thickness increased for the larger
region relative to the bed radius was smaller. The com- bed. They related data from many different beds, Fig.
parisons were not exact since the cross sections com- 11, with the equivalent bed diameter, taken as the
pared for the two beds were at different heights. hydraulic diameter, using the form
Zhang et al. (1991) carried out investigations with
¼ 0:05De0:74 ð4Þ
three different fast bed systems with diameters of 32,
90, and 300 mm. They found that the radial voidage The thickness,
, was found to be insensitive to particle
distribution, given as a ratio to the cross-sectional concentration, gas velocity, and height within the fur-
average, was independent of bed diameter and solids nace. This suggests that the thickness results from a
recycle rate. The similarity does not hold at transition balance of solids internal circulation that is generally
Figure 10 Mixing coefficients for different vessel diameters; Hovmand and Davidson (1971) review Stewart’s cri-
ML , longitudinal dispersion of fluid; MB , gas backmixing; terion for the transition from bubbling to slug flow,
MS : longitudinal solid dispersion. (From Van Deemter, u umf
1980.) ¼ 0:2 ð5Þ
0:35ðgDÞ1=2
much higher than net throughflow. If the local solids and show it gives good agreement with most experi-
flux profile, as a ratio of the cross-sectional average, is ments. When u umf is larger than the value found
roughly invariant over the cross section for these larger from Eq. (5), the bed will be in slug bed. Thus a small
Figure 11 Empirical correlation and experimental data of thickness of downflowing layer at the wall of a CFB as a function of
the equivalent bed diameter. (From Zhang et al., 1995.)
Since large pilot plants are costly, it may not be feasible 4.1 Development of Scaling Parameters
to undertake an extensive scale-up program.
Furthermore, it is difficult to observe or measure There are several approaches to developing the correct
detailed behavior within large hot beds. Thus many scaling relationships. Probably the most straightfor-
test results from beds at elevated temperatures and ward is the nondimensionalization of the governing
pressures are confined to overall operating characteris-
tics, leaving the designers to speculate on the cause of
shifts in performance with bed size. Full-scale cold
models are also expensive and require lengthy construc-
tion to modify the bed geometry. As we will see below,
the full-scale cold test may not correctly simulate the
hydrodynamics of the actual process at elevated tem-
perature and pressure. Indeed, a familiar occurrence in
atmospheric bubbling fluidized bed combustors is the
marked difference in flow behavior between a bed flui-
dized with cold air and the same bed, using the same
particles, run at normal operating conditions.
A technique, that can assist in the scale-up of com-
mercial plant designs is the use of scale models. A scale
model is an experimental model that is smaller than the
hot commercial bed but that has identical hydrody-
namic behavior. Usually the scale model is fluidized
with air at ambient conditions and requires particles
of a different size and density than those used in the
commercial bed. The scale model relies on the theory
of similitude, sometimes through the use of
Buckingham’s pi theorem, to design a model that Figure 12 Use of scale models with different bed diameters
gives identical hydrodynamic behavior to the commer- to simulate the influence of diameter on the hydrodynamics
cial bed. Such a method is used in the wind tunnel of a hot commercial reactor.
stant. These calculations were carried out over a should be valid for particle diameters of about 0.2 mm
range of particle Reynolds numbers, RepE , based on or less for a pressurized bubbling bed with u0 /umf of 10
the full scaling law, or exact, model. In the simplified and uslip =U0 of 0.3. At u0 /umf of 1000 and uslip =U0 of
scaling relationships, the Reynolds number is not one-fiftieth, the one-sixteenth scale model should be
maintained constant. The concern is how much the valid for pressurized beds with particles up to 1 mm
drag coefficient is impacted by the shift in Reynolds in diameter. These conclusions apply when the parti-
number. It was found that the particle Reynolds num- cle-to-fluid drag term is given by the Ergun equation or
ber for the one-fourth scale simplified model similar relationships and the scaled particles are not so
remained roughly equal to 0.34 RepE over a wide small that interparticle surface forces come into play.
range of values for RepE , whereas the particle
Reynolds number for the one-sixteenth scale model 5.2 Clusters
was roughly 0.12 RepE .
Using these Reynolds number scale factors, the In the freeboard of a bubbling bed or in the upper
error in the dimensionless drag coefficient L=s u0 portion of a circulating bed where particles generally
using the simplified scaling models is shown on Fig. are considered to act in clusters or groups, a similar
15 for u0 /umf of 10, plotted as a function of RepE examination of scaling of the gas-to-solid drag can
based on parameters for the exact scaled bed. For a be made. Consider all the particles grouped into
particle Reynolds number of 1000 or less, which corre- clusters with an effective diameter dc and the clus-
sponds to pressurized beds with particles of 1 mm or ters occupying a volume fraction c of the bed
less, the error in the drag coefficient with the simplified volume. The cluster-to-gas drag will be represented
scaling laws is 20% or less for a one-quarter length by the drag coefficient for a solid sphere of diameter
scale model. The error is 40% or less for a one- d c , CD .
sixteenth length scale model. At u0 /umf of 1000 and If the reduced scale models faithfully reproduce the
uslip= umf of one-fiftieth the errors for the one-sixteenth dynamics of the exact case, the cluster dimensions
scale model are 20% or less for RepE less than 103 . For should scale directly with the linear dimensions of the
particles of 0.2 mm or less, corresponding to a bed. Thus, a one-quarter linear scale model that has a
Reynolds number of 100 or less, the errors in drag velocity one-half that of the exact case will have a
coefficient are minimal. When the Ergun equation cluster Reynolds number (Redc ) one-eighth that of
applies for the drag coefficient, a one-quarter scale the exact bed. From the relationship of CD with Re
model based on the simplified scaling laws should be the change of CD with model scale at a given Reynolds
valid for any conditions. A one-sixteenth scale model number of the exact bed can be determined. Figure 16
shows the shift in CD using the CD relationship of diameter, equal in magnitude to the bed diameter or
White (1974) for length scales of one-fourth, one- one order of magnitude smaller. From these considera-
eighth, and one-sixteenth of the exact bed length as a tions, the Reynolds number based on the cluster dia-
function of the cluster Reynolds number of the exact meter should be 104 or larger in an atmospheric
bed. Also shown in the figure is the typical Reynolds combustor with a cluster diameter of 0.2 m. The cluster
number of an atmospheric combustor with a 0.3 m Reynolds number should be 105 or larger in a pressur-
cluster diameter, approximately 1:5 104 . In a bub- ized combustor. From Fig. 16 it can be seen that a one-
bling bed, the cluster diameter in the freeboard should quarter scale or an eighth scale model should have
be at least equal in size to the diameter of bubbles drag coefficients similar to those of the exact bed.
erupting at the bed surface. For beds with horizontal For pressurized beds, the drag coefficients should be
tubes, the bubble diameter will be equal to or larger very close in magnitude.
than the horizontal tube spacing. In a bubbling bed Considering the drag acting on a single particle in a
without tubes, the bubbles and clusters can be much dispersed flow, the validity of the simplified relation-
larger. In an open circulating bed, the cluster diameter ships can be examined by comparing the terminal velo-
is more difficult to determine. It is reasonable to city predicted by the simplified relationship and the
assume that its diameter is proportional to the bed exact value.
The errors in ut /umf is shown in Fig. 17 for simpli- Note that this is a subset of the simplified scaling laws
fied scale models at two different linear dimensions. presented above with the solid-to-gas density ratio
Scaling a combustor with comparatively small parti- removed.
cles, 0.2 mm or less, gives good agreement for ut /umf Recent results indicate that in most scaling interac-
even at one-sixteenth linear scale, while for large par- tions, elimination of the ratio of solid-to-gas density
ticles a linear scale of one-fourth gives fair agreement ratio from the set of scaling parameters leads to dis-
for ut /umf . Since u0 /umf is held constant in the simpli- crepancies between the scale models and the larger
fied scaling laws, close agreement of ut /umf also results beds being simulated; see Glicksman et al. (1998).
in close agreement of ut /u0 .
tribution, and solid or gas flow fluctuations in the the model, along with the gas pressure and tempera-
return loop do not influence the riser behavior; see ture, determines the values of f and m. The particle
Rhodes and Laussman (1992). density for the model is chosen to match the density
Proper conditions must be chosen to design a scale ratio, so that
model to match the dimensionless parameters of the
commercial bed. To model a gas fluidized commercial f f
¼ ð11Þ
bed, a scale model using air at standard conditions is s m p c
most convenient, although several investigators have
used other gases; see Fitzgerald and Crane (1980), where the subscript m is for the model and c is for the
Fitzgerald et al. (1984), Chang and Louge (1992) or commercial bed. For the remaining parameters, the
pressurized scale models, Almstedt and Zakkay form of Eq. (6) will be chosen for the dimensionless
(1990), Di Felice et al. (1992 a,b). The gas chosen for parameters. Combining the Reynolds number based
Figure 17 Error in terminal velocity using simplified scaling at 1013 kPa and 800 C. (From Glicksman et al., 1994.)
so that the ratio of solids flow to solids density scales as The simplified scaling relationships, Eq. (8) offer some
the ratio of the cube root of the kinematic viscosity. flexibility in the model design, since fewer parameters
Once the model fluid and its pressure and tempera- must be matched than with the full set of scaling rela-
ture are chosen, which sets the gas density and viscos- tionships. When the fluidizing gas, and the pressure
ity, there is only one unique set of parameters for the and temperature of the scale model, are chosen, the
model that gives similarity when using the full set of gas density and viscosity for the scale model is set.
dimensionless parameters. The dependent variables, as The model must still be geometrically similar to the
nondimensionalized as X=D, y=D, u=u0 are the same in commercial bed. There is still one free parameter.
the respective dimensionless time and spatial coordi- Generally, this will be the linear scale of the model.
nates of the model as the commercial bed. The spatial For the simplified scaling relationships, the gas-to-
variables are nondimensionalized by the bed diameter, solid density ratio must be maintained constant,
so that the dimensional and spatial coordinates of the
f f
model are proportional to the two-thirds power of the ¼ ð20Þ
kinematic viscosity, as given by Eq. (13), s m p c
Scale model,
Given Commercial bed full scaling laws
Temperature ( C) 850 25
Gas viscosity (105 kg=ms) 4.45 1.81
Density (kg=m3 ) 0.314 1.20
Derived from scaling laws
Solid density sc 3.82sc
Bed diameter, length, etc. Dc 0.225Dc
Particle diameter dpc 0.225dpc
Superficial velocity u0c 0.47u0c
Volumetric solid flux ðGs =s Þc 0.47ðGs =s Þc
Time tc 0.47tc
Frequency fc 2.13fc
With f of the model set by the fluidized gas and its full scaling relationship. With U0 and s set the solids
state, the solid density in the model follows from Eq. recycle rate can be determined by
(20). Choosing the length coordinate of the model, Dm ,
the new free parameter, the superficial velocity in the Gs Gs
¼ ð23Þ
model is determined so that the Froude number s uo m s u o c
remains the same,
Gm u0m Dm
! ! ¼ sm ¼ fm ð24Þ
u20 u20 Gsc sc u0c fc Dc
¼ ð21Þ
D gD
m c Finally, the mean particle size for the model as well as
the sphericity and particle size distribution must be
so that
determined. The particle size is determined by the
1=2 need for equal values of u0 /umf between the model
ucm Dm
¼ ð22Þ and the commercial bed:
u0c Dc
u0 u0
Note that in the simplified case the velocity scaling is ¼ ð25Þ
not uniquely tied to just the gas properties as it is in the umf umf
Scale model,
Given Commercial bed full scaling laws
Temperature ( C) 850 25
Gas viscosity (105 kg=ms) 4.45 1.81
Density (kg=m3 ) 3.14 1.20
Pressure (bar) 10 1
Derived from scaling laws
Solid density sc 0.382sc
Bed diameter, length, etc. Dc 1.05Dc
Particle diameter dpc 1.05dpc
Superficial velocity u0c 1.01u0c
Volumetric solid flux ðGs =s Þc 1.01ðGs =s Þc
Time tc 1.01tc
Frequency fc 0.98fc
Scale model,
Given Commercial bed full scaling laws
Temperature ( C) 850 20
Gas viscosity (105 kg=ms) 4.45 1.19
Density (kg=m3 ) 3.14 4.34
Pressure (bar) 10 1
Derived from scaling laws
Solid density sc 1.38sc
Bed diameter, length, etc. Dc .334Dc
Particle diameter dpc .334dpc
Superficial velocity u0c .58u0c
Volumetric solid flux ðGs =s Þc .58ðGs =s Þc
Time tc .58tc
Frequency fc 1.7fc
1=2 1=4
ucm Dm dpm Dm
ðumf Þm ¼ ðumf Þc ¼ ðumf Þc ð26Þ ¼ ð31Þ
u0c Dc dpc Dc
In general, umf is a function of the particle diameter when both ðRedp Þm and ðRedp Þc > 1000.
and gas properties, as well as and emf . Once the Tables 5 and 6 show the values of the mean particle
fluidizing gas and the length of scale of the model are diameter for models of an atmospheric and pressurized
chosen, the proper particle diameter is that which gives commercial bed, for different selected linear scale
the value of umf needed in Eq. (26). ratios between the model and the commercial bed.
If both the model and commercial bed are in the By the use of the simplified scaling parameters, the
region where the respective Reynolds number based linear scale factor can be changed as shown in Tables 5
on particle diameter and gas density is very low, then and 6. Note that as the linear scale factors are changed
a single algebraic relationship can be developed. In the particle diameter changes much more slowly. The
that region, scale model of the 10 atmosphere bed has a mean par-
ticle diameter that is quite close to the mean particle
s dp2 diameter of the commercial bed. The model particles
umf ð27Þ
m have a substantially lower density in this case.
1=4 It is not clear where cohesive forces will become
dpm fc mm 1=2 Dm
¼ ¼ ð28Þ important. The use of very dense particles (for the
dpc fc mm Dc models of the one atmosphere bed) will cause a shift
of the boundary of cohesive influence as given, for
when both ðRedp Þm and ðRedp Þc < 20. example, by Geldart’s classification. However, ade-
When the Reynolds number of the model and commer- quate experimental data are still lacking with such
cial bed are both very large, dense fine particles to set the limits of cohesive influ-
2 s ence.
Umf dp ð29Þ Note that for completeness the nondimensional par-
f
ticle size distribution, the sphericity, and the internal
Combining Eq. (29) with Eqs. (22) and (25), we find angle of friction (for slugging and spouting beds)
should also be matched between the two beds.
u2mfm dpm ðs =f Þm u20m Dm
¼ ¼ 2 ¼ ð30Þ
u2mfc dpc ðs =f Þc u0c Dc 7.3 Hydrodynamic Scaling of Bubbling Beds
Since the gas-to-solid density ratio of the model and Experiments using scaled models of bubbling beds
the commercial beds must be the same to satisfy the have been carried out since 1980 using the scaling rela-
simplified scaling relationships Eq. (30) becomes tionships presented in previous sections. The earlier
Commercial beds Particle diameter of model with bed linear scale factor
work was concerned with experimental testing and ver- experiments, additional scaling parameters were unin-
ification of the scaling relationships. Hot bed behavior tentionally matched.
was compared to cold scaled models in several studies. Fitzgerald and Crane (1980) were among of the first
In others, cold beds of different sizes were compared to to evaluate the full set of hydrodynamic scaling para-
each other. For bubbling beds there is the sense of meters. They compared the hydrodynamics of two
what constitutes sufficient verification of the scaling scaled beds using pressure fluctuation measurements
relationships. Since bubbles are the prime motive and movies. In one bed cork particles were fluidized
agents for both gas and solids displacement, a detailed with air; the other bed used sand fluidized with pres-
verification should involve comparison of bubble prop- surized refrigerant 12 vapor. Movies showed qualita-
erties through the large bed and the scale model. This tive agreement between bubble growth and the solids
is preferable to comparison of overall performance flow in the beds.
parameters for a bed based on input and exhaust mea- Fitzgerald et al. (1984) measured pressure fluctua-
surements. tions in an atmospheric fluidized bed combustor and a
quarter-scale cold model. The full set of scaling para-
meters was matched between the beds. The autocorre-
7.4 Verification of Scaling Relationships for Bubbling lation function of the pressure fluctuations was similar
and Slugging Beds for the two beds but not within the 95% confidence
levels they had anticipated. The amplitude of the auto-
Most early experiments devoted to verifying the scaling correlation function and the experimentally deter-
relationships have dealt with the full set of scaling mined time-scaling factor differed from the expected
relationships. Recent experiments have dealt with the value. They suggested that the differences could be
simplified set of dimensionless parameters. In some due to electrostatic effects.
Table 6 Scale model of 10 atm commercial hot bed using the simplified scaling
relationships
Commercial beds Particle diameter of model with bed linear scale factor
Commercial bed t ¼ 800 C, P ¼ 1 atm, s ¼ 2500 kg=m3 gas:air. Model bed s ¼ 896 kg=m3
gas:air at STP.
of Glicksman et al. (1993b), who found it essential to The earliest scaling studies were directed at atmo-
match the density ratio when scaling circulating flui- spheric bubbling bed combustors. To date, a rich vari-
dized bed hydrodynamics. ety of questions have been addressed.
Jones and Glicksman (1986) constructed a model
of the 20 MW bubbling bed pilot plant jointly spon-
8 APPLICATIONS OF SCALING TO sored by the Tennessee Valley Authority and the
COMMERCIAL BUBBLING FLUIDIZED BED Electric Power Research Institute (EPRI) at
UNITS Paducah, Kentucky. Figure 23 shows a photograph
of the model of the in-bed tubes installed in the scale
A substantial number of experimental investigations model. The model, which is roughly 100 by 120 cm in
have demonstrated the validity of scaling. This has cross section, simulates two-thirds of the entire 20
increased awareness of the concept and confidence in MW pilot plant. Care was taken to match the pilot
its application. Scaling has many useful applications. plant tube bundle geometry and distributor design.
The dynamic characteristics of different bed designs Steel grit particles with the same dimensionless size
can be quickly compared. The influence of bed dia- distribution and sphericity as the hot bed material
meter on hydrodynamic behavior can be studied by were used. The full set of scaling parameters was
the use of several different models. The models allow matched in the model and the combustor. The largest
easy experimental examination of existing operating discrepancy was in the solid-to-gas density ratio,
characteristics. The beds also can be used to confirm which was 18% smaller in the model than in the
quickly the influence of proposed modifications in pilot plant.
operating parameters and bed geometry. Since the The measured bubble velocity for an actively bub-
models usually operate at ambient conditions, it is pos- bling bed was found to agree closely with the drift flux
sible to instrument them to observe detailed behavior. form proposed by Davidson and Harrison (1963). In
This allows a better understanding of the fundamental contrast, the volumetric flow rate of the bubbles was
physics as well as the identification of hydrodynamic found to be far less than that predicted by the two-
factors needed for proper correlation of performance. phase hypothesis; see Fig. 24.
wear rate. The second row of tubes is shielded by the tion, bubble frequency, vertical bubble size, jet half-
first row, reducing the incidence of strikes by large angle, and others. They also found considerable gas
high-velocity bubbles. In the interface between an leakage from the bubbles to the emulsion phase. This
actively fluidized and a slumped bed, there is a flow data covers much larger bed sizes than had been
of particles down the slumped surface, which causes examined before.
higher wear. The relative changes in wear rate with
tube location as well as the circumferential location
with highest wear agreed closely between the cold 9 HYDRODYNAMIC SCALING OF
model and the 20 MW combustor. CIRCULATING BEDS
Other boiler manufacturers have used scale models
to aid in the design of large bubbling bed combustors. Given the success in scaling bubbling beds, research
Yang et al. (1995) used atmospheric cold models 30 has progressed to scaling of circulating fast beds. The
cm and 3 m in diameter to simulate a large jetting initial research has focused on the verification of both
fluidized bed gasifier to be operated at about 12 atm. the full and the simplified scaling relationships for cir-
They used the full set of scaling relations that culating beds. The verification using data from com-
required the model to be about the same dimensions bustors is complicated by the difficulty in accurately
as the hot bed. A half-round bed with a transparent measuring the recycle rate of solids, an important para-
wall was used to photograph the jet behavior. They meter. Figure 28 presents the range of scaling para-
measured and developed correlations for jet penetra- meters for experimental studies undertaken for
circulating in terms of the Froude number and the Ishii and Murakami (1991) evaluated Horio et al.
Reynolds number based on particle diameter and the (1989) CFB scaling relationships using two cold CFB
superficial gas velocity. models. Solids flux, pressure drop, and optical probe
Horio et al. (1989) experimentally verified their pro- measurements were used to measure a large number of
posed circulating fluidized bed scaling laws. The solid- hydrodynamic parameters to serve as the basis for the
to-gas density ratio was not varied in the tests; thus comparison. Fair to good similarity was obtained
they effectively verified the simplified set of scaling between the beds.
laws. Two cold scaled CFBs, fluidized using ambient Tsukada et al. (1991) applied Horio et al. (1989)
air, were used in the verification. Good agreement in CFB scaling laws at several different elevated pressures
the axial solid fraction profiles was obtained for most (viscous limit scaling laws). A single bed and bed mate-
of the conditions tested. An optical probe was used to rial were used in the study. A pressure vessel was used
verify similarity in the annular flow structures and the to vary the gas pressure. They found that as the pres-
cluster velocities. sure was increased the axial solid fraction profile chan-
ged, indicating a change in the hydrodynamics. In this
study, the gas-to-solid density ratio changes with pres-
sure level. It is likely based on the recent results of
Glicksman et al. (1993b) that the change in gas-to-
solid density ratio led to the influence of the pressure
level on the bed hydrodynamics. This points out a
deficiency of the viscous limit, Horio’s (1989) scaling
relationship.
Glicksman et al. (1991a) made scaling comparisons
between an experimental circulating fluidized bed com-
bustor and a scaled cold model based on the full set of
scaling laws. Due to uncertainties in the hot bed solid
circulation measurements, the cold bed solids flux was
Figure 26 Particle accumulation in slumped zone adjacent
to active bed: air velocity through active region ¼ 151 cm=s; adjusted until the average bed solid fraction matched
air velocity through slumped region ¼ 7:3 cm=s; numbers that of the hot bed. Good agreement was obtained
above heater are heat transfer coefficient in W=m2 C after between the vertical solid fraction profiles except
15–30 min. (From Glicksman et al., 1992.) near the top of the beds. Disagreement may have
been due to the differences between the gas-to-solid A modified set of the simplified scaling relationships
density ratio in the cold model and the hot bed. were also evaluated by Glicksman et al. (1993b) in a
Good agreement was also obtained in the comparison series of comparison tests using circulating beds in
of the probability density distribution and the Fourier which the gas-to-solid density ratio was not held con-
transform of the pressure fluctuations. stant. The average solid fraction profiles, Fig. 29, solid
Glicksman et al. (1993a) evaluated the full set of fraction probability density functions, and power spec-
scaling laws for circulating fluidized beds. Solid frac- tral densities were all in poor agreement. It is believed
tion data were obtained from the 2.5 MWth Studsvik the beds were operating near the point of incipient
atmospheric CFB prototype. The full set of scaling choking condition as predicted by the Yang (1983)
laws was evaluated through solid fraction profile com- correlation. Because this correlation indicates that
parisons between Studsvik and a one-quarter scale cold choking is a strong function of the solid-to-gas density
model. Fairly good agreement was obtained; the pro- ratio, it was concluded that this parameter must be
files most closely matched in the top of the beds. matched to model bed hydrodynamics near the bound-
Differences between the profiles were attributed to ary between different flow regimes.
uncertainty in the hot bed solid flux measurements The simplified scaling relationships were used by
and to the mismatch in the solid-to-gas density ratio. Glicksman et al. (1993b) to compare two geometrically
similar beds, one having linear dimensions four times ment; see Fig. 30. The probability density functions and
larger than the other. In one series of tests, properly power spectral densities also agreed well.
sized plastic particles were used in both beds; in another Glicksman et al. (1993b) verified the simplified scal-
test series, glass particles were used in the two beds. The ing laws for hot beds by comparing the solid fraction
average solid fraction profiles showed excellent agree- profiles for the Studsvik bed, the one-quarter scale cold
Figure 29 Solid fraction profiles, glass/plastic viscous limit scaling; density ratio mismatched: low velocity case. (From
Glicksman et al., 1993a.)
model, and a one-sixteenth scale cold model. The one- tions of the full set of scaling laws. The density ratio
sixteenth scale model had a cross sectional area of was not matched exactly between the hot bed and the
16 cm2 to simulate a 2.5 MW combustor! The average two cold beds, which may have affected the agreement.
solid fraction profiles were in good agreement for most Figure 31 provides a typical comparison of the solid
of the conditions tested. The agreement was excellent fraction profiles in the three beds. The authors con-
between the one-quarter scale cold model, which used cluded that the simplified set of scaling laws, which
the full set of scaling laws, and the one-sixteenth scale includes the solid-to-gas density ratio, gives acceptable
model that used the simplified set of scaling laws. Thus results over a wide range of particle densities and bed
any disagreement between the Studsvik bed and the sizes, even when the length ratio is as small as one-
one-sixteenth scale model is not due to the simplifica- sixteenth for an atmospheric combustor.
Figure 31 Solid fraction profiles, hot bed scaling with simplified scaling laws, low velocity, 49% primary air. (From Glicksman
et al., 1993a.)
Figure 32 Solid fraction profile comparison: between pressurized circulating fluidized bed combustor and one-half size scale
model based on simplified scaling. Two different operation conditions. (From Glicksman et al., 1995.)
Ye-Mon Chen
Shell Global Solutions US, Houston, Texas, U.S.A.
1 INTRODUCTION the end of the riser. The product vapor goes to the
main fractionation and gas plants for product separa-
1.1 What Is FCC? tion. The spent catalyst enters the stripper where steam
is introduced to further recover entrained/adsorbed
Fluid catalytic cracking (FCC) is the primary conver- hydrocarbon products. The spent catalyst then passes
sion unit in most refineries in the United States. It through a spent catalyst transfer line and enters the
converts low-value heavy components of crude oil regenerator. As the catalyst activity is restored by
into a variety of high-value lighter products. A modern burning off the coke layer on spent catalyst, a large
FCC unit consists of three major sections: the reactor/ amount of heat is released, which heats up the regen-
regenerator section, the main fractionation/gas plant erated catalyst. The flue gas from the regenerator goes
section, and the flue gas cleaning/power recovery sec- to flue gas cleaning and the power recovery system.
tion. This chapter focuses mainly on the reactor/regen- The regenerated catalyst then returns to the reactor
erator section of the FCC process, paying special riser. This completes the catalyst circulation loop.
attention to its relationship with fluidization and
fluid–particle systems. Readers who are interested in 1.2 How the FCC Unit Fits in a Typical US
broader aspects of the FCC process can refer to several Refinery
recent publications, such as Sadeghbeigi (1995) and
Wilson (1997). The FCC unit usually has the second highest through-
In the reactor/regenerator section of the FCC pro- put after the crude distillation unit in most U.S. refi-
cess, liquid hydrocarbon feedstock is preheated, mixed neries. Crude oil contains hydrocarbons of different
with steam, and injected into the riser reactor through boiling points ranging from light gas to residue.
feed nozzles. Hot regenerated catalyst is drawn from Typically, crude oil is first distilled in an atmospheric,
the regenerator and contacts the hydrocarbon feed. or crude, distillation unit to produce a wide range of
The two are thoroughly mixed in the lower riser by hydrocarbon products according to the different boil-
fine atomization through the feed nozzles. The tem- ing points in the original crude. The residue from the
perature of regenerated catalyst drops rapidly as the atmospheric distillation unit, the atmospheric residue,
feed vaporizes and reacts, converting the feedstock into is then fed to the vacuum distillation unit, which pro-
lighter products while a coke layer deposits on the duces vacuum gas oil and vacuum residue. The vacuum
spent catalyst, temporarily deactivating the FCC cata- gas oil from the vacuum distillation unit is the tradi-
lyst. Product vapor and spent catalyst are separated by tional FCC feedstock with typical boiling points ran-
riser termination devices, such as cyclone systems, at ging from 650 to 1050 F. Depending on the type of
axis and the feed injection, plays a significant tion. Most of the modern FCC units have the side
role in mixing the feed with catalyst and hence entry configuration, as shown in Fig. 1, in which
achieving a desirable temperature profile in the regenerated catalyst is lifted from the riser bottom
riser. Most FCC units have feed nozzles to the feed injection point by fluidizing gas, normally
installed through riser shrouds at a fixed steam. The section between the riser bottom and the
angle. A new feed nozzle design by Chen et feed injection point is called the catalyst preaccelera-
al. (1999) as shown in Fig. 4, enables an FCC tion zone. A typical lift steam velocity for the preac-
unit to adjust the feed injection angle while celeration is in the range between 3.0 to 10.0 ft/s,
using the existing riser shrouds. This enables which is in the turbulent to fast fluidization regime.
an FCC unit to optimize the mixing of feed The length for the preacceleration zone is in the range
with catalyst by adjusting the feed injection from 6 to 20 feet. At the end of preacceleration, mul-
angle to achieve the best performance of the tiple feed nozzles located on the circumference of the
unit. riser are used to inject the feedstock. With proper
designs of the preacceleration and feed nozzles, back-
Commercial experience has confirmed that using better mixing of catalyst in the lower riser is minimized with
feed nozzles can substantially reduce dry gas and the side entry configuration.
increase gasoline yield. These results are in line with Most of the older FCC units have the bottom entry
the expectation that thermal cracking reactions, which configuration. In this case, a single or multiple nozzles
are the primary source for dry gas, are reduced. As a located at the bottom of the riser are used to introduce
result, catalytic cracking reactions are maximized, and the feedstock directly into the region where hot regen-
more desirable cracking products are produced. erated catalyst enters the riser. The catalyst condition
Depending on unit constraints and market demands, in the riser bottom is denser, chaotic, and highly ero-
other steps can be taken, such as increasing riser tem- sive. Earlier feed nozzles with the bottom entry config-
perature, increasing catalyst activity, or adding ZSM-5 uration emphasize mechanical robustness, but with
catalyst, in order to take full advantage of the dry gas primitive feed atomization and poor performance.
reduction. Because of this disadvantage many bottom entry
The second component of the feed injection system FCC units have been converted to side entry units in
is the design of the low riser transition prior to feed recent years. However, recent commercial experience
injection. There are two common configurations of has demonstrated that, with proper feed nozzle design
the lower riser to prepare for feed injection; one is (Chen, 2002), a bottom entry configuration can achieve
bottom entry and the other is side entry configura- feed contacting and riser performance similar to the
Designing a standpipe is simple from mechanical higher rising velocity, these large bubbles have a better
viewpoint, and yet very complex from hydrodynamic chance to escape from the inlet hopper, thus minimiz-
viewpoint. A standpipe is typically a vertical, or nearly ing gas entrainment into the standpipe. The complex
vertical, column with a constant diameter, which is the nature of the design is that a poorly designed inlet
simplest mechanical structure of the entire FCC unit. hopper can either entrain too much gas or cause deflui-
There are only two design elements that influence the dization. The overall hydrodynamic behaviors of the
standpipe flow besides the slide valve, which controls two cases are again very similar, i.e., the standpipe
the outlet end of the standpipe. The first design ele- cannot build proper pressure and the catalyst flow is
ment is aeration along the standpipe. As pressure unstable. If catalyst defluidizes in the inlet hopper, put-
increases down the standpipe, the gas phase within ting a local aeration into the hopper can solve the
the fluidized catalyst is compressed and must be problem (Chen, 1986). However, if the hopper draws
made up by additional fluidizing gas. Otherwise, a too much gas into the standpipe, any additional aera-
long standpipe can become defluidized at the lower tion will only worsen the situation. One particular pro-
end and restrict the catalyst flow. The complex nature blem with the conventional inlet hopper is that the
of the aeration design is that an overaerated standpipe amount of gas entrainment cannot be controlled.
has a hydrodynamic behavior similar to that of an Thus a standpipe can suddenly become unstable as
underaerated standpipe, i.e., it cannot build proper the catalyst circulation rate is increased or decreased
pressure, and the catalyst flow is unstable. In order beyond a certain rate.
to distinguish the two cases, one must look beyond A new standpipe inlet design has recently been
the superficial symptoms. In the overaerated case the developed by Chen and Brosten (2001), as shown in
instability is caused by the formation and release of Fig. 8. Instead of using an inlet hopper, this new inlet
large bubbles, whereas in the underaerated case it is design uses a disk positioned directly below the stand-
caused by defluidization leading to a stick-slip flow. pipe inlet. The concept is to trap fluidizing gas from
The second design element is the standpipe inlet. below, causing a local partial defluidization above the
Because both stripper and regenerator fluidized beds disk and forming a dense bed region near the proxi-
are operated at relatively high gas flow rates, excessive mity of the standpipe inlet. A small amount of fluidi-
gas can be drawn into the standpipe through the inlet, zation gas can be introduced above the disk to control
which is highly undesirable. To address the gas entrain- the fluidization condition of the standpipe inlet region
ment issue, the conventional design is to add an inlet independently of regenerator and stripper fluidization
hopper at the top of the standpipe. The typical inlet conditions, which are set by process requirements.
hopper size is about 2 to 2.5 times the standpipe dia- Commercial experience has demonstrated that both
meter. The concept of the inlet hopper is to provide catalyst circulation rate and stability can be signifi-
enough residence time for small entrained bubbles to cantly improved by simply replacing the conventional
coalesce into larger ones. Since larger bubbles have a standpipe inlet with the better inlet design of Fig. 8.
REFERENCES
Bedaw RE. Atomizing spray nozzle for mixing a liquid with a Dries H, Patel M, van Dijk N. New advances in third-stage
gas. U.S. Patent 5,240,183, 1993. separators. World Refining 30–34, Oct. 2000.
Chen Y. A theoretical investigation of an aerated hopper Geldart D. Types of gas fluidization. Power Technol
flow. J Chinese Inst Chem Eng 17:195, 1986. 7:285–292, 1973.
Chen Y. Feed nozzle. U.S. Patent 6,387,247, 2002. Gwyn JE. Entrance, exit and wall effects on gas/particulate
Chen Y, Brosten D. Standpipe inlet. U.S. Patent, 6,228,328, solids flow regimes. Proceedings of 4th International Conf
2001. on Circulating Fluid Beds, Somerset, Pennsylvania, 1993,
Chen Y, Dewitz TS. Feed nozzle. U.S. Patent 5,794,857, pp 679–684.
1998. Harauch J. Nozzle for use in fluidized catalytic cracking.
Chen Y, Rangachari R, Jackson R. Theoretical and experi- European Patent EP 0-717-095-A2, 1995.
mental investigation of fluid and particle flow in a vertical Herbst JA. Fluid catalytic cracking process and apparatus
standpipe. Ind Eng Chem Fundam 23:354, 1984. for more effective regeneration of zeolite catalyst. U.S.
Chen Y, Brosten D, Nielson JW. Feed nozzle. U.S. Patent Patent 4,814,068, 1989.
5,979,799, 1999. Kauff DA, Bartholic CA, Steves CA, Keim MR. Successful
de Kruijff GT, van Hattem A. Apparatus for solids–fluid application of the MSCC process. Paper AM-96-27,
separation. U.S. Patent 5,1998,005, 1993. NPRA meeting, 1996.
Richard A. Newby
Siemens Westinghouse Power Corporation, Pittsburgh, Pennsylvania, U.S.A.
Kellogg Fuel gas for combustion 20,000 MJ/h development unit testing Sierra Pacific Power
KRW gasifier turbine power completed on wide range of coals; Co., 1994;
800,000 MJ/h coal demonstration plant Demuth and Smith,
startup initiated 1998; 1998
air-blow operation;
2100–2400 kPa operating pressure
IGT Coke oven fuel gas 105,000 MJ/h commercial plant Bryan and Hoppe,
U-gas gasifier operational; 1998
air-blow operation;
160–220 kPa operating pressure
Rheinbraum AG Fuel gas for combustion Lignite operation only; Renzenbrink et al.
HTW gasifier turbine power; 450,000 MJ/h demonstration project 1998
syngas for methanol completed;
air and oxygen-blown operations
Foster Wheeler Fuel gas and char for 10,000 MJ/h development unit testing Robertson, 1995
Development Corp. combustion turbine completed on range of coals;
power 50,000 MJ/h pilot plant construction
completed 1998;
air-blow operation
British Coal Fuel gas for turbine 40,000 MJ/h pilot plant operation Dawes et al., 1997
power; completed 1992;
char for steam generation air-blow operation;
integrated gas cleaning and turbine
Tsinghua Fuel gas for town gas; 5,000 MJ/h development unit operated Lu et al., 1996
University char for pyrolysis heat on Chinese coals;
and steam air-blown operation;
100,000 MJ/h demonstration plant in
design
fed in either dry or paste form; the paste form is drain the raw char product. The gasifier must interface
claimed to improve feed reliability and cost. If the with the pressurized air delivery system; with the pres-
application is to provide a pressurized fuel gas to fire surized coal and sorbent preparation and feed systems;
a combustion turbine, the raw fuel gas must be well with the raw fuel gas pressurized processing system;
cleaned of particulate and alkali species to protect the and with the pressurized char processing system.
combustion turbine. If the pressurized fuel gas is used Coals may be fed in either dry or paste forms. The
to fire a process heater, such high levels of cleaning raw fuel gas will have a low heating value of about
may not be needed. Coals will generally result in a 3.3–5.6 MJ/Nm3 (LHV), similar to the raw fuel gas
raw fuel gas that needs to be desulfurized, and the issued from the pressurized air blown bubbling bed
fuel gas ammonia content may be a concern for stack partial oxidation gasifier.
gas NOx emissions. The third scheme shown, Fig. 1C, is a low-pressure
Figure 1B is representative of pressurized bubbling bubbling bed coal pyrolysis process based on a parallel
bed carbonization of coal. Here, cyclone fines recycling circulating bed char combustor that provides hot cir-
is not used, and the carbonizer vessel and cyclone will culating solids to the gasifier. The bubbling bed pyro-
Carbona Corp Fuel gas for combustion 100,000 MJ/h pilot unit operated on DeLong et al.,
(IGT Renugas gasifier) turbine power range of biomass fuels; 1995
integrated gas cleaning;
750,000 MJ/h alfalfa-stem plantation
unit in design
Cratech gasifier Fuel gas for small 10,000 MJ/h pilot unit operating on Craig and Purvis,
combustion turbines cotton gin mill waste; 1998
integrated gas cleaning;
100 kW turbine operation planned
Foster Wheeler Fuel gas for combustion 70,000 MJ/h demonstration testing Stahl and Neergaard,
Energia Oy turbine power ongoing since 1995; 1998
Bioflow gasifier integrated with 6 MWe combustion
turbine operation;
wood chip fuel
Lurgi gasifier Fuel gas for process 350,000 MJ/h unit for cement kiln fuel Hirschfelder and
heat; gas in operation; Vierrath, 1998;
fuel gas for combustion 160,000 MJ/h unit for combustion turbine DeLange and Barbucci,
turbine power in design; 1998
air-blown operation
Foster Wheeler Fuel gas for process 55,000–230,000 MJ/h commercial Nieminen and Kivila,
Energia Oy heat and steam units in operation; 1998
lime kiln process heat and steam;
air-blown operation
Termiska Processor Fuel gas for process Two, 50,000 MJ/h units in commercial Barducci et al., 1995;
AB (TPS) gasifier heat and steam; operation for steam generation; Pitcher et al., 1998
fuel gas for combustion 300,000 MJ/h eucalyptus unit designed
turbine power for combustion turbine plant in Brazil;
100,000 MJ/h wood chip unit for
combustion turbine in UK in
construction
Battelle gasifier Fuel gas for process 145,000 MJ/h wood chip in Farris et al., 1998
heat and steam; operation for steam generation;
fuel gas for combustion future conversion for combustion
turbine power turbine operation planned
lyzer is fluidized by steam, and its design and operation must be processed to meet the requirements of those
must be closely integrated with the circulating bed applications. The raw fuel gas may also be cooled,
combustor. The pyrolysis gasifier must also interface scrubbed and compressed to act as a combustion tur-
with low-pressure coal and sorbent preparation and bine fuel gas. The product fuel gas, after cooling and
feed systems. The low-pressure raw fuel gas can be steam condensation, will have a medium heating value
used for process heat or for steam generation, and it of about 12–16 MJ/Nm3 (LHV). The fluidized bed
combustor produces substantial steam for export, and gasifier must interface with the pressurized air delivery
a separate coal stream may be fed to the fluidized bed system (only air being considered feasible with biomass
combustor to raise additional steam. fuel gasification applications); with the pressurized fuel
Figure 2 shows representations of biomass fuel flui- and inert feed systems; with the raw fuel gas pressur-
dized bed gasifiers. Four types are illustrated: (A) pres- ized processing system; and with the pressurized ash
surized bubbling bed partial oxidation of biomass processing system. Biomass fuels will be fed only in
fuels; (B) pressurized circulating bed partial oxidation dry form.
of biomass fuels; (C) low-pressure, circulating bed par- The application is to provide a pressurized fuel gas
tial oxidation of biomass fuels; and (D) low-pressure to fire a combustion turbine, and the raw fuel gas must
circulating bed pyrolysis. For all of these, proper bio- be well cleaned of particulate and alkali metal species
mass fuel drying and sizing is critical to their successful to protect the combustion turbine. Here, it is expected
operation. that tar generation will be small, and special provision
Figure 2A is representative of pressurized bubbling for tar cracking or removal will not be needed. The fuel
bed air blown partial oxidation of biomass fuel. The gas must be cooled to the extent needed to achieve this
sketch indicates that fines recycling might be used, with cleaning, probably to a maximum temperature of
nonmechanical valves to control the reinjection of fines 650 C. Fuel gas ammonia content may be a particular
into the fluidized bed gasifier. The vessel will drain a concern for NOx stack emissions when using some
low-carbon ash product. The sketch also suggests that biomass fuels. Biomass fuels generally will not require
overbed air injection might be used for fuel gas partial fuel gas desulfurization, but some agricultural planta-
oxidation as a means for fuel gas tar destruction. The tion crops do have significant sulfur content and may
require gas desulfurization. The raw fuel gas will have process challenges of pressurizing and feeding biomass
a low heating value similar to those raw fuel gases fuels at low pressure are much less than in the two
produced in pressurized air blown coal gasifiers. previous high-pressure schemes. The low-pressure
Figure 2B is a pressurized circulating bed air blown raw fuel gas may be cooled to reduce hot gas piping
partial oxidation of biomass fuel. Here biomass fuel is expense and used to fire a conventional boiler–steam
fed as a dry solid into a circulating bed of inert parti- generator with little gas cleaning. Alternatively, the
cles. While circulating bed gasification may not be raw fuel gas might be cooled and scrubbed so that it
practical for most coals, the high reactivity of biomass can be compressed to provide a combustion turbine
fuels makes circulating bed gasification an effective fuel gas. It is expected again that tar generation will
process. The raw fuel gas must be processed so that be small in most cases.
it can serve as a combustion turbine fuel gas. Again, it The fourth scheme, Fig. 2D, is a low-pressure bio-
is expected that tar generation will be small, and spe- mass pyrolysis process based on parallel circulating
cial provision for tar cracking or removal will not be beds, one a pyrolyzer and the other a char combustor.
needed. Large bed circulation rates are used to main- The high reactivity of biomass fuels make this process
tain vigorous gas–particle mixing and nearly uniform feasible in relatively compact vessels. The raw fuel gas
temperatures throughout the gasifier. The raw fuel gas will have a medium heating value of about 13.0–16.7
heating value and carbon utilization will be in the same MJ/Nm3 (LHV). The low-pressure raw fuel gas can be
range as the other air blown gasifiers. used to fire a conventional boiler after cooling, with
The third scheme, Fig. 2C, is low-pressure (near little fuel gas cleaning; or it may be cooled, scrubbed,
atmospheric pressure) circulating bed air blown partial and compressed to act as a combustion turbine fuel
oxidation of biomass fuel. Here low-pressure biomass gas. The concentrated pyrolysis fuel gas will generally
fuel is fed as a dry solid into a circulating bed of inert contain relatively high tar content, and tar cracking, or
particles fluidized by low-pressure air and steam. The recovery, will need to be incorporated into the raw fuel
gas processing system for the combustion turbine Generally, the bubbling bed gasifiers have mean bed
application. Fuel gas tar content may also be a process particles characteristic of Geldart group B classifica-
concern for the boiler fuel gas application. tion (Geldart, 1973), with no overlap into the A or D
New techniques for pyrolysis of biomass fuels by groups. The pressurized bubbling bed gasifier appears
indirect heat input are currently at various stages of to approach the turbulent fluidization regime much
development and demonstration for small-capacity more closely than does low-pressure bubbling bed gasi-
fuel gas supply applications: the BrightStar fication, indicating the potential for more vigorous flui-
Technologies low-pressure steam fluidized pyrolyzer dization, but also more extensive particle elutriation.
using fuel firing to heat the windbox (Chau and The low-pressure and pressurized circulating bed gasi-
Rovner, 1995); the MTCI low-pressure steam fluidized fiers operate well above the terminal velocity, although
pulse combustion pyrolyzer with immersed heat trans- the low-pressure operating region indicates that low-
fer surfaces in the bed (Monsour et al., 1995); and the pressure circulating bed gasifiers may operate very near
Iowa State University cyclic (gasification–oxidation), the transition between turbulent and fast fluidization.
steam fluidized pyrolyzer with high-temperature The single gasifier module maximum fuel capacity is
phase change materials contained in tubes immersed sometimes applied as an indicator of the feasibility of
in the bed (Stobbe et al., 1996). These new pyrolyzer competing configurations. A key factor influencing the
configurations all have significant materials, scale-up, single gasifier module fuel capacity of fluidized bed
and process integration challenges to overcome. gasification technologies is the cross-sectional area of
In AFBC applications, the convective heat transfer topped-PFBC, a fluidized bed carbonizer (see Sec. 3),
surface is arranged so that maximum steam genera- operated at 815–930 C, generates a low–heating value
tion is accomplished by cooling the product combus- fuel gas and a char product. The char is combusted in a
tion gas down to the lowest permissible stack pressurized fluidized bed combustor at 870–930 C,
temperature. In PFBC operation, the product com- producing a hot combustion product gas stream con-
bustion gas is maintained at its highest permissible taining excess oxygen. Both fluid bed reactors use in-
temperature, that of the fluidized bed combustor, so situ sulfur removal with limestone. The carbonizer fuel
that maximum turbine power and efficiency can be gas and the fluidized bed combustor product gas com-
obtained. PFBC operates at bed temperatures bine in a low-NOx topping combustor, resulting in an
(815–870 C) much lower than conventional combus- expansion gas temperature characteristic of modern
tion turbine inlet temperatures (1150–1430 C), so the heavy-duty gas turbines. The two parallel gas streams
turbine performance in PFBC is relatively poor, but it from the carbonizer and combustor are cleaned at, or
still contributes significantly to the power plant gen- near, their generation temperatures so that the com-
erating output and efficiency. bined streams meet both the gas turbine protection
Another form of PFBC, topped-PFBC, has been requirements and the power plant environmental emis-
devised to maximize the gas turbine performance. In sion standards. The hot gas cleaning system performs
the performance of both the underbed and the overbed horizontal heat transfer surface within the dense flui-
feed systems. Nonuniform char–bed mixing and vola- dized bed, achieving high heat transfer coefficients
tiles–oxygen mixing may result in bed material agglom- (250–400 W=m2 C) but also facing the possibility of
eration, high carbon losses, volatiles breakthrough steam tube erosion. Heat transfer surface is also
with above-bed burning, reduced sulfur removal, and placed above the dilute freeboard zone where convec-
other undesirable results. tive steam generation occurs. Bubbling bed combus-
Typical bubbling bed temperatures are 815–870 C, tor dense beds are generally about 1–1.5 m deep,
and excess air levels range from 20 to 25%, with allowing them to achieve acceptable pressure drops
lower reactivity coals requiring higher excess air, (15–25 kPa), while the splash zone and freeboard
higher temperatures, and fewer coal feed points than height is about 10 m from the top of the dense bed.
higher reactivity coals. Bubbling bed AFBC places The combustion gas velocity above the transport dis-
Behzad Jazayeri
Fluor Daniel, Inc., Aliso Viejo, California, U.S.A.
This chapter provides an overview of some of the more content (10 to 45 micron) of around 25–45% at about
important commercial applications of fluidized bed 66 cm/s (Pell and Jordan, 1987), and possibly as high
technology not covered in the other sections. Recent as 100 cm/s. BP has introduced several new catalysts
developments with high potentials of commercializa- with increasing selectivity to acrylonitrile. With these
tion in the near future are also discussed. changes, propylene conversion is now approaching
100%, and per pass reactor yield to propylene is
about 85%. The reaction is highly exothermic, releas-
1 CHEMICAL SYNTHESIS ing about 670–750 kJ/mol. The heat of reaction is
removed by direct generation of high pressure steam
1.1 Acrylonitrile in serpentine coils located inside the reactor. The reac-
tor hydraulic diameter (Volk et al., 1962) is reported as
The Sohio Company (now BP Chemicals) introduced 100 to 150 cm.
in 1950s what is now considered to be the most suc- Sohio initially tested the process in both a small
cessful application of fluidized beds for chemical synth- bench scale unit and a circa 61 cm pilot unit before
esis (Fig. 1). In this application, air, propane, and building the first commercial plant (circa 300 cm). Over
ammonia are reacted together in a turbulent fluid 51 units have been built, representing over 95% of the
bed of Geldart group A (Chapter 3) catalyst to pro- world’s acrylonitrile capacity. The largest units now
duce acrylonitrile by the reaction: utilize reactors with estimated diameters of 800 to pos-
sibly 1000 cm.
3
C3 H8 þ NH3 þ O2 ! C2 H3 CN þ 3H2 O ð1Þ In the process, air is introduced uniformly into the
2 bottom of the reactor via a distributor plate. The pro-
The features of the original process as reported by pylene and ammonia are introduced into the fluidized
Kunii and Levenspiel (1991a) and some recent devel- bed above the distributor plate via a separate sparger.
opments follow. Reactor operating conditions are 400 The design of these gas distribution systems has
to 500 C at about 1.7 atmosphere, with a reactor con- evolved over the years improving the reactor yield
tact time of 5 to 20 seconds. In addition to acryloni- (Ohta and Yokura, 1996.)
trile, carbon monoxide, carbon dioxide and water, The original units used several parallel sets of three-
small quantities of hydrogen cyanide, acetonitrile, stage cyclones for catalyst recovery. Recently, one
and acrolein are also produced. The original plants cyclone vendor has announced the use of multiple
were designed to operate at around 52 cm/s superficial cyclones made of a single stage followed by two second
velocity. Modern plants are now operating with fines stages.
dized bed catalyst discussed above. Details of the cat- reactor where it is reduced by reaction with butane.
alyst structure are provided by Contractor et al. (1978). The reactor operates at high velocities with a relatively
A significant departure from conventional fluidized short contact time. Per pass conversion of butane is
bed design is the use of the catalyst as oxygen carrier. therefore low, requiring recycling of unconverted
The catalyst is oxidized in a separate regeneration ves- butane. Butane feed concentration can be high, possi-
sel by contact with air. It is then circulated to a riser bly over 20 mol%. To avoid catalyst overreduction in
permitting reactor operation within the flammability fixed bed plants due to limited availability of coal tar
envelope (Graham, 1970). The reactor operates at naphthalene. Attempts to develop an o-xylene-based
around 2.7 atmosphere and between 345 and 385 C, fluid bed process have been hampered by the unavail-
with a velocity of between 30 and 60 cm/s and a con- ability of a suitable attrition resistance catalyst support
tact time of 10 to 20 seconds (Graham and Way, 1962). (Bolthrunis, 1989).
The catalyst is Geldart group A with fines content (10 Excellent accounts of the development, operation,
to 45 micron) of 28 % (Johnsson et al., 1987). The and safety aspect of this process have been provided
reactor is believed to operate within the turbulent elsewhere (Miseralis et al., 1991; Graham, 1970) and
regime. Besides carbon monoxide, carbon dioxide, are summarized here.
and water, small quantities of maleic anhydride, The original Sherwin Williams first-generation reac-
naphthoquinone, and benzoic acid are also formed as tor used an external shell-and-tube catalyst cooler and
by-products. external ceramic filter elements (Fig. 4a). Reaction heat
The fluidized bed process was commercialized by was removed by circulating catalyst through the exter-
Sherwin Williams in the mid-1940s but is now known nal cooler via a standpipe. The cooled catalyst was
as the Badger phthalic anhydride process. Over 20 returned back to the reactor through the distributor
reactors have been built, with the largest having a plate. This operation was never satisfactory and suf-
capacity of 100,000 metric tons per year. The last reac- fered from both mechanical and operational problems.
tor was built in China in the mid-1980s. The process is Reactor throughput was often limited due to insuffi-
now considered noncompetitive with o-xylene-based cient heat removal capability. In the mid-1950s, Badger
suggested experimenting with internal serpentine cool- ments. Catalyst was purged periodically from the filters
ing coils as a way to increase the heat removal capabil- to prevent buildup of very fine dust (below 10 micron)
ity of the reactor. This experiment proved to be very in the reactor. The fine dust is produced by attrition,
successful. Heat removal was increased so much that and at high enough levels it can affect the fluidity of the
the external cooler was no longer needed. The behavior catalyst. Ceramic filter elements are susceptible to
of the fluidized bed was also markedly improved with cracking from thermal shock and mechanical stress.
the addition of the coils. The second generation reac- They were eventually replaced by filter elements con-
tors were designed with internal cooling coils only. structed of glass-wool-wound steel pipe cores and
Medium to high pressure steam was generated directly finally micrometalic filters with marked improvement
in the coils. in operating reliability. With the improvement in
Entrained catalyst leaving the bed was originally cyclone design and operation, the filters were even-
removed by externally mounted ceramic filter elements. tually replaced by internal multistage cyclones.
The hot reactor offgas is susceptible to afterburn Figure 4b shows a third-generation phthalic anhydride
(Graham and Way, 1962). Long runs of pipes between reactor. Two fluidized beds are stacked one on top of
the reactor and filter housings were used to cool the gas the other. The bottom reactor is used for reaction. The
to about 260 C before the catalyst fines were removed. top reactor is a quench zone. This quench zone is
The recovered dust was blown back to the reactor with necessary to prevent afterburn of the reactor offgas
heated air to prevent thermal shock of the filter ele- after the entrained catalyst is removed by the cyclones.
Key reactor design issues are catalyst addition been increased by as much as 200% by operating in
point, polymer withdrawal point, distributor plate the condensing mode. The condensing technology is
design, aspect ratio, operating velocity, and tempera- offered by Union Carbide, BP, and Exxon. The recycle
ture control. Temperature uniformity is critical as the gas is cooled below its dew point to form some liquid.
reactor operates close to the polymer melting point. The latent heat of the liquid is then used to increase the
Maldistribution can result in polymer sheet formation reactor capacity. Union Carbide and Exxon have
or bed collapse; under severe conditions, there can be joined forces to offer supercondensing operation with
complete solidification of the reactor. as much as 50% liquid in the total feed. BP claims
The upper section of the reactor is expanded to operation with up to circa 40% liquid. In the Union
reduce velocity and minimize polymer entrainment. Carbide and Exxon process, the gas and liquid are fed
In the UNIPOL process, the gas and the small amount together through the distributor plate. In the BP pro-
of entrained polymer are cooled and recycled back to cess, the gas and liquid are separated, and the liquid is
the process. In the BP design, a cyclone is used first to fed above the distributor plate (Newton et al., 1995).
recover the polymer before the gas is cooled and Swing reactor operation, where the operating con-
recycled. ditions are changed to make different polymers, have
Reactor capacity is limited by the system heat been considered theoretically but do not appear to
removal capability. Internal cooling coils cannot be have been used in practice. Dual reactors in series
used, as these will be quickly rendered useless by poly- have been used for impact polypropylene copolymer
mer film formation. Temperature control in the first production (Burdett, 1992).
generation reactors was accomplished by cooling of The polymer product and unreacted catalyst is with-
the recycle gas. Capacity of the newer plants have drawn from the reactor and degassed with nitrogen in
4.6 U-CAN
5 SEMICONDUCTOR SILICON
In 1948, Montedison and Lurgi piloted a fluidized bed Fluidized beds have been used for the roasting of ores
for synthesis of AlF3 from aluminum hydroxide since 1947. Over 550 units have been built to date.
AlðOHÞ3 and hydrofluoric acid by the reaction Both KTI/Dorr–Oliver and Lurgi offer conventional
AlðOHÞ3 þ3HF ! AlF3 þ 3H2 O ð7Þ single or double stage fluidized bed roasters.
Roasting applications include iron pyrite and pyrrho-
The pilot plant produced 28 ton/day of AlF3 with a tite for sulfuric acid, roasts for arsenic removal and
purity of 92 to 94% (Reh, 1971). The process has since roasts of metal sulfides for recovery of zinc, copper,
been commercialized successfully. Moisture-free alumi- cobalt, nickel, gold, tin, and molybdenum. Both slurry
num hydroxide of about 45 micron average size is fed feed and dry feed systems are used. Refer to Kunii and
into the expanded upper section of a two-zone fluidized Levenspiel (1991b) and Bunk (1990) for additional
bed (Fig. 14.) The hydroxide contacts superheated information.
hydrofluoric acid in a zone of increased solid concen- In 1989, Lurgi commercialized a 575 ton/day wet
tration at the bottom of the lower reactor. Hot flue gas feed slurry circulating fluidized bed refractory gold
is introduced immediately below the expanded section roaster operating at about 640 C and near ambient
to provide the endothermic heat of reaction and pressure (Peinemann, 1990). A second unit of identical
increase operating velocity to around 140 cm/s, affect- design was added in 1990. A third unit with dry feed
ing circulation of the solids through the cyclone to the handling and a capacity of about 2000 ton/day gold
containing refractory ore was started in 1991. KTI/ between which catalyst is circulated. Methanol conver-
Dorr–Oliver also has a circulating unit in operation sion is essentially 100%. The reaction produces a small
quantity of coke that will slowly deactivate the cata-
lyst. To maintain activity, catalyst is circulated to the
7 OTHER PROCESSES regenerator where the coke is burnt off with air. The
heat of reaction is removed by steam generation. As
7.1 UOP/Hydro Methanol to Olefins the reactor severity is increased, the ethylene/propylene
ratio and the coke yield increase, while the overall ole-
UOP and Norsk Hydro have jointly developed and fin yield drops (Vora et al., 1996). A 0.5 MT/day
piloted a fluid bed process for the conversion of metha- demonstration unit was operated in the mid-1990s.
nol to olefins (Fig. 16.) The process uses Union UOP has announced the successful completion of the
Carbide’s (now part of UOP) SAPO-34 catalyst. A demonstration unit and the availability of this process
two fluidized bed reactor/regenerator system is used, for licensing.
about 538 C. By recycling, up to 40 weight percent of Chlorinated hydrocarbon liquid streams from the
the total feed can be converted to propylene and iso- oxychlorination process (this chapter), containing
butylene. The reactor product slate can be changed by finely divided carbonaceous solids and metallic corro-
adjusting its operating temperature. sion products in solution and suspension are pumped
into the lower portion of the fluidized bed. Air is used
to fluidize the bed and to oxidize the feed. The reaction
7.3 Catalytic Oxidation of Chlorinated Byproducts occurs at below 540 C. Combustion of the feed is
essentially complete with no significant breakthrough
The Catoxid fluidized bed process was developed by of elemental chlorine and only minor breakthrough of
BF Goodrich (now Geon) in collaboration with chlorinated hydrocarbons (Benson, 1979). The heat of
Badger. The first unit was started in 1974. At least 4 combustion is recovered by generating medium to high
units have been built. pressure steam inside coils placed in the bed. The pro-
cess requires auxiliary fuel only when the feed contains 7.4 Isophthalonitrile
significantly more than 70% chlorine. The auxiliary
fuel is required to sustain the correct bed temperature GB Biosciences (now part of Zenneca) has been oper-
and satisfy the hydrogen–chlorine ratio for hydrogen ating a dual fluidized bed process to produce isophtha-
chloride production. lonitrile by ammoxidation of m-xylene, by the reaction
The combustion gas leaving the Catoxid process
contains hydrogen chloride, water, nitrogen, oxygen, C8 H10 þ 2NH3 þ 3O2 ! C8 H4 N2 þ 6H2 O ð8Þ
and carbon oxides and can be recycled directly
back to the oxychlorination reactor, as shown in The process was commercialized in 1976. A vanadium
Fig. 3. catalyst with Geldard group A/C characteristics is
The Catoxid catalyst is robust and is compatible used. The catalyst is circulated between a reactor oper-
with the oxychlorination reaction. Catalyst makeup is ating at 400 C and a reoxidizer operating at 427 C.
required only for attrition losses. The catoxid catalyst The reactor operating pressure is not reported but is
dust moves through the oxychlorination fluid bed expected to be 1 to 2 atmosphere. Catalyst losses are
reactor and leaves the system with the oxy catalyst kept to a minimum by the use of sintered metal filters
fines. (Fig. 18).
Gabriel I. Tardos
The City College of the City University of New York, New York, U.S.A.
Paul R. Mort
Procter & Gamble, Cincinnati, Ohio, U.S.A.
Figure 2 Schematic representation of a fluid bed granulation circuit. Shown in the circuit is a fluid bed grinder–granulator. 1.
Fluid bed houseing, 2. Fluidized bed, 3. Grinding roller, 4. Drive, 5. Spray head, 6. Binder solution, 7. Pump, 8. Fluidization air
fan, 9. Air heater, 10. Air control valve, 11. Cyclone, 12 and 13. Rotary valves, 14. Filter bag-house. (After Dencs and Ormos,
1993.)
3 MICROSCOPIC PHENOMENA
Eq. (7) is negligibly small and that the granule’s where C is a constant coefficient that incorporates all
internal strength is given by ð Þ ¼ k n . parameters in Eq. (9) except the diameter and shear rate.
Equation (9) can then be rewritten as Taking the logarithm of the above expression yields
n
logðacr Þ ¼ 1 logð Þ þ log C ¼ m logð Þ þ c
2
a2cr 2 ¼ C n ð17Þ ð18Þ
Vessel diameter (cm) 12.5 23 50 Here we assumed that the shear rate, , is propor-
Cross-sectional area (cm2 ) 122.7 415.5 1963.5 tional to the impeller tip speed, Ui , and therefore the
Spray area (cm2 ) 38.5 132.7 594.0 quantity Ui =Uf2 is some modified shear rate. The aver-
Vessel volume ratio 1 6.23 64 age value of the slope for all granulators, large and
Powder feed wt. (kg) 0.36 2.23 22.9
small, is approximately m ¼ 0:25, which gives a rheo-
Spray area (cm2 )/feed (kg) 106.9 59.5 25.9
logical flow index of n ¼ 1:5. Such values are common
Source: Watano et al., 1995. for concentrated slurries of viscous liquids and small
solid particles and characterize a shear thickening
behavior. The most important conclusion from the
This expression shows that on a log–log plot, the slope above considerations is that both the slope, m, and
of the curve of granule diameter versus shear rate gives the coefficient, c, take approximately the same value
the shear index, n, of the binder through the coefficient for all size granulators operated under different condi-
m, while the intercept gives a measure of the critical tions, especially at the high-shear range of the impeller
Stokes number through the value of c. where the power-law term of the viscosity dominates.
The experimental data shown in Fig. 12 include This implies that the binder viscosity has the same role
median granule size as a function of varying tip in both small and large units and that the Stokes num-
speed at three independent gas fluidization velocities ber, based on deformation, is the relevant dimension-
(Uf ¼ 0:4, 0.6, 0.8 m/s) at three machine scales. less parameter. It also implies that the Stokes number
Equation (18) predicts reduced growth at high propel- may be the appropriate scaling factor when such
ler tip speed. This effect, however, is counteracted devices are scaled up.
somewhat by increased gas fluidization velocity that
cushions particles from the impact of the propeller. 7.2.3 Agglomerate Consolidation
Data regression shows that the effect of air fluidization
velocity (Uf ) relative to the propeller tip speed is Experimental support for the expression derived for
approximately to the 2 power. The full set of data granule consolidation [Eq. (13)] is found again in the
in Fig. 12 fit the log–log trend predicted by work of Watano et al. (1995). Their data is shown in
Fig. 13 where the relative density ratio is given as a
function of the impeller tip speed (Ui ) for all three
Figure 14 Experimental data by Dencs et al. (1993) and Dencs (1998); agglomerate size versus modified shear rate and
fluidization velocity. Data for sucrose and zinc sulfate solutions in small and large fluid beds.
Figure 16 Simulated granulation with increasing amounts of binder. Cumulative particle size distribution vs. relative agglom-
erate diameter. Stokes number St ¼ 0:1 and capillary number Ca ¼ 1:0.
This last equation yields [with Eq. (24) and a ¼ 1] the ginally formalized in Ennis et al. (1991) is extended in
largest possible agglomerate that will survive in the the present paper to totally deformable particles where
granulator under given conditions. the binder viscosity is replaced by an equivalent visc-
osity of the malleable surface. Furthermore, the con-
cept of energy dissipation is also applied to determine a
CONCLUSIONS criterion of ‘‘green’’ particle deformation and breakup
and to granule consolidation. A set of dimensionless
A comprehensive model of granulation was presented, numbers that characterize granule growth and the
one that takes an agglomeration process from binder deformation process are defined. It is shown that
selection and testing to granule growth, deformation, these numbers (Stokes numbers) delimit different gran-
and consolidation. An instrument and procedures were ulation regimes, and critical values are found that can
described to measure the critical amount of binder that be used to calculate the critical granule sizes that are
ensures that coalescence and particle growth will characteristic at the transition from one regime to the
occur. It was also shown how the pendular bridge other. Several experimental results from the literature
apparatus (PBA), developed earlier by the present are used to illustrate the usefulness of the theoretical
authors, can be used to test the bridge-forming char- approach. A set of very careful experiments performed
acteristics of the binder while at the same time deter- by the present authors in a specially constructed gran-
mine binder spreading and penetration rates and the ulator with constant shear are used to measure critical
critical time of binder strengthening. These operations Stokes numbers. Employing these values, it is possible
can be performed before any granulation runs are car- to predict the outcome of a granulation process pro-
ried out, and binder characteristics can be modified to vided that the shear in the device is essentially constant
achieve proper wetting, bridge formation characteris- and that binder bridge formation is not superimposed
tics, and binder strengthening (viscosity increase). on granule growth.
A theoretical model of particle/granule growth Finally, it was shown that rapid granular flow simu-
based on simple principles of energy dissipation in lations (RGFS) into which the effect of binder viscosity
the liquid (binder) bridge between two colliding parti- and surface tension are introduced, readily predict the
cles was developed. When the initial energy of the par- outcome of granulation. These simulations support the
ticles is smaller than the dissipated energy, particle simplified theoretical models of granule growth by coa-
coalesce and growth will take place. This concept, ori- lescence and granule deformation and breakup and
Arun S. Mujumdar
National University of Singapore, Singapore
Sakamon Devahastin
King Mongkut’s University of Technology Thonburi, Bangkok, Thailand
High drying rates due to excellent gas–particle con- Table 1 compares the key characteristics of the flui-
tact leading to high heat and mass transfer rates dized bed dryer with those of some of the competing
Smaller flow area dryers. Note that the flash dryer is applicable only for
Higher thermal efficiency, especially if part of the the removal of surface moisture from smaller size par-
thermal energy for drying is supplied by internal ticles, while the fluidized bed is recommended for
heat exchangers particles requiring longer drying times, e.g., 10–60 min-
Lower capital and maintenance costs, compared to utes compared to only 10–30 seconds for flash drying.
rotary dryers Some of the characteristics listed in Table 1 are subject
Ease of control to modification by changes in the design of the hard-
ware as well as the operating conditions. It is impor- in batches) or continuously. Rectangular and other
tant to point out that for removing organic solvents or noncircular cross section vessels may be employed if
handling toxic or flammable solids, the conventional needed, however. For most operations, fluidized beds
hot air fluidized bed dryer is not a good choice, since operate at slight negative pressure to avoid leakage of
there is danger of fire or explosion if flammability lim- hot and humid gas into the ambient; it is generally not
its are exceeded. Also, owing to the inherent need to cost-effective to fabricate fluidized bed dryers as pres-
handle large volumes of air due to the need to achieve sure vessels or totally leak-proof.
fluidization, complete collection of the toxic material is Figure 1 provides a coarse classification scheme that
not economically feasible. covers most (but not all) fluidized bed dryers available
commercially. Some special variants of the conven-
tional fluidized bed may be needed for special applica-
2 CLASSIFICATION AND SELECTION tions. It should be noted at the outset that the well-
CRITERIA mixed dryer type (Fig. 2) has the advantage of being
able to handle higher moisture content feed which may
The simplest and most common fluidized bed dryer is a not be fluidizable under normal conditions. As the wet
circular cross section vessel in which the particulate feed enters a well-mixed fluidized bed of a lower aver-
material is dried convectively either in the batch age moisture content, the danger of defluidization or
mode (for smaller quantities or if the feed is produced formation of large chunks (or agglomerates) is
Note that when removing an organic solvent it is Finally, some remarks on capital costs of fluidized
normally necessary to use an inert gas (e.g., nitrogen); bed dryers are appropriate. Typically, the equipment
the solvent is recovered by condensation and the inert cost rises as the 0.60 to 0.70 power of the capacity.
gas reheated indirectly and returned. A small ‘‘bleed’’ The cost of the ancillary equipment, e.g., feeders,
is necessary to avoid the buildup of impurities in the blowers, gas cleaning equipment, can exceed the
circulating gas stream. When drying solids that can cost of the dryer itself several times, depending on
undergo combustion, care must be taken to avoid con- specific applications. It is therefore important to
ditions susceptible to fire or explosion. Suitably cost the entire drying system rather than just the
designed rupture disks should be installed and the oxy- dryer.
gen level in the drying system (including cyclones and
baghouses) must be controlled to be outside the explo-
sion limits. All internal metal parts must be grounded 3 BASICS OF DRYING KINETICS
to avoid the potential of buildup of electrostatic
charges by the fine dry dust. Figure 5 represents a typical textbook drying rate
Control of fluidized bed dryers is discussed by curve obtained from a batch of wet particles fluidized
Jumah et al. (1995). More recently, Liptak (1998) has with hot air of fixed temperature, humidity, and flow
discussed advanced control strategies for continuous rate. After an initial transient period (generally a
fluidized bed dryers. Attempts have also been reported heatup period) if the particles are ‘‘very wet,’’ i.e., cov-
in the literature on the use of fuzzy logic and artificial ered with a liquid film, the drying rate depends entirely
neural nets to control fluidized bed drying and granu- on the external heat and mass transfer rates, e.g., a rise
lation. in air temperature or air flow rate or a reduction in air
Figure 4 A schematic flow diagram of the whole fluidized bed drying system.
humidity will lead to an increase in the drying rate. follows the adiabatic saturation line (in both constant
This rate continues unchanged as long as the liquid and falling rate periods).
film covers particle surfaces fully, since the drying con- It is useful to summarize the key results of numer-
ditions are held constant. Thus the so-called constant ous experimental studies on fluidized bed drying
rate drying period exists as long as the drying condi- kinetics. The following generalizations apply mainly
tions remain unchanged. Once the particle surfaces to the conventional hot gas fluidized adiabatic dryer.
become partially dried, the rate of drying is reduced
since in the area not covered with liquid film the rate of 3.1 Effect of Bed Height
drying will depend also on the rate at which the inter-
nal liquid arrives at the surface prior to vaporization. For materials with high mobility of internal moisture
The moisture content at which this drop in drying rate (e.g., iron ore, ion exchange resins, silica gel), most
first occurs is called the critical moisture content. drying takes place close to the distributor plate.
Unfortunately, this critical value is not a material Increasing bed height beyond a particular value has
property in that it is also a function of the drying
conditions for a given material. It must be determined
experimentally. Once the surface film disappears, the Table 4 Approximate Critical Moisture Contents for
surface attains its equilibrium moisture content corre- Various Materials
sponding to the temperature and humidity conditions
Critical moisture content
it is exposed to. [See Mujumdar and Menon (1995) for
Material (kg water/kg dry solid)
details of the definitions and terminology in psychro-
metry and drying.] Approximate critical moisture con- Salt crystals, rock salt, 0.05–0.10
tent values for some selected materials are given in sand, wool
Table 4. Brick clay, kaolin, crushed 0.10–0.20
For a purely convective adiabatic dryer, neglecting sand
sensible heat effects, the surface temperature attained Pigments, paper, soil, worsted 0.20–0.40
by the drying particles in the constant rate period is the wool fabric
Several foods, copper 0.40–0.80
so-called wet-bulb temperature corresponding to the
carbonate, sludges
gas conditions they are exposed to. Below the critical Chrome leather, vegetables, > 0:80
moisture content the particle temperature will rise fruits, gelatin, gels
above the wet-bulb temperature. The gas temperature
Hz) allows ‘‘pseudo-fluidization’’ of the bed with To intensify heat and mass transfer rates for rapid
rather low air flow rates. In this case, the requirements drying of surface-wet particles, a centrifuge-type device
of hydrodynamics and heat/mass transfer are effec- may be used so that the drag force due to the fluidizing
tively decoupled. The gas velocity can be chosen to gas can be balanced with an ‘‘artificial gravity’’ gener-
fit the needs of the drying kinetics; it must be signifi- ated by rotating the bed on a vertical axis. The rotating
cantly less than the minimum fluidization velocity, fluidized bed equipment is complex, and the decrease
since above such a velocity the bed is no longer in in drying times for most materials is normally not high
continuous contact with the vibrating distributor enough or essential enough to justify the cost and com-
plate. Furthermore, since the vibrational energy trans- plexity.
mitted is attenuated with distance from the plate, the
bed depth is limited to about 10–15 cm. Vibrated bed 4.7 Spouted Bed Dryers (SBDs)
dryers can also be used to reduce attrition by gentle
processing. A schematic sketch of this dryer is shown Spouted bed dryers are found suitable for drying of
in Fig. 7. Most vibrated fluidized bed dryers are con- Geldart’s type D particles, which are too coarse and
tinuous units; the vibration vector is applied at a small dense to fluidize well without channeling. Unlike flui-
angle to the vertical to assist with conveying of the dized beds where the particle motion is random, the
material from the feed end to the exit weir. movement of particles in spouted beds is a regular
recirculatory motion. Both batch and continuous
4.5 Mechanically Agitated Fluidized Bed Dryers modes of operation are possible. Owing to their limited
processing capability per unit floor area and their high
Several versions of such dryers are in use. For drying power consumption, spouted beds have not yet found
of pastes or sludges, one variant uses a cylindrical ves- major commercial applications. They are more com-
sel with a fast spinning agitator at the bottom onto monly used for roasting (e.g., coffee/cocoa beans,
ble heat required to bring the product temperature ms Cps Tin Tp ðt ¼ 0Þ
t¼ log ð2Þ
from its wet-bulb value (at the end of the surface dry- Gg Cpg Tin Tp ðtÞ
ing period in a purely convective dryer) to the final
temperature that (from laboratory experiments) the In the case when the amount of moisture to be
product must reach at the end of drying to achieve removed is large and the effect of latent heat cannot
the desired moisture content. In such cases, the batch be neglected, different procedures must be employed.
drying time in the final drying period can be estimated Kunii and Levenspiel (1991) have presented simplified
reasonably accurately by a simple transient energy bal- procedures for analysis and design of batch and con-
ance to determine the time required to heat up the tinuous fluidized bed dryers. As expected, different cal-
particles from the wet-bulb to the final desired tem- culation procedures are required to calculate the
perature. In most cases, it is also necessary to cool drying time for the constant rate and falling rate per-
the product down to a lower temperature for storage iods. When the moisture diffusivity is high, the particle
or packaging. Indeed, the same design equation can be size small, and the drying conditions mild, much of the
used to estimate the cool down time. The governing drying will occur in the constant rate period. In prac-
equation for the energy balance is (van’t Land, 1991) tice, fluidized bed dryers are used when a significant
dTp Gg Cpg ½Tin Tp ðtÞ portion of the moisture is internal, in which case a
¼ ð1Þ simple diffusion model can be used to estimate the
dt ms Cps
drying time for the falling rate period. Interested read-
Upon integration, this yields the heat-up time, which is ers are referred to the design procedures and the exam-
the dryer residence time needed for the falling rate ple design problems presented by Kunii and Levenspiel
period drying: (1991).
Figure 8 Decision tree for selection of fluidized bed dryers for granular feedstock. Note: Other dryer types are also suited for
several of the above applications. If the feedstock contains a solvent, fluidized bed dryers are not recommended in general;
indirect or vacuum dryers are preferred instead.
cal literature on modeling of the heat and mass transfer 6 NOVEL FLUIDIZED BED DRYERS
processes in fluidized bed drying of particles as well as
fluidized bed drying of slurries using beds of inert par- Numerous innovative designs and operational modifi-
ticles. Borde et al. (1997) have presented a simple cal- cations have been proposed in the literature while
culation procedure for the design of a vibrated bed fewer are readily available commercially.
dryer. The review presented by Pakowski et al. (1984) Tables 7, 8, and 9 summarize the key features of the
is also very useful to obtain a clear idea about the flow ‘‘innovative’’ dryer concepts relative to the more con-
and heat/mass transfer characteristics of vibrated flui- ventional FBDs. Many of the ideas presented are new
dized beds. For the design and analysis of conventional and have not been investigated even at the laboratory
spouted beds as well as certain modified spouted beds stage. It must be noted that not all the novel ideas are
the reader is referred to the work of Passos et al. necessarily better, but many are worth exploring, since
(1987). For the special case of intermittently spouted they appear to offer some distinct benefits over the
beds for grain drying, Jumah et al. (1996) present a more conventional ones.
simple diffusion-based model for the estimation of dry-
ing times. Their model may be used only if the entire 6.1 Modified Fluidized Bed Dryers
drying occurs in the falling rate period and the diffu-
sion model adequately describes the falling rate period. Depending on special requirements it is possible to
Unfortunately, no degree of sophistication in mod- develop modified versions of the traditional fluidized
eling can predict product quality parameters. Thus bed dryers. Two-stage fluidized bed dryers consisting
pilot testing is essential prior to scale-up to full-scale of several dryers are now in commercial use. Some
dryer design. examples of these combinations are given in Table 10.
Conventional Innovative
Conventional Innovative
Pulsing the fluidizing gas flow has been demon- along the length of the dryer. The final section may
strated to enhance drying performance in some cases generally act as a fluid bed cooler to reduce the product
at the expense of added capital costs. Periodic local temperature to a safe value for storage or packaging
fluidization of the bed by moving the distributor (Bahu, 1997). If the product is highly hygroscopic, and
zone at regular intervals (so-called pulsated fluidized low final moisture contents are desired, the cooling
beds) has the advantage of reduced energy and air stage may require dehumidified air.
consumption when batch drying particulate materials In the following we will review only a few innovative
with most of the heat and mass transfer resistances FBD techniques. For further details, see Kudra and
within the particles (Gawrzynski and Glaser, 1996). Mujumdar (1995).
For larger particles (e.g., Geldart’s type D), intermit-
tent spouting of the bed with a rotating spouting jet
has been shown by Jumah et al. (1996) to reduce 6.2 Superheated Steam Drying in Fluidized Beds
energy consumption at only a marginal increase in dry-
ing time for batch drying. Indeed, in this case, it is also Although the idea was first published as early as 1898,
possible to introduce another type of intermittency, the potential for superheated steam as a drying med-
e.g., periodic heating of the spouting air when drying ium was not exploited industrially for at least six dec-
in the batch mode. ades thereafter. Superheated steam drying offers one or
To conserve energy while reducing the drying time more of the following advantages (Mujumdar, 1995):
in batch fluidized bed drying one may remove the sur-
face moisture using a higher gas temperature initially; No fire or explosion hazards
the gas temperature can be reduced progressively as the No oxidative damage
material dries so that it does not exceed its maximum Ability to operate at vacuum or high-pressure oper-
permissible temperature (Devahastin and Mujumdar, ating conditions
1999). This can be achieved with a fuzzy logic control- Ease of recovery of latent heat supplied for evapora-
ler or with a model-based control scheme. In plug-flow tion; effectively a multiple effect operation is fea-
dryers this idea can be extended by temperature-zoning sible
Conventional Innovative
Type Remarks
Similar FBD stages stacked one below the other Reduces floor area/bed depth of each stage; low product
moisture content possible
Flash dryer stage preceding FBD stage Fast removal of surface moisture; reduced stickiness leads
to easy fluidization/smaller fluidized bed dryer
Spray dryer stage followed by FBD stage Significantly reduced spray dryer size
Well-mixed FBD followed by plug-flow FBD Ease of fluidization for high moisture content feed; uniform/
stage underneath in same vessel low final product moisture content possible
A higher drying rate above a critical ‘‘inversion tem- be used to lower the operating temperature, thus redu-
perature’’ when removing surface moisture cing the probability of thermal degradation; it may
Better quality product under certain conditions allow operation in conditions out of the flammability
Closed system operation to minimize air pollution limits. Arnaldos et al. (1997) have presented experi-
mental results on the hydrodynamic behavior of
Offsetting some of these advantages are some limita-
vacuum fluidized bed dryers and their applications
tions, e.g.,
for drying granular solids and coating of particles.
Higher product temperature Note that the minimum fluidization velocity increases
Higher capital costs compared to hot air drying as the pressure decreases. The flow behavior at lower
Possibility of air infiltration making heat recovery pressures is quite different from that at atmospheric or
from exhaust steam difficult by compression or high-pressure operation. The fluidization is progres-
condensation sive, and the concept of minimum fluidization velocity
loses some of its significance. Experimental measure-
Among industrial scale applications of superheated ments of the velocity for complete fluidization are
steam fluidized bed drying one may cite needed for design purposes.
Drying of pulverized coal (Faber et al., 1986) Little work is published on vacuum fluidized bed
Drying of high-moisture lignite using a fluidized bed drying. Arnaldos et al. (1997) report vacuum fluidized
with internal heat exchangers to improve energy bed drying data at 200, 400, and 500 mbar pressures.
efficiency and reduce dryer size (Potter et al., Silica ðdp ¼ 975 m, s ¼ 1650 kg=m3 ; s ¼ 0:7) and
1986) millet (dp ¼ 1800 m, s ¼ 1600 kg=m3 ; s ¼ 0:90)
Drying of pulps, bagasse, sludges, hog fuel, spent were used as test materials. Interestingly, the highly
grain from breweries, etc., in a pressurized super- porous silica particles dried faster as the pressure was
heated steam dryer (3–5 bars) manufactured by lowered, while the denser millet particles dried slower
Niro A/S, Denmark (Jensen, 1992) since much of its resistance to drying is internal. Thus
the effect of lower pressure depends on the internal
As may be expected, the net energy consumption of heat and moisture resistance of the particles. Because
such dryers can be very small. Up to 90 percent fuel of their inherently higher capital and operating costs,
saving over conventional dryers has been reported for vacuum fluidized bed dryers have found application
drying of beet pulp in a high-pressure fluidized bed mainly in the pharmaceutical industry.
dryer (Bosse and Valentin, 1988).
Owing to the high potential of thermal degradation Fluidized bed dryers, both conventional and innova-
and fire or explosion hazards if the wet solids contain tive, will continue to find increasing applications in
an organic solvent—common occurrences in the fine various industries. For further details, the reader is
chemical and pharmaceutical industries—low-pressure asked to refer to Vanecek et al. (1965), Gupta and
operation offers several advantages (Mujumdar, 1995). Mujumdar (1983), Pakowski et al. (1984), Reay and
Similar problems may arise in coating operations. To Baker (1985), Hovmand (1995), Keey (1992), and
alleviate this problem, low-pressure fluidization may Mujumdar (1995), which contains an extensive anno-
2 APPLICATIONS
Figure 1 Schematic of typical circulating fluidized bed Major commercial applications of circulating fluidized
system. beds are listed in Table 1, together with key references.
These applications include solid-catalyzed gas reac-
tions and gas–solid reactions, as well as physical opera-
solid contactors such as packed beds and rotary kilns tions. While catalytic cracking and solid-fuel
frequently include combustion are the predominant catalytic and gas–
solid reaction applications, respectively, there are also
High gas throughputs
a number of other processes where the unique charac-
Limited backmixing of gas
teristics of circulating fluidized beds are being
Long and controllable residence time of particles
exploited. Further details on the most important appli-
Temperature uniformity, without ‘‘hot spots’’
cations are presented in Chapters 14 through 16.
Flexibility in handling particles of widely differing
sizes, densities, and shapes
Effective contacting between gas and particles 3 FLOW REGIMES
Lack of bypassing of gas with minimal mass trans-
fer limitations 3.1 Onset of Fast Fluidization
Opportunity for separate and complementary
operation (e.g., catalyst regeneration or particle A circulating fluidized bed (CFB) is operated in the
cooling) in the return loop transport mode, with solids carried over from the top
Disadvantages often include: of the riser separated and returned to the bottom of the
riser via a standpipe and feeding or control device. The
Need for very tall vessel: small scale CFB processes transition from low-velocity fluidization to transport
are therefore seldom viable operation occurs when significant solids entrainment
Substantial backmixing of solid particles commences with increasing superficial gas velocity.
Internals (e.g., baffles, heat transfer surfaces) not At least seven methods have been proposed (Bi et al.,
viable because of wear/attrition 2000) to quantify the transition. The criteria can be
Wall wastage sometimes a serious problem divided into two groups, one based on solids entrain-
Suspension-to-surface heat transfer less favorable ment and the other on solids concentration profiles.
than for low-velocity fluidization
Lateral gradients can be considerable 3.1.1 Transport Velocity, Utr , based on Phase
Losses of particles due to entrainment. Diagrams
Extensive reviews of all aspects of circulating flui- A critical solid circulation rate may exist when a sharp
dized beds up to about 1996 are provided by the books change in pressure drop over the lower part of a riser
edited by Kwauk (1994) and Grace et al. (1997). More takes place when varying the solid circulation flux at a
Fluid catalytic cracking Avidan, 1997; Sec. 8.1 of this handbook Hundreds of units worldwide; mainstay of
petroleum refining
Fischer–Tropsch synthesis Shingles & McDonald, 1988; Steynberg Applied for many years as Synthol process in
et al., 1991; Matsen, 1997 South Africa
Maleic anhydride Matsen, 1997; Contractor, 1999 One commercial reactor in Spain
Combustion of coal, Li and Zhang, 1994; Brereton, 1997; Widespread usage for power generation and
biomass, wastes, off-gases Lee, 1997; Basu, 1999; Plass, 2001 boilers in Europe, North America, and Asia
Gasification Hirschfelder & Vierrath, 1999; Plass, 2001 Commercial units gaining a foothold, especially
in Europe
Calcination (e.g., of Reh, 1971, 1986; Schmidt, 1999 Lurgi units used widely
aluminium trihydrate
and carbonates)
Catalyst regeneration Chen et al., 1994 Applied in China
Roasting of ores Dry & Beeby, 1997; Pienermann et al., Applied in Australia
1992
Reduction of iron ore Dry & Beeby, 1997; Husain et al., 1999; Lurgi plant in Trinidad
Plass, 2001
Smelter off-gas treatment Hiltunen & Moyöhänen, 1992 One plant in Australia supplied by Ahlstrom
Flue gas dry scrubbing of Graf, 1999; Mayer-Schwinning & Commercial units since the 1970s, primarily in
HF, HCl, SO2 , dioxins, Herden, 1999 Europe
mercury, etc.
given gas velocity, Ug (see Fig. 2). As U increases Below this velocity, a distinct interface exists between
beyond a certain point, the sharp change in the pressure the top-dilute and bottom-dense regions. Beyond this
gradient disappears. The gas velocity at this critical velocity, the interface becomes relatively diffuse.
point, defined as the transport velocity Utr , marks the To predict Utr , several correlations have been devel-
onset of fast fluidization (Yerushalmi and Cankurt, oped, most of the form
1979). An examination of pressure gradient profiles
Retr ¼ kArn ð1Þ
(Bi, 2002) reveals that Utr varies with height. Utr may
indicate a transition of axial voidage profiles in the riser. Typical k and n values are given in Bi et al. (2000), with
k ¼ 2:28 and n ¼ 0:419 from Bi and Fan (1992). In
view of the difficulty of determining Utr experimen-
tally, and the variation of Utr with measurement loca-
tion and riser geometry, caution needs to be exercised
when Eq. (1) is used to predict Utr .
Figure 4 Transitions between dense suspension upflow, fast Figure 5 Flow regime diagram based on pressure gradient
fluidization, and pneumatic transport with increasing solids and superficial gas velocity coordinates. (Adapted from Zenz,
circulation flux at a constant gas velocity. 1949.)
4.1.1 Definitions and Measurement Techniques Figure 6 Flow regime diagram based on solids circulation
rate and superficial gas velocity. (Adapted from Bi and
The voidage is the fraction of volume occupied by the Grace, 1996.)
gas in a gas–solid suspension. Voidage is a key variable
in circulating fluidized bed risers. Several different voi-
dages are in common usage: In some cases we are interested in the fluctuations
themselves, as well as in the time-averaged quantities;
Instantaneous local voidage in a measuring volume
we can then write, for example,
that is small with respect to the equipment but
large with respect to individual particles. This is ðH
1
designated ", or "ðx; y; z; tÞ to indicate variation ½" ¼ "av ðzÞ ð10Þ
H 0
with position and time.
Local time-average voidage ", "ðr; zÞ, or "ðx; y; zÞ. "ðr; zÞ ¼ "ðr; zÞ þ " 0 ðr; zÞ ð11Þ
Cross-sectional average voidage at a given height,
denoted here by "av or "av ðzÞ. where the prime denotes a fluctuating component, i.e.,
Overall average voidage for the riser, ["], i.e., frac- the instantaneous deviation from the time mean. It is
tion of entire riser volume occupied by gas. also common to report or discuss the volumetric con-
centration of solid particles, often called the ‘‘solids
The overbar signifies a time average, whereas the sub- holdup.’’ This terminology is used throughout this
script ‘‘av’’ and square brackets denote spatial chapter and denoted by "s , i.e., "s ¼ 1 ". Still another
averages. Since we are mostly interested in steady way of designating the concentration is to refer to a
flows, time-averaging is often also carried out. For ‘‘suspension density,’’ i.e., susp ¼ p ð1 "Þ þ g ".
steady axisymmetric flow in a cylindrical riser of radius Clearly one can obtain local time-averages and spa-
R and overall height H, tial-averages values of "s and susp in an analogous
ð manner to obtaining these quantities for ".
1 t
"ðr; zÞ ¼ lim "ðr; zÞ dt ð8Þ Several techniques have been found to be useful for
t!1 t 0
determining local voidages/solids holdups:
ð
1 R Optical Probes for the measurement of voidages and
"av ðzÞ ¼ "ðr; zÞ2r dr ð9Þ
R2 0 solids holdups in CFB risers were pioneered by Qin
Exit
Author Particles dp , m U, m/s Gs , kg=m2 s D, m H, m geometry Other effects studied
Li & Kwauk, 1980 iron, alumina, 54–105 0.8–5.6 16–135 0.09 8 smooth
FCC, pyrite
Rhodes & Geldart, 1986 alumina 64 4 45–115 0.152 6 abrupt
Bader et al., 1988 catalyst 76 4.3, 9.1 147 0.305 12.2 smooth
Mori et al., 1992 FCC 54 1.0–1.5 2.4–29 0.05–0.1 1.5–5.5 various column geometry
Bai et al., 1994 FCC, sand, 70, 321, 1.5–3.0 7–128 0.097, 3.0 smooth binary mixtures
alumina 633 0.15
Brereton & Grace, 1994 sand 148 3.7–9.2 9–89 0.152 9.3 various exit geometry, secondary air
Karri & Knowlton, 1997 catalyst 76 3.0–8.1 98–586 0.3 13 smooth pressure
Mastellone & Arena, 1999a glass, sand, 67–310 3–6 35–144 0.12 5.75 smooth particle properties
FCC
Issangya et al., 1999 FCC 70 4–8 18–425 0.076 6.1 smooth
solids holdup is highest near the bottom of the riser column with a constricted exit, the solids holdup
and decreases monotonically with height right to the increases throughout, with a particularly large buildup
top when there is no constriction or sharp bend at the occurring near the top exit. The influence on the shape
exit. The holdup is seen to increase at all heights with of voidage profiles of a wide range of variables, includ-
increasing Gs at constant Ug . Except at very low Gs or
very high Ug , where particles tend to be carried right
through the riser in pneumatic conveying, the curves
tend to be sigmoidal in shape, approaching an asymp-
totic value at high heights. This shape, first shown by
Li and Kwauk (1980), was once considered to be uni-
versal. Many risers, however, have constricted top exits
that affect the solids concentration profiles signifi-
cantly. An abrupt or constricted exit acts as an internal
separator, with the particles in the CFB riser being
large and heavy enough that they do not readily follow
the gas as its path varies in direction while negotiating
the exit. A significant fraction of the particles arriving
at the top (typically 20–80% depending on the exit
shape, Ug , and Gs ) are therefore internally separated
from the gas and sent back down the riser, primarily
along the outer walls. With constricted exits, the con-
gregation of particles at the top causes an increase in
particle concentration near the exit (e.g., Jin et al., Figure 8 Axial profiles of solids holdup in riser of diameter
1988; Brereton and Grace, 1994; Pugsley et al., 1997). 152 mm and height 9.3 m with a constricted exit.
The influence of a constricted exit is shown schemati- dp ¼ 148 m, p ¼ 2650 kg=m3 , air as fluidizing gas.
cally in Fig. 8. Note that as Ug and Gs increase in this (Brereton, 1987.)
Table 3 Representative Data Showing Radial Profiles of Local Voidage in CFB Units
to the superficial gas velocity, while the magnitudes of that are ascending is close to 100% toward the axis of
downward velocities are significantly lower. the column, while near the wall, typically 10–40% of
There are substantial velocity fluctuations, espe- particles travel upward, with the rest descending (Zhou
cially near the wall, with the time mean fluctuating et al., 1995a). For rectangular columns, the magnitude
component typically reaching a maximum near the of the downwards mean particle velocity and the frac-
edge of the wall layer (Yang et al., 1992; Caloz et al., tion of particles that are descending when averaged
1999). Flows near the wall undergo reversals, i.e., the over time are greater in the corners than elsewhere
particle velocity there is sometimes upward and some- along the outer wall.
times downward. The fraction of individual particles Zhou et al. (1995b) inferred horizontal components
of velocity for particles crossing the axis of a riser of
square cross section. The measured lateral components
increased to a maximum of 40% of the superficial
gas velocity, Ug , in the middle of the riser and then
decreased toward the top.
Couturier et al., 1991 3:15 3:15 23:8 Pressure profile, solids flux profile
Leckner et al., 1991 1:7 1:4 13:5 Solids flux near wall
Werther, 1994 various Pressure profiles, solids flux, wall layer thickness
Zhang et al., 1995 1:7 1:4 13:5 Particle flux profiles, wall layer thickness
Lafranechère & Jestin, 1995 9.65 eq. dia: 36:9 Pressure profiles, solids flux profiles, wall layer
thickness
Hage & Werther, 1997 5:1 5:1 28:0 Local particle concentration, local particle velocity
Zhang et al., 1997 1:7 1:4 13:5 Particle momentum, concentration and velocity
Caloz et al., 1999 14:7 11:5 35 Particle velocity, fluctuations
Lin et al., 1999 4 6:2 22:8 Recirculation rate, particle velocities
Schouten et al., 1999 1:2 0:8 9:0 Bottom bed hydrodynamics
Sternéus et al., 1999 1:7 1:4 13:5 Turbulence
Johnsson & Johnsson, 2001 1:7 1:4 13:5 Local particle concentration, probability density,
intermittency index
Johnsson et al., 2002 21:1 9:9 42:5 Local particle momentum, local particle concentration
the power of 0.6 (Werther, 1994), not to the For Rep ¼ g dp Ug =g < 4, the above list may be
power of 1. reduced to three groups (Glicksman et al., 1993; van
Large-scale risers, at least those used for combus- der Meer et al., 1999), the Froude number (Frd) and
tion processes, commonly operate under very dimensionless solids flux Gs /(p Ug ) from the above list,
dilute conditions at the top (e.g., see Werther, plus either Ug =Umf or Ug =ut , e.g.,
1994 and Sec. 5.1 below).
Ug Gs Ug
Turbulence is likely to play a lesser overall role as D Frd ¼ pffiffiffiffiffiffiffi ðRep < 4Þ ð32Þ
increases (Senior and Grace, 1998). gdp p Ug ut
Because of experimental limitations, it is uncommon in For Rep ¼ g dp Ug =g > 1000, Rep may be dropped
experimental CFB studies to vary the riser diameter, or from those listed in Eq. (31) above, leaving a subset
even the H=D ratio, while keeping all other variables of four groups to match. D=dp can be replaced in this
fixed. four-group set by Ug =Umf or Ug =ut (Horio et al., 1989;
There have been a number of efforts to introduce a Glicksman et al., 1993; van der Meer et al., 1999),
sufficient set of dimensionless groups which, when yielding, for example,
matched and coupled with geometric similarity, assure Ug Gs p
dynamic similarity between small- and large-scale Frd ¼ pffiffiffiffiffiffiffi
gdp p U g g
units. This work is well summarized by van der Meer ð33Þ
et al. (1999). For full hydrodynamic scaling, they Ug
ðRep > 1000Þ
recommend that at least five dimensionless groups be ut
matched:
A number of other possible variables are ignored in
Ug g dp Ug Gs arriving at the above lists, the most notable being the
Frd ¼ pffiffiffiffiffiffiffi Rep ¼ particle shape, the particle size distribution, the coeffi-
gdp g p Ug
ð31Þ cients of restitution and friction for the particles them-
p D selves and between the particles and the wall, and any
g dp property variations due to temperature or concentra-
tion gradients. Advanced fluid dynamic models (e.g.,
with dp being the Sauter mean particle diameter. In Sinclair and Jackson, 1989; Senior and Grace, 1998)
place of the third group above, Kehlenbeck et al. and experimental results (Chang and Louge, 1992) sug-
(2001) suggest a dimensionless mass turnover, invol- gest that interparticle collisions and other interactions
ving the solids inventory. are of considerable importance for the conditions
6 MIXING
toward the outer wall, partially offsetting the upward
flux there. Typically the local net flux is 1.5–2.5Gs at 6.1 Gas Mixing
the axis of the column (Issangya et al., 1998; Karri and
Knowlton, 1999; Liu, 2001) falling to nearly 0 at the Vertical and horizontal gas mixing are of considerable
wall. Results from a column of diameter 0.3 m at high importance in predicting conversions and selectivities
Ug in Fig. 21 indicate that the radial net flux profiles in CFB reactors, especially for catalytic reactions. Gas
can reverse, showing a minimum at the axis of the mixing depends on the flow regime, flow pattern, par-
column. ticle properties, and riser configuration.
Figure 20 Local time mean particle velocity and standard deviation of particle velocity at z ¼ 4:2 m in 76 mm diameter, 6.4 m
diameter column operated at Ug ¼ 6 m/s, Gs ¼ 417 kg/m2 s. (Liu, 2001.)
With the whole riser treated as a closed vessel with the RTD of the entire riser, tracer is introduced below
well-mixed tracer injection at the inlet and sampling at the gas distributor, with samples taken at the riser exit
the outlet, the RTD curve can be obtained from a pulse to ensure well-mixed injection and sampling. Transient
or a step-change test (Schügerl, 1967; Brereton et al., one-dimensional dispersion can be described by
1988; Liu et al., 1999). An effective dispersion coeffi-
cient, Dge , can then be derived by fitting. Both vertical @C @C @2 C
þ Ug Dge 2 ¼ 0 ð34Þ
and horizontal mixing/dispersion of gases contribute @t @z @z
to gas dispersion. In steady state tracer experiments, With closed–closed boundary conditions, Eq. (34) can
a steady flow of tracer gas is introduced into the riser at be solved. With relatively low dispersion, the variance
a point, and the tracer concentration is measured either and the effective dispersion coefficient are related
downstream or upstream. Ideally, the injection velocity (Levenspiel, 1998) by
should equal the gas velocity in the riser (Bader et al.,
1988). Both adsorbing and nonadsorbing tracers can
2t Dge "av Dge "av 2 Ug H
be used. When an adsorbing tracer is selected, the equi- ¼2 2 1 exp
2 Ug H Ug H Dge "av
librium adsorption on the particle surface needs to be
ð35Þ
identified (Kunii and Levenspiel, 1991). Based on the
tracer concentration measured upstream of the injec- The effective dispersion coefficient Dge was reported
tion point, a one-dimensional backmixing coefficient, to be in the range 1 to 12.7 m2 /s for FCC (Dry and
Dgb , can be derived (Kunii and Levenspiel, 1991). The White, 1989; Liu et al., 1999) and (group B) sand
radial dispersion coefficient, Dgr , is obtained by analyz- (Brereton et al., 1988; Wang et al., 1992). Dge increases
ing radial concentration profiles measured downstream with increasing Gs at relatively low solids circulation
of the injection point (Bader et al., 1988). To charac- rates, then tends to level off (Dry and White, 1989) or
terize the mixing in the riser fully, all three coefficients to decrease after reaching a maximum at
need to be studied. Gs 200 kg=m2 s (Liu et al., 1999). Liu et al. (1999)
related the decrease of Dge for Gs > 200 kg/m2 s to
the transition from fast fluidization to dense suspen-
6.1.1 Gas RTD and Effective Gas Dispersion
sion upflow (see Sec. 5).
The overall mixing behavior of the riser can be char- Figure 22 shows that the Peclet number, Pege , based
acterized by the residence time distribution (RTD) on the effective dispersion coefficient, decreases with
curve or an effective dispersion coefficient. To obtain increasing solids flux, especially for Gs < 100 kg/m2 s.
van Zoonen (1962) Pegr ¼ 500 1000 For small columns only.
Koenigsdorff and Pegr ¼ 150 þ 56; 000ð1 "av Þ For dilute flow only with ð1 "av Þ < 0:01.
Werther (1995)
1:3
Gayan et al. (1997) ðGs þ g Ug ÞD 1:14 dp That Dgr decreases with increasing Gs at
Dgr ¼ 1:4 106 ðGs þ 1Þ D1:85 high solids flux is inconsistent with
g p
experimental data.
0:96 3:73 Ug Gs
Namkung and Udp dp Ug Usl ¼
Kim (2000) ¼ 153:1 "av p ð1 "av Þ
Dgr D Usl
The axial dispersion coefficient is obtained by plotting In all reported pulse tests, the tracer was injected on
lnðC=Ci ) vs. the distance, z, from the source. the axis, with the sampling port downstream on the
Experimentally it is difficult to generate a plane axis. The tracer concentration then reflects the local
source of tracer gas. A point source of injection has response on the axis, not the response from the entire
thus been used in all experiments. When tracer is cross section. The dispersion coefficient, Dgd , obtained
injected at the axis, little tracer is detected upstream in this way is the same as the one-dimensional disper-
(Li and Weinstein, 1989; Namkung and Kim, 1998). sion coefficient defined by equation (37) only for risers
On the other hand, high tracer concentration and sig- of small diameter because of the strong radial variation
nificant radial variation were detected when tracer was of tracer concentrations downstream of the injection
injected into the annulus. These findings suggest that point.
gas backmixing is significant in the annulus region but Dgd generally increases with increasing Gs at given
negligible in the dilute core region. gas velocities for Gs < 200 kg/m2 s. For Gs > 200 kg/
To determine a one-dimensional average back- m2 s, van Zoonen (1962) reported a decrease of Dgd
mixing coefficient based on equation (41), an average with increasing solids flux. van Zoonen (1962) and
tracer concentration is required. With axial symmetry, Luo and Yang (1990) found that Dgd increases with
ðR increasing Ug at constant Gs for Gs > 50 kg/m2 s. Li
2 and Wu (1991) and Bai et al. (1992) reported the oppo-
C¼ 2 CðrÞ
ug ðrÞ
r
dr ð42Þ
R ðUg ="av Þ 0 site trend for Gs < 50 kg/m2 s.
Luo and Yang (1990) found that Dgd decreases with
Dgb generally increases with increasing Gs at given
increasing particle size at given Ug and Gs . The effect of
Ug . Li and Weinstein (1989) found that Dgb increased
column diameter is unclear. Li and Wu (1991) corre-
with increasing Ug at constant Gs , while Namkung and
lated their data by
Kim (1998) found the opposite trend, possibly because
their data were obtained at relatively low fluxes Dgd ¼ 0:195"4:11
av ð46Þ
(Gs < 40 kg=m2 s), while Li and Weinstein (1989)
reached relatively high solids fluxes (up to 200 kg/m2 s). which ignores any effect of particle properties. On the
other hand, Luo and Yang (1990) correlated their Dgd
6.1.4 Axial Gas Dispersion data as functions of operating parameters and gas and
particle properties as
Axial gas dispersion can be studied with pulse or step-
change injection (van Zoonen, 1962; Li and Wu, 1991; Dgd g g Ug dp 1:6 Gs 0:62 p 1:2
¼ 0:11 ð47Þ
Luo and Yang, 1990; Martin et al., 1992; Bai et al., g g p ut g
1992b). Transient one-dimensional dispersion can be
described (Li and Wu, 1991) by
6.1.5 Interphase Mass Transfer
2
@C @C @C
þ Ug Dgd 2 ¼ 0 ð43Þ Brereton et al. (1988) and Namkung and Kim (1998)
@t @z @z considered a core–annulus structure with exchange
For pulse injection with open-open boundary condi- between the two regions. With one-dimensional flow
tions, the solution of equation (43) is assumed in both regions, axial dispersion in the annu-
" # lus and plug flow in the core, one can write
1 ð1 Þ2
C ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi exp @Cc @C 2K
2 ð"av Dgd =Ug L 4ð"av Dgd =Ug LÞ þ ugc c þ ca ðCc Ca Þ ¼ 0 ð48Þ
@t @z Rc
ð44Þ
@Ca @2 C 2K R
where L is the distance between the injection and the Dgza 2a 2 ca c2 ðCc Ca Þ ¼ 0 ð49Þ
@t @z R Rc
sampling point for single point sampling or the dis-
tance between the two sampling locations for double- for the core and annulus regions, respectively. Both the
point sampling. The variance of the RTD curve is axial dispersion coefficient in the annulus region, Dgza ,
with
t vp L
¼ and Pepz ¼ ð52Þ
Dpz Figure 24 Effect of solids circulation flux on axial solids
dispersion. Data: & sand, dp ¼ 90 m (Smolder and
where is the mean residence time, EðÞ is the dimen- Baeyens, 2000b), ~ (Diguet, 1996); ^ FCC, dp ¼ 54 m
sionless RTD function, and L is the distance between (Du et al., 1999), * FCC, dp ¼ 56 m (van Zoonen, 1962),
the tracer injection and detection points. The mean ~ sand, dp ¼ 275 m (Patience et al., 1991).
Figure 27 Heat transfer coefficient vs. suspension density for large-scale CFB risers. For operating conditions see Table 7.
wall conduction resistances, has been developed by Xie Vertical Length of Heat Transfer Surface: Short
et al. (2003). heat transfer surfaces lead to overestimates of heat
transfer to the wall (Wu et al., 1991; Nag and Moral,
Descending Particles: To investigate the effect of 1990) because clusters travel along the heat transfer
the descending wall layer of particles on the heat surface over considerable distances, causing the heat
transfer coefficient at the membrane water wall, Gol- transfer coefficient to decrease as temperature equili-
riz (1994) installed an angled deflector 0.2 m into the bration occurs between the cluster and the surface.
combustion chamber of a 12 MWth combustor 0.5, A number of researchers have measured descending
1.0, 1.5, and 2.0 m above a heat flux meter located cluster velocities in the wall layer for a variety of flow
4.0 m above the distributor. The heat transfer coeffi- conditions and different wall geometries, with most
cient increased and the character of the signal chan- data between 0.5 and 2.0 m/s. The data have been
ged with the obstacle diverting clusters and strands well summarized by Griffith and Louge (1998) and
away from the membrane wall, exposing the surface are correlated by equation (22).
to hotter particles. The measured heat transfer coeffi- The contact length for clusters, Lc , can be 1.5 m or
cient increased by 55% with the deflector 0.5 m more along the fins of membrane water walls (Wu et
above the heat flux meter. The enhancement of the al., 1989; Golriz and Leckner, 1992). Lc is much
heat transfer decreased as the distance between the shorter for smooth walls (Wu et al., 1991; Burki et
meter and deflector increased, disappearing at a se- al., 1993; Glicksman, 1997). The experimental results
paration of 2 m. At 10 m above the distributor plate suggest that the cluster residence time at the wall is
where the suspension was more dilute, heat transfer 0:1 to 1 s for smooth walls and 0:5 to 3 s for
was unaffected by a deflector 1.0 m above the heat sheltered regions, including the fins of membrane
flux meter. Hyre and Glicksman (1995) found similar walls. However, for short heat transfer surfaces, e.g.,
augmentation by deflecting descending particles from 0.01 m long, Lc is only of order 0.01 s. Given the
a plane wall. Cao et al. (1994) increased the heat influence of length of heat transfer surface, consider-
transfer by 100% by placing a large horizontal ring able caution is required when applying data from short
in a CFB. surfaces to much longer surfaces (Grace, 1990, 1996).
hov ¼ 580:36
sus 2 kg=m3
*
For more details see Golriz and Grace (2002).
Figure 32 Predicted heat transfer vs. experimental data. Dashed lines indicate 25% deviations.
Figure 33 Nonuniform radial profiles of concentration determined experimentally in CFB reactors: (a) concentration of ozone
undergoing decomposition in 0:25 m dia 10:5 m tall riser (Ouyang et al., 1995a); (b) concentration of NOx in
0:15 0:15 m 7:3 m CFB pilot scale combustor (Brereton et al., 1995).
Table 9 Examples of Reactions Considered in CFB Reactor Studies and Dimensions of Columns Used in the
Experimental Studies
M ¼ model,
Reaction E ¼ Expt’l References Riser dimensions
A. Catalytic Reactions
Claus reaction M Puchyr et al., 1996, 1997 NA
Fischer–Tropsch process E Shingles & McDonald, 1988 0.1 m id 13 m, then scaled up by
factors of 500 & 2.5
Fluid catalytic cracking M, E Fligner et al., 1994 0.3 m id 12 m
M Gao et al., 1999 NA
Methanol-to-olefins M Schoenfelder et al., 1994 NA
M Gupta et al., 1999 NA
Maleic anhydride from n-butane M Pugsley et al., 1992 NA
E, M Golbig and Werther, 1996 21 mm i.d.
Oxidative coupling of CH4 M Pugsley et al., 1996 NA
Ozone decomposition E Jiang et al., 1991 0.10 m id 6:3 m
M, E Ouyang & Potter, 1993, 1995a,b 0.25 m id 10:9 m
M, E Schoenfelder et al., 1996 0.4 m id 15:6 m
M, E Mao et al., 1999 0.25 m id 10:5 m
E Bolland & Nicolai, 1999 411 mm id 8:5 m
Partial oxidation of methane M Pugsley & Malcus, 1997 NA
to syngas
Steam methane reforming M Matsukata et al., 1995 NA
B. Gas-Solid Reactions
Calcination E Legros et al., 1992 0:15 m sq: 7:3 m
Combustion M, E Li et al., 1995 0.20 m square
M, E Talukdar & Basu, 1995a,b 0:20 m sq: 6:5 m
0:51 m sq: 14:3 m
3:96 m sq: 21:4 m
M, E Mattisson & Lyngfelt, 1998 2:5 m2 13:5 m
11 m2 17 m
56 m2 33:5 m
M, E Hartge et al., 1999 1:6 m sq: 13:5 m
M, E Torii et al., 1999 0:4 m sq: 20 m
Gasification M, E Jennen et al., 1999 0:3 m id 8 m
Roasting E Luckos & den Hoed, 2001 150 mm id 6:0 m
NA ¼ not applicable.
Axial dispersion
Interregion Core radius/
Authors Core Annulus Axial gradients mass transf. column radius
CA ¼ core/annulus; FF ¼ fast fluidization; NS ¼ not specified; PF ¼ plug flow; PM ¼ perfect mixing; TIS ¼ tanks-in-series.
Figure 34 Conversion vs. dimensionless rate constant from simple one-dimensional, two-region, axially uniform model for first-
order chemical reaction with different overall voidages of 0.90 and 0.96 and different values of the dimensionless core–annulus
interregion mass transfer coefficient.
series of equations are written that combine hydrody- Amos et al. (1993) included radial dispersion in the
namic balances with empirical correlations to give core zone, but the core occupied the entire cross
methods for predicting the evolution of voidages and section, with the solid particles recycling in a
flows as a function of the axial coordinate, z. This separate region beyond the walls of the vessel.
approach also incorporates correlations for a slip fac- A cluster/gas model (Fligner et al., 1994) with two
tor, discussed in Sec. 4.6.2 above. phases—a cluster phase containing all the parti-
cles as spherical clusters, each of voidage "mf , and
9.4 More Sophisticated Models a gas phase devoid of particles. Mass transfer is
assumed to be controlled by the resistance at the
The core–annulus models treated in the previous sec- outer surface of the clusters.
tion improve on the one-dimensional models covered Kruse et al. (1995) extended the model of
in the preceding section by making some allowance for Shoenfelder et al. (1994) to include terms for
the difference in behavior between the relatively dense radial dispersion in both the core and the annu-
wall region and the dilute core of fast fluidized beds. lus, with the radial dispersion coefficient assumed
However, the hydrodynamics are represented in a rela- to be identical in both zones. These models were
tively crude manner. As illustrated in Fig. 33, experi- further extended by Shoenfelder et al. (1996a) to
mental results show that reactant concentration varies provide for continuous variations throughout the
continuously across the entire cross section of the riser, entire cross section.
rather than there being a sharp discontinuity at a core– Ju (1995) developed a Monte Carlo model for a
annulus boundary. Hence models are needed that pro- CFB combustor with the riser divided into 40
vide for continuous variation across the riser, or at cells, 20 in the core and 20 in the annulus.
least a greater number of intervals in the lateral direc- Particles are introduced and tracked one by
tion. Such models include the following: one, with particles in the core only able to
move upwards or sideways subject to the laws
The model of Werther et al. (1992) still assumes a of chance, while those in the annulus can only
core–annulus structure, but allows for radial dis- move downwards or sideways. Some particles
persion in the central core, while disregarding reaching the top are assumed to be reflected
gradients in the annular wall zone. back down the riser. Particles devolatilize and
Wen-Ching Yang
Siemens Westinghouse Power Corporation, Pittsburgh, Pennsylvania, U.S.A.
Olazar and his associates (1992, 1993a–c, 1995, 1998) ðReÞDo;ms =Reynolds number at minimum
studied the design, operation, and performance of a spouting base of Do
conical spouted bed and found that the conical =included angle of the cone
spouted bed is especially useful for hard-to-handle
solids that are irregular in texture or sticky. A conical 2.8.2 Bed Voidage Along the Spout Axis
spouted bed is depicted in Fig. 3. The conical spouted
bed exhibits pronounced axial and radial voidage pro- The bed voidage along the spout axis at r ¼ 0 was
files that are quite different from the cylindrical found to be parabolic and dependent on the system
spouted beds. variables employing an optical fiber probe (San Jose
et al., 1998). It can be calculated from
2.8.1 Minimum Spouting Velocity
z 2
"ð0Þ ¼ 1 E ð16Þ
1:68 0:57 H
Db
ðReÞDo ;ms ¼ 0:126Ar 0:5
tan ð15Þ where E is an empirical parameter varying between 0.3
Do 2
and 0.6 and is empirically correlated as follows
where Db =upper diameter of the stagnant bed 0:12 0:97
Di =diameter of the bed bottom Db Ho U 0:7 0:25
E ¼ 1:20 ð17Þ
Do =diameter of the inlet Do Di Ums
The recirculating fluidized bed with a draft tube con- There is considerably more operation and design
cept is illustrated in Fig. 4. This concept was first called flexibility for a recirculating fluidized bed with a
a recirculating fluidized bed by Yang and Keairns draft tube. The downcomer region can be separately
(1974). Several other names have also been used to aerated. The gas distribution between the draft tube
describe the same concept: the fluid-lift solids recircu- and the downcomer can be adjusted by changing the
lator (Buchanan and Wilson, 1965), the spouted fluid design parameters at the draft tube inlet. Because the
bed with a draft tube (Yang and Keairns, 1983; draft tube velocity and the downcomer aeration can be
Hadzismajlovic et al., 1992), the internally circulating individually adjusted, the solid circulation rate and
fluidized bed (Milne et al., 1992; Lee and Kim, 1992); particle residence time in the bed can be easily con-
or simply a circulating fluidized bed (LaNauze, 1976). trolled. Stable operation over a wide range of operat-
The addition of a tubular insert, a draft tube, in a ing conditions, solids circulation rates up to 100 metric
spouted fluid bed changes the operational and design tons per hour, and a solids loading of 50 (weight of
characteristics of an ordinary spouted bed. Notably, solids/weight of air) in the draft tube have recently
there is no limitation on the so-called ‘‘maximum spou- been reported by Hadzismajlovic et al. (1992). They
table bed height.’’ Theoretically, a recirculating flui- used a 95.3 cm diameter bed with a 25 cm diameter
dized bed with a draft tube can have any bed height draft tube using 3.6 mm polyethylene particles. A
desirable. The so-called ‘‘minimum spouting velocity’’ detailed discussion of the recirculating fluidized bed
will also be less for a recirculating fluidized bed with a with a draft tube is recently published by Yang (1999).
draft tube because the gas in the draft tube is confined Operating conditions for a recirculating fluidized
and does not leak out along the spout height as in an bed can be flexible as well. The bed height can be
ordinary spouted bed. lower than the draft tube top or just cover the draft
where
Rcf ¼ ðUcf Þp =ðUcf Þatm ð42Þ
4.1.1 Momentum Dissipation of a Gas-Solid Experimentally observed jet half-angles range from 8
Two-Phase Jet to 12 for the experimental data mentioned above.
These compare to 10 suggested by Anagbo (1980)
Gas velocity profiles in a gas–solid two-phase jet inside
for a bubbling jet in liquid.
a fluidized bed were determined using a pitot tube by
Yang and Keairns (1980). The velocity profiles were
4.1.3 Bubble Dynamics
integrated graphically, and gas entrainment into a jet
was found to occur primarily at the base of the jet. A To describe the jet adequately, the bubble size gener-
reasonably consistent universal velocity profile can be ated by the jet needs to be studied. A substantial
obtained by plotting ðUjr Ujb Þ=ðUjm Ujb Þ vs. r=r1=2 , amount of gas leaks from the bubble to the emulsion
comparable with the Tollmien solution for a circular phase during the bubble formation stage, particularly
homogeneous jet in an infinite medium (Abramovich, when the bed is less than minimally fluidized. A model
1963; Rajaratnam, 1976). developed on the basis of this mechanism predicted the
of the response being an indication of the magnitude of The region has a thickness of approximately 0.25 m.
the solids circulation rate. The number of major peaks Between these two regions the solids flow is alterna-
per unit time is closely related to the actual bubble tively upward and downward, depending on the
frequency in the bed. The force probe data allow the approach and departure of large bubbles. No stagnant
identification of three major solids flow regions in the region was evident anywhere in the bed.
3 m model as shown in Fig. 8. At the central portion of In addition to the three solids circulation regions
the bed, the solids flow is induced upward primarily by readily identifiable, the approximate jet penetration
jetting action at the lower bed height and by large depth and bubble size can also be obtained from Fig.
bubbles at the higher bed height. At the outer region 8. The jetting region can be taken to be the maximum
next to the vessel wall, the solids flow is all downward. average value of jet penetration depth. From the jet
T. M. Knowlton
Particulate Solid Research, Inc., Chicago, Illinois, U.S.A.
Figure 4 Optimum standpipe operating regions for group A and group B solids.
There are two basic types of standpipe configurations, nearly all standpipes used in fluidized bed standpipe
the overflow standpipe (Fig. 6A) and the underflow systems are constrained at the bottom. The standpipes
standpipe (Fig. 6B). The overflow standpipe is so discussed in this chapter will all be bottom-constrained
named because the solids overflow from the top of standpipes.
the fluidized bed into the standpipe, and there is no In any gas–solids flow system, a pressure drop loop
bed of solids above the standpipe. In the underflow can be defined so that the sum of the pressure drop
standpipe, the solids are introduced into the standpipe
from the underside, or bottom, of the bed or hopper,
and a bed of solids is present above the standpipe. With
this definition, a cyclone dipleg is classified as an over-
flow standpipe because there is no bed of solids above
the entrance to the dipleg.
With two types of standpipe configurations and the
two typical standpipe flow regimes (fluidized and
packed bed), there are four different types of stand-
pipes:
however, and at high mass fluxes through the dipleg, standpipe (reducing the mass flux), or reducing the
this dilute-phase pressure drop can be large enough length of the standpipe, if possible (increasing P=Lg).
that with a sufficiently long dipleg and a low pressure Wirth (1995) conducted a study of a standpipe oper-
drop across the dipleg, the dilute phase alone can gen- ating between a cyclone and a loop seal at the bottom
erate the required sealing pressure drop. This situation of a standpipe. He also found that if the solids mass
is shown schematically in Fig. 9. flux in the standpipe increased above a certain value,
When a dipleg operates in streaming flow, substan- the standpipe lost its dense-phase seal at the bottom.
tial amounts of gas can be dragged down the dipleg by This increased mass flux was accompanied by increas-
the flowing solids. The amount of gas being transferred ing amounts of gas flowing through the standpipe. If
may be very large (as much as 1/3 of the gas entering the solids mass flux was increased further beyond a
the cyclone can be transferred down the dipleg under certain threshold value, an even greater flow of gas
certain conditions; PSRI, 1995). In many cases, it is was observed in the standpipe, with the result that
not desirable to have so much gas being transferred even the upward-flowing part of the loop seal became
down the standpipe with the solids. By increasing the dilute. Wirth found that this situation could be con-
pressure drop across the standpipe (increasing the trolled by decreasing the solids flux through the stand-
P=Lg across the standpipe) or reducing the solids pipe. This could be accomplished in two ways, by
mass flux through the standpipe, the amount of gas decreasing the rate of solids flowing around the unit,
being transferred down the standpipe can be substan- or more practically, by increasing the diameter of the
tially reduced. Practically, this means that the gas flow standpipe. Increasing the diameter of the standpipe
down the standpipe or dipleg can be decreased by decreases the solids mass flux in the downcomer and
immersing the standpipe further in the fluidized bed therefore the dilute-phase density in the standpipe.
(increasing P=Lg), increasing the diameter of the When the dilute-phase density is decreased below a
critical value, a dense phase appears at the bottom of Thus for this packed bed underflow standpipe case, the
the standpipe. standpipe must generate a pressure drop greater than
Streaming flow is also more likely to occur with long P2 -P1 . This is shown as Case I in the pressure diagram
standpipes. Therefore to prevent this type of situation of Fig. 10B.
from occurring and causing poor loop seal operation in If the gas flow rate through the column is
combustors, it is recommended that a large-diameter increased, Pd will increase. If Pv remains constant,
downcomer be used. then Psp must also increase to balance the pressure
In Fig. 10A, solids are being transferred through an drop loop. This is shown as Case II in the pressure
underflow standpipe (operating in the packed bed diagram of Fig. 10B. Unlike the overflow standpipe
mode) from the upper fluidized bed to the freeboard case, the solids level in the standpipe cannot rise to
of the lower fluidized bed against the differential pres- increase the pressure drop in the standpipe. However,
sure P2 -P1 . The differential pressure P2 -P1 consists of the P=Lg in the standpipe must increase in order to
the pressure drop across the gas distributor of the balance the pressure drop around the loop. This
upper fluidized bed Pd . However, there is also a pres- occurs in a packed bed standpipe because of an
sure drop across the solids flow control valve Pv . increase in vr in the standpipe. This can be visualized
Therefore the standpipe pressure drop Psp must with the aid of Fig. 10C.
equal the sum of Pd and Pv , i.e., For Case I, the pressure drop in the bed was satis-
fied by having the standpipe operate at point I on the
P
Psp ¼ Hsp ¼ Pd þ Pv ¼ P2 -P1 þ Pv P=Lg versus vr curve, as shown in Fig. 10C. When
Lg the pressure drop across the distributor increased,
ð7Þ the vr in the standpipe adjusted to generate a higher
catalyst is transferred to a regenerator where the car- ing manner. In standpipe 1, slide valve 1 controls the
bon is burned off of the catalyst. This heats up the temperature at the outlet of the riser by varying the
catalyst and also restores its reactivity. However, flow of regenerated solids to the riser. The slide valve
before the catalyst is returned to the regenerator via in standpipe 2 controls the fluidized solids level in the
an angled standpipe/slide valve combination, hydro- stripper (King, 1992). Generally, slide valves in FCC
carbon vapors are removed from the catalyst in a standpipes are operated with a pressure drop between
steam stripper. From the regenerator, the catalyst is about 15 kPa (the minimum required for good control
transferred down another angled standpipe/slide of the catalyst) and 100 kPa. Higher pressure drops
valve combination and is injected into the bottom of result in excessive valve wear. Good valve design will
the riser. result in the valve operating between 25 and 75% open,
In most FCC units, the solids flow rate around the with the valve port area about 25 to 50% of the stand-
system is controlled by a slide valve. FCC units pipe open area. A typical design is to for the valve to
designed by the Kellogg Brown and Root Company operate in the middle of the ranges shown above (i.e.,
use another type of valve called a cone, or plug at about 50 kPa pressure drop and 50% open) as
valve, to control the solids flow rate (Wrench et al., reported by King (1992).
1985). Underflow fluidized standpipes in FCC units are
Solids flow in standpipes in FCC units such as those operated in a vertical configuration, a completely
shown in Fig. 14 are generally controlled in the follow- angled configuration, or a hybrid configuration in
which both vertical and angled sections are present ing up along the upper portion of the standpipe while
(Fig. 15). Angling a standpipe is a convenient way to the solids flowed down along the bottom portion of
transfer solids between two points that are separated the standpipe (Fig. 16). The pressure buildup in the
horizontally as well as vertically. However, it has been hybrid standpipe was lower than that in the vertical
found (Karri and Knowlton, 1993; Yaslik, 1993) that standpipe, and Karri and Knowlton (1993) reported
long angled underflow fluidized standpipes do not per- that the maximum solids mass flux possible in a
form as well as vertical standpipes. hybrid angled underflow fluidized standpipe was less
Sauer et. al. (1984) and Karri and Knowlton than that attainable in a vertical underflow fluidized
(1993) studied hybrid angled standpipe operation standpipe (Fig. 17). The principal reason for this is
using transparent standpipes to allow visual observa- that the rising bubbles in the angled section of the
tion of the flow. Both found the gas and solids standpipe become relatively large at a low solids
separated in the standpipe, with the gas bubbles flow- flow rate (and low aeration rate). At a certain solids
mass flux, the bubbles become large enough to bridge FCC units with Geldart group A catalyst. However,
across the vertical section at the top of the standpipe, these standpipes are relatively short and are designed
hindering the solids flow. When this occurs, the max- so that the mass flux through them is not too high, so
imum solids flow rate in the hybrid angled standpipe that they can be operated satisfactorily. Yaslik (1993)
has been achieved. found that a long angled standpipe had a limited solids
Karri et. al. (1995) showed that the solids flow rate circulation rate relative to vertical standpipes. Thus
through a hybrid angled standpipe can be increased if a when operating a hybrid angled standpipe or a true
bypass line (Fig. 18) is added between the top of the angled standpipe it is essential (1) to keep the solids
angled section of the standpipe and the freeboard of mass flux through the standpipe below a value that will
the bed above it. The effect of adding the bypass line at lead to slugging, and (2) to keep the line as short as
the top of the standpipe can be seen in Fig. 17. The possible so that the large gas slugs will not have as
bypass line allows the bubble gas from the angled sec- great a length in which to form.
tion to bypass the vertical section of the pipe so that
large bubbles are not formed there. Thus the solids 2.5 Nonmechanical Solids Flow Devices
flow rate can be increased. Karri et. al. (1995) reported
that if the bypass was used, the solids flow rate could A nonmechanical solids flow device is one that uses
be increased to such a value that the solids velocity in only aeration gas in conjunction with its geometrical
the hybrid standpipe became greater than the bubble shape to cause particulate solids to flow through it.
rise velocity, and the bubbles were carried down the Nonmechanical solids flow control devices have sev-
standpipe with the solids. When the bubbles were being eral advantages over mechanical solids flow devices:
carried down the standpipe by the solids, the bypass
line could then be closed and the standpipe would 1. They have no moving mechanical parts, which
operate without slugging in the vertical section. are subject to wear and/or seizure. This feature is espe-
Even though vertical standpipes can transfer solids cially beneficial when operating at elevated tempera-
more efficiently than hybrid angled standpipes, true tures and pressures.
angled standpipes (those containing no vertical sec- 2. They are inexpensive because they are con-
tion) are commonly operated satisfactorily in large structed of ordinary pipe and fittings.
T. M. Knowlton
Particulate Solid Research, Inc., Chicago, Illinois, U.S.A.
2.2 Flow Patterns in Cyclones The rotating vortex in the main body of the cyclone
below the gas outlet tube sets up a tangential velocity
In a tangential-inlet reverse-flow cyclone, the cyclone that varies with the radius of the cyclone. As the
inlet translates the linear inlet gas flow into a rotating radius, rw , decreases from the wall toward the center
vortex flow. As shown in Fig. 1, the gas–solids mixture of the cyclone, the tangential velocity increases to a
enters an annulus region between the outer wall of the maximum that occurs at a radius slightly less than
cyclone and the outer wall of the gas outlet tube. As the outer radius of the gas outlet tube, ro , as shown
the gas–solids mixture spirals downwards, it sets up a in Fig. 4. At radii much smaller than that of the gas
vortex with an axial direction downward toward the outlet tube, the tangential gas velocity decreases to a
solids outlet. much lower value at the center.
The inner vortex (often called the core of the vortex)
rotates at a much higher velocity than the outer vortex.
In the absence of solids, the radius of this inner vortex
has been measured to be 0.4 to 0:8ro . With axial-inlet
cyclones, the inner core vortex is aligned with the axis
of the gas outlet tube. With tangential or volute
cyclone inlets, however, the vortex is not exactly
aligned with the axis. The nonsymmetric entry of the
tangential or volute inlet causes the axis of the vortex
to be slightly eccentric from the axis of the cyclone.
This means that the bottom of the vortex is displaced
some distance away from the axis, and can pluck off
and reentrain dust from the solids sliding down the
cyclone cone if the vortex gets too close to the wall
of the cyclone cone.
At the bottom of the vortex, there is substantial
turbulence as the gas flow reverses and flows up the
Figure 3 Schematic drawing of tangential, volute, and axial middle of the cyclone into the gas outlet tube. As indi-
inlet cyclones. cated above, if this region is too close to the wall of the
3.1 Series
is only 0.025 m/s for a flux in the dipleg of 20 kg/s- Increasing the pressure drop across the cyclones will
m2 . Gas can easily flow up the dipleg against this very improve the solids distribution but is not always
low velocity. Conversely, in first stage cyclone diplegs sufficient to ensure equal solids flow into each cyclone.
where the solids flux may be as high as 750 kg/s-m2 , Typically, when the number of parallel cyclones is
the solids velocity is so high (approximately 0.94 m/s small (less than 6 to 8), each cyclone will have an indi-
at a dipleg density of 800 kg/m3 ) that gas is carried vidual inlet and outlet duct. However, when the num-
down the dipleg with the solids. ber of cyclones exceeds this number, the cyclones
Because of these recurring problems with second- generally have common inlet and outlet plenums and
stage cyclone diplegs, companies sometimes choose to a common collection bin. Parallel arrangements of this
use a single high-efficiency cyclone in place of two type frequently result in the operating problems of
cyclone stages. Even though the overall efficiency equalizing the gas and solid flow rates into each
using a single cyclone is slightly less than using two cyclone, and preventing backflow from the common
cyclone stages, they find that this cyclone arrangement chamber into one or more cyclones and reentrainment
optimizes the operation of the process by minimizing of solids from the common collection hopper back into
upsets and reducing downtime using a single high- one or more cyclone outlets. To prevent these two
efficiency cyclone. It is also a much simpler system as problems, a device is often placed at the outlet of
well and saves on initial capital costs. each parallel cyclone. This device (generally proprie-
tary) reduces the area available for flow and acts as a
3.2 Parallel ‘‘check valve’’ to prevent backflow and dust reentrain-
ment into the cyclone.
Several small cyclones are placed in parallel when it is Placing a large number of cyclones in parallel in a
not possible to fit a single large cyclone into the common bin can result in distribution problems
available height, or when extremely high centrifugal because it will be easier for the gas and solids to flow
forces are required. It is difficult to ensure equal dis- through the closest cyclone than one located some dis-
tribution of gas and solids into parallel cyclones. This tance away from the inlet. Multiple inlets to the com-
can lead to cyclone inefficiencies and increased wear on mon vessel reduce this problem, but result in increased
the cyclones taking the bulk of the solids flow. complexity and cost.
xi ¼ the weight fraction of each individual When dividing the overall particle size distribution
entering the cyclone into several cuts, it is advisable
cut size to use a minimum of at least 10 cuts. It is also better
Eoi ¼ the collection efficiency of the cyclone not to divide the distribution into equal size cuts.
for each individual cut size, %: Instead, it is better to bias the cuts so that the small
end of the particle size distribution has more cuts than
6. Modify the low-loading cyclone collection effi- the large end. This is because small particles are more
ciency for the effect of loading. difficult to collect than large particles, and it is the
7. Determine cyclone pressure drop. small particles that will determine the collection effi-
A detailed description of this approach is described ciency of the cyclone. For each dpi in the partition,
below. there will also be a corresponding weight fraction xi ,
for each cut.
12.1.1 Estimation of Ns
12.1.3 Calculate dpi =Dp;th and Determine Eoi
The number of effective spiral paths taken by the gas in
the outer vortex of the cyclone can be estimated by the After Dp;th has been calculated and the particle size
curve in Fig. 9. The number of effective cyclone spirals distribution has been partitioned, the ratio of
Figure 9 The effect of gas velocity on the number of solid spirals in a cyclone. For maximum velocity, use greater of inlet or
outlet velocity.
X
n
Eo ¼ xi Eoi
i¼1
George E. Klinzing
University of Pittsburgh, Pittsburgh, Pennsylvania, U.S.A.
2.11 Bends also met with some success in analyzing the pressure
loss around bends. One factor that is important in
Most conveying systems have bends or connections. If bend analysis in pneumatic conveying is to avoid hav-
one has a system that is relatively long (> 300 ft) and ing three bends in quick succession in the design. This
has few bends, the presence of the bends has a minor arrangement will most often lead to blockages and
effect on the overall system pressure loss. If a system unsteady operation.
has several bends and is of relatively short length
(< 300 ft), the bend pressure loss is significant. 2.12 Choking Conditions
Again, for the analysis of bend pressure loss, the linear
combination of the effects of the gas and the effects of Choking conditions for vertical flow is likened to the
the solids is applied. The classic work of Ito (1959, saltation condition in horizontal flow. The most com-
1960) for single-phase flow is used to establish the plete analysis of choking to date has been carried out
base for the pressure loss in bends. The work of by Yang (1975), who correlated a vast amount of exist-
Schuchart (1969), although performed on a limited ing literature data. These two expressions must be
material, has been employed with success to determine solved simultaneously:
the solids contribution to the bend pressure loss.
4:7
2gD "choking 1
Pbend 2RB 1:15 2 ¼ 0:01
¼ 210 ð15Þ uchoking ut
P solids D
up 1 "choking ¼ uchoking ut 1 "choking Þ
The design suggestion of using the traditional pressure
loss factors for single-phase flow in two-phase flow has ð16Þ
Figure 3 Phase diagram-type plot of reduced pressure analog versus the volume analog at constant reduced gas flow analog.
Note similarities of strand flow to liquid phase, flow above strand flow to vapor phase, and homogeneous flow to gas phase.
If one experiences excessive line wear, one can In order to design a dilute-phase pneumatic conveying
system, one needs to have a mechanism to select and
1. T-bends can be used.
calculate a number of factors and components. The
2. The material-to-air ratio can be increased.
first question is, What type of system would be appro-
3. Step pipe diameters through the system can be
priate for the conveying operation? Should one use a
carried out to maintain a constant velocity
vacuum or pressure system? Is the dilute-phase the
throughout the system.
mode most appropriate for transport? Using an expert
When material breakage is present, one can system approach, our laboratory has developed a pro-
gram entitled NUSELECT (Pneumatic Conveying
1. Reduce the conveying velocity Consultants) to help answer these types of questions.
2. Eliminate bends, cyclones, and impact points After selecting the system, one needs to choose the
3. Reduce the velocity before the terminal point of appropriate feeder to inject the material into the pipe-
the system. line for transport. Again, we have addressed this ques-
George E. Klinzing
University of Pittsburgh, Pittsburgh, Pennsylvania, U.S.A.
Figure 2 Effect of electrostatics charging on the pressure drop behavior of PVC particles in 0.0254 m (I.D.) system—
Ws ¼ 1:14 kg=min.
set of experiments is shifted to the right, thereby The first minimum occurring at the lower gas velo-
changing the previous minimum in the gas velocity city of the double minimum increased with increasing
that occurred without electrostatics. Operating at solids flow rate (Figs. 4 and 5). This increase in pres-
what was thought of as the previous minimum would sure drop, as the gas velocity was in the range of 4.5–
have increased the pressure drop significantly, which 7.6 m/s, occurred owing to the clustering of the part-
could result in reduced conveying capacity. Three dif- icles caused by the electrostatic generation of the glass
ferent solids flow rates were investigated at several gas beads with the Lucite tubing. This clustering caused
velocities, with the higher solids flow rates having the the pressure drop in the system to be equivalent to
most pronounced results. that of larger particles. The greater the charging of
The solids were mainly in a homogeneous transport the particles, the larger and more defined the clusters
state during the initial tests, except at the lowest gas were. This was also evident by the tracking of the
velocity. During the second series of tests, the flow P=L vs. time for a given Ug and Ws . The pressure
patterns changed from a homogeneous flow at the drop would continue to increase until a discharge, of
start of the experiments at low gas velocities to a the system, as evident by sight and sound, occurs. At
more clustered or packet flow at the intermediate test this point the pressure drop would fall back to a cer-
velocities. The particles formed clusters for gas veloci- tain level and continue to increase again with time until
ties in the range of 4.5–6.0 m/s for the particular set of there is another discharge. A certain residual charge
conditions investigated. resided in the particles and on the pipe.
Operating at a gas velocity for what seemed to be a experiments. Figure 6 shows the flow patterns observed
minimum for the pressure drop in the presence of with electrostatic charging.
electrostatics could result in the pressure drop being
increased substantially. This would lead to undersized 5.4 Electrostatics with Larger Glass Beads
equipment for the transport of the material and would
result in less material transported than the system was Electrostatics was also observed when 545 m glass
actually designed for. beads were used as the transport solid. This time the
The double minimum condition has also been pressure drop was not observed or investigated as a
reported by Zaltash (1987) and Klinzing et al. (1986). function of gas velocity, but rather pressure fluctua-
Klinzing suggests that the smaller particles behave as a tions were observed with time. Electrostatic discharges
cluster at the dilute-phase regime, and the clusters tend were observed with arcs discharging from a few centi-
to break up as the choking regime is approached, meters to a meter in length accompanied by a powerful
resulting in the particles behaving as smaller particles ‘‘snapping’’ noise.
and thus causing an oscillation in the pressure drop. The pressure drop was seen to decrease suddenly
This described flow phenomenon was observed in the upon discharge, rise sharply prior to another discharge
of the solids, and fall back to the original level citor, since the solids were isolated from any conduct-
(pressure drop) after discharge. Table 1 shows a cyclic ing medium due to the Lucite walls, and therefore it
example of this for the 545 m glass beads at a gas contained a charge. The pressure drop decreases just
velocity of 6.6 m/s and a solids feed rate of 1100 g/ prior to the charging portion of the curve. Since the
min. Standard deviation of the pressure drop was also particles are large (545 m) in comparison with the 9
found to increase with the increased charging of the 7 m glass beads, it is hypothesized that the particles
system. are weakly held together or attracted by electrostatic
The rise and fall of the pressure signal is similar to charges, causing a higher pressure drop than if they
that of a voltage vs. time plot observed during the were individual entities. The dip in the pressure drop
charging and discharging of a capacitor (Floyd, curve prior to the charging and discharging portion of
1987). Figure 7 shows the similarity of the curves. the pressure drop curve could be due to the particles
In effect, the transport of the solids causes the parti- separating (behaving as individual particles) and pos-
cles to become charged, and the charge accumulates sibly losing some of the charge. Then upon losing the
on the solids until the solids reach their dielectric charge or becoming separated, there is an increasing
breakdown and discharge across an air (or nitrogen) force that causes the particles to reattract or cluster
gap. The collector could be considered a large capa- together.
P, Pa Time
Times between readings, i.e., tiþ1 ti ¼ t, are approximately the same
distribution in the plant where particles are usually Cheremisinoff (1986), Grace and Baeyens (1986),
circulated between columns pneumatically or by grav- which all concentrate on bubbling bed measurements,
ity flow. In many cases the reaction conditions of each and Werther et al. (1990, 1993), Soo et al. (1994), Yates
column are quite different. The gas flow for each col- and Simons (1994), Nieuwland et al. (1996), Louge
umn is usually separated by the so-called material seal- (1997), Chen (1999) or Werther (1999), although
ing, where pressure balance takes a major role. many of them are rather limited in scope.
Additionally, the state of solids and their distribution
in the column can be detected from pressure gradient
and pressure fluctuation. As long as the solids are flui-
2 DETECTION PRINCIPLES AND
dized, their weight is suspended by fluid, and the solids
DETECTORS
holdup can be detected from the static pressure gradi-
ent. In many cases, when a bed is defluidized (e.g., by
2.1 Temperature Sensors/Probes
sintering), the bed pressure drop suddenly decreases.
On the other hand, the pressure fluctuations indicate
2.1.1 Thermocouples
good or poor bubbling as well as the flow regime tran-
sitions. Thermocouples are the most common temperature
To obtain information of particular interest in the measurement devices. A detailed description of
relevant column, it is required to choose appropriate thermocouples and other temperature measuring
sensing as well as data processing systems and to orga- instruments and sensors can be found in Holman
nize proper measurements to obtain necessary infor- (1994) or Stephenson et al. (1999). Maximum tempera-
mation with a high accuracy. Accordingly, in the ture limit, up to which a thermocouple can be used
present chapter it is intended to provide information without damage, depends strongly on its diameter
to engineers, first on available sensors and probes as and the surrounding environment. The thicker the
well as relevant data necessary to use them, second on thermocouple, the higher the temperature application
data processing methods, and third on experimental limit, but on the other hand the longer the response
methods, particularly for fluidization measurements time.
and monitoring. In Sec. 2, a variety of sensors, probes,
and tracers are introduced. In Sec. 3, some information
on data processing and visual imaging is provided. In 2.1.2 Optical Pyrometer
Sec. 4, fluidized bed diagnostic techniques are The optical pyrometer is a noninvasive temperature
reviewed. measurement device utilizing Planck’s law to obtain
For further information readers can see temperature, T½K, from radiant energy flux,
Cheremisinoff and Cheremisinoff (1984), Eð
Þ½W, of a particular wavelength
:
Solids
Flow regime Particle Gas Solids mass Gas/solids
& structure characteristics Pressure Temperature Voidage Velocity velocity flux composition
Pressure sensors B, P, C B, P, C B, P, C B, P, C B, P, C
Thermocouples B, P, C B, P, C
Pyrometers B, P, C
Suction probes B, P B, P, C B, P B, P, C, B, P, C
Gs meters B, P, C B, P B, P, C
Capacitance probes B, P B, P B, P
Optical fiber probe B, P, C B, P, C B, P, C B, P B, P, C B, P B, P
Laser sheeting B B B B
LDV/PIV B, P B, P B, P B, P B, P B, P
Tomography B B B, P
Solid/gas tracers B B B B, P B, P
Camera observation B, P, C B, P, C B, P, C B, P, C B, P
(IR camera)
Gas sampling & injection B, P, C B, P B, P, C
where "ð
Þ is emissivity of an object, Since "ð
1 Þ "ð
2 Þ, Eq. (3) can be rewritten and solved
C1 ¼ 3:74 1016 Wm2 , and C2 ¼ 0:0114 mK. for T as
For shorter wavelengths and for lower tempera- !1
1 1
51
tures, the above equation is simplified to the formula T C2 ln X 5 ð4Þ
called Wien’s law:
2
1
2
C1 "ð
Þ C2
Eð
Þ ¼
exp ð2Þ 2.1.3 Infrared Camera/Thermometer
5
T
Another noninvasive tool to measure temperature is
The schematic of an optical pyrometer is shown in the infrared (IR) thermometer, which usually has an
Fig. 2. The radiation of a target body, whose tempera- optical response time below 0.1 s. The measurement
ture is measured, is viewed through the lens and filters. principle is based on determination of the thermal
The aim of the filters is to absorb the radiation and radiation Q of a target body. The radiation is propor-
reduce its intensity. A lamp is placed in the optical path tional to the temperature of the body according to the
of the incoming radiation. By adjusting the lamp cur- Stefan–Boltzmann law
rent, the color of the filament is changed so that it ð
matches that of the incoming radiation. Temperature Q ¼ Eð
Þ d
¼ "T 4 ð5Þ
calibration is made in terms of the lamp heating cur-
rent. Additionally, the filter is installed in the eyepiece where is the Stefan–Boltzmann constant,
to ensure that comparisons are made for essentially ¼ 5:67 108 W=m2 K4 .
monochromatic radiation, thus eliminating uncertain- MgF2 , ZnSe, and Sapphire (Al2 O3 ) glasses are used
ties resulting from variation of radiation properties as IR transparent media windows, e.g., MgF2 : wave-
with wavelength. Based on a similar principle, Ross length 3–5:4 m, 95% of IR passing through; ZnSe:
et al. (1981) determined the burning char temperature 0.5–20 m, 90%; Al2 O3 : 0.2–4 m, 85% transparency
by comparing its monochromatic photograph with for IR.
that of a calibrated filament. Since the IR thermometer must adapt itself to the
To avoid the uncertainties of temperature measure- temperature of the surroundings, it may not measure
ment associated with the uncertainty of emissivity, a the temperature accurately without enough time to
two-color pyrometer is recommended, which analyzes adapt. The measurement results are also affected by
the body radiation properties for two wavelengths. In the angle of optical axis to the surface as shown in
the two-color or ratio pyrometer the radiation is Fig. 3. When the angle is larger than 40 , an IR
detected at two separate wavelengths for which emis- camera placed at position A gives temperature values
sivities of the surface can be considered nearly the higher than the real ones.
same. Thus the ratio of the thermal radiation sensor IR cameras have been applied to so-called inverse
output calculated for two wavelengths by Wien’s law is heat transfer problems, which concern the identifica-
obtained as tion of unknown temperature distribution or thermal
resistance distribution inside a body from the observed Botterill et al., 1984) or by separating the convective
surface temperature distribution (Pasquetti and Le component from the total heat flux by using quartz
Niliot, 1991; Le Niliot and Gallet, 1998). windows (Basu and Knouche, 1988) or ZnSe windows
(Ozhaynak et al., 1984). Quartz windows have rather
2.2 Heat Flux Probes narrow transmittance bands for infrared radiation
(roughly 0.15–3:5 m wavelengths compared to 0.5–20
The overall heat flux in fluidized beds consists of a m for ZnSe) but are resistant to thermal shocks. To
radiative and a convective component. Assuming, for measure simultaneously the radiative and total heat
simplicity, that the probe surface and the suspension flux, Luan et al. (1999) proposed the multifunctional
flow are gray, parallel, and separated by an opaque probe shown in Fig. 4c. They determined the radiative
gas, one may express the total bed-to-wall heat flux as component of heat flux both by using a ZnSe window
to separate the two fluxes and by estimating the radia-
qtotal ¼ qrad þ qconv tive flux from comparison of the measurement from
ð6Þ two probes of different emissivities. More on heat flux
4
ðTbed Twall
4
Þ
¼ þ hðTbed Twall Þ measurement can be found in Mickley and Fairbanks
ð1="bed þ 1="wall 1Þ (1955), Botterill (1975), Mathur and Saxena (1987),
Leckner and Andersson (1992), Goedicke and Reh
where h is the heat transfer coefficient, T is tempera-
(1992), Werdermann and Werther (1993), Couturier
ture, " is emissivity, is the Stefan–Boltzmann con-
et al. (1993), and Molerus and Wirth (1997).
stant, and q is the heat flux.
Since heat transfer in fluidized beds is dominated by
2.3 Gas/Solid Pressure and Force Sensors/Probes
surface renewal of bed materials, the bed-to-wall heat
transfer coefficient for a vertical surface is much larger
In most cases the pressure is detected using diaphragm
at its leading edge. Accordingly, the heat transfer coef-
sensors classified in terms of sensing element as metal
ficient determined by small probes can be abnormally
strain gauge, piezoelectric semiconductor, electric
large. Furthermore, depending on gas and solids
capacitance, reluctance, and LVDT (linear variable-
motion, surface renewal rate, and the time fraction of
differential transformers) sensors. The structure of a
coverage by bed material, the heat transfer coefficient
silicon semiconductor sensor, where the pressure is
varies for different segments of the surface, particularly
detected by a silicon diaphragm, is shown in Fig. 5.
for horizontal tubes. The requirement to determine the
In most cases these sensors can be used at temperatures
instantaneous heat transfer coefficient is the low ther-
lower than 70 C.
mal inertia of the probe, such as that used e.g., by
Mickley et al. (1961) or Ma and Zhu (1999) and
2.3.1 Gas Pressure
shown in Fig. 4a–b, respectively.
In some experiments it is necessary to separate the Gas pressure sensing is the most basic issue in fluidiza-
radiative and convective heat fluxes. This can be done tion and solids transport. The key points of the sensing
by using two probes of different emissivities (e.g., system are the volumetric capacity of its tubing and the
ability to prevent the tube from solids plugging. Figure where h is the damping ratio p is the amplitude of
6 shows a simplified pressure measurement system con- pressure signal impressed on the sensor, p0 is the pres-
sisting of a leading tube of diameter dpp , length L and sure at the tube inlet, ! is frequency, and !n is the
volume V at the end where the pressure transducer is resonance frequency.
placed. The resonance frequency, !n , damping ratio, h, and
If laminar friction resistance is assumed for the flow phase angle of the signal, , are given by
in the tube, the pressure–amplitude ratio is given by sffiffiffiffiffiffiffiffiffiffiffiffiffiffi
(Holman, 1994) 3r2 c2
!n ¼ ð8Þ
4LV
p 1 rffiffiffiffiffiffiffiffiffiffi
¼
0:5 ð7Þ
p
2 2 3LV
0
1 ð!=!n Þ2 þ4h2 ð!=!n Þ2 h¼ ð9Þ
f cr3
2.3.2 Manometer
Manometers are classical and still useful gas pressure
measurement devices used also to calibrate semi-
conductor pressure sensors. Applying the manometers
Figure 5 Silicon semiconductor piezo-resistive pressure sen- as pressure measurement devices, caution should be
sor. (Chau et al., 1999, with permission of CRC Press.)
taken to prevent the overflowing as well as the
spontaneous oscillation of the liquids. To obtain
pressure distribution in one glance, the manometers
2hð!=!n Þ
¼ tan1 ð10Þ can be assembled and connected to a common-
1 ð!=!n Þ2 reservoir.
Figure 7 Effect of tube diameter, dpp , on pressure fluctuations. Particles: 68 m FCC, purge gas velocity: 1 m/s, pressure port
length: 2 m dead volume, 450 mm3 . (Xie and Geldart, 1997.) (With permission of Elsevier Science)
Figure 8 Frequency response of a particle pressure sensor to the gas pressure change. (a) Setup to test the dynamic response; (b)
gas pressure (measured by a pressure transducer) and the phase shift. (Campbell and Wang, 1990.) (With permission of Institute
of Physics)
2.4.3 Photodetectors
various types (gas, solid, semiconductor) as listed in
Table 3. The types and various applications of photodetectors
The ranges of output wavelengths of major com- are shown in Fig. 13. They are classified into two major
mercial LEDs and lasers applied to fluidized bed mea- categories: point sensors and image sensors. The struc-
surement are shown in Table 4. An example of an tures and characteristics of the typical point sensors,
emission spectrum of an LED in comparison with i.e., CdS cells, photodiodes, and photomultipliers, are
those of various lamps is presented in Fig. 11. The shown in Fig. 14. The characteristics of CdS cells and
most commonly used gas lasers are He-Ne, CO2 , and photodiodes are sensitive to temperature and require
Ar ion lasers. In the case of solid-state lasers the light is compensation circuits.
emitted by atoms, which are fixed within a glassy mate-
rial or crystal (e.g., Nd YAG laser, where the laser 2.4.4 Optical Fibers
medium is an yttrium-aluminum-garnet matrix with
trivalent neodymium ions present as impurities). Optical fibers are flexible fibrous light transfer media
Semiconductor lasers made usually of AlGaInP, used in optical systems. The fibers are made of plastics
GaAlAs, InGaAsP, or lead salt materials use the spe- or quartz glass. Quartz fibers are much more expensive
cial properties of a p-n junction. More precise data can than plastic ones and not so flexible. However, they are
be found e.g., in Weber (1999). temperature resistant (up to roughly 1000 C) and
If the phase matching condition is satisfied, the transmit light more efficiently, particularly in the UV
OPO (optical parametric oscillator) laser can continu- range. Whatever the material, the optical fiber has a
ously emit light of various wavelengths within the center core with a refractive index 1 , surrounded by a
range indicated by the equation layer of another material (called cladding) with a
refractive index 2 . 2 must be lower than 1 in order
s þ
i ¼
p ð16Þ to keep the light inside the core area. Plastic core is
usually made of PS (polystyrene) or PMMC (poly-
where
s ,
i , and
p are the wavelengths of signal, idler, methylmethacrylate). Plastic cladding is usually made
and pumping waves, respectively. of teflon or silicone. Figure 15 shows the transmission
LED Lasers
Peak
wavelength
Color Material (nm) Type Material Wavelength Power Operation
characteristics of PMMA fibers and quartz fibers ver- nonbundle-type probe or the diameter of a fiber bundle
sus the wavelength. probe should be five to ten times larger than the par-
ticle diameter.
The main difficulty in using optical fiber probes is
2.4.5 Optical Fiber Probe
associated with their proper calibration. For solids
Types of optical fiber probes are presented in Table 5, velocity measurements Ohki and Shirai (1976)
and some examples of probes are shown in Fig. 16. In calibrated their probe (cf. Table 5) by dipping its
the case of the reflection-type optical fiber probe, the top, about 2 mm, into a thin bed of the same mate-
volume from which the probe can detect the informa- rial placed on a rotating disk (a music record disk
tion depends on fiber configuration. The intensity dis- player was used) and obtained a cross-correlation of
tribution of reflected light as a function of the distance the signals received from two fibers. The cross-cor-
between a fiber tip and a flat surface is shown in Fig. relation can be successfully applied for the analysis
17. The size of the core of the optical fiber should be of velocity data unless the velocity fluctuates too
selected depending on the maximum diameter of par- often.
ticles being investigated. If individual particle passages To measure solids concentration, the calibration of
are to be detected, the fiber core diameter should be the probe is more difficult, because it requires a
less than or equal to the particle diameter. On the other homogeneous suspension. In many cases, the suspen-
hand, if only the passage of solids or the solids con- sion tends to contain agglomerates or clusters, which
centration is to be detected, the fiber diameter of a may seriously affect the output. Various ways to cali-
brate the probes can be found, e.g., in Matsuno et al. similar relationship exists for gas–solid systems, they
(1983), Hartge et al. (1986), Lischer and Louge compared the cross section average concentration
(1992), Amos et al. (1996), or Zhang et al. (1998). from the probe with the -ray adsorption or pressure
Matsuno et al. (1983) calibrated their optical fiber gradient data to obtain the calibration coefficient for
probe using free-falling glass beads assuming that quartz sand particles. Lischer and Louge (1992)
the particle velocity equaled the terminal velocity ut . inserted an optical fiber into the center of a capaci-
For solids concentration up to 50%, Hartge et al. tance wall probe and calibrated the optical probe by
(1986) confirmed a linear relationship between parti- comparing the two signals. Herbert et al. (1994)
cle volume concentration and light reflection intensity applied a method similar to that of Matsuno et al.
for the liquid–solid fluidized bed. Assuming that a (1983) to FCC particles but used particle velocity
determined by optical probe measurement. They a stationary source. This phenomenon is called the
found a clear effect of an antielectrostatic additive Doppler effect, and its frequency shift !i is expressed
(Larostat 519, PPG/Mazer Chemicals) in dispersing as
particles, as shown in Fig. 18 (in their original
paper Figs. 4a and 4b should be reversed). They !i ¼ ðks ki Þvp ð17Þ
obtained a power-law relationship between solids
volume fraction and light reflection. Zhang et al. where vp is particle velocity and ks and ki are the wave
(1998) also reported similar probe calibration. number vectors of emitted and scattered light, respec-
Rensner and Werther (1993) determined the tively.
effective measuring volume of a single fiber optical Since in many cases the velocity vp is not sufficiently
probe (d ¼ 0:6 mm) for FCC and quartz sand as a large to make a large shift, the interference of scattered
function of particle concentration. In both cases (i.e., lights from two laser beams at an angle is usually
quartz and FCC), the 50% transmission length was utilized. For such a situation let k1 and k2 denote the
less than 1 mm for the solids volume fraction "p ¼ wave number vectors for the two reflected lights, which
0:002 and less than 0.1 mm for "p ¼ 0:2. Concerning result in two different Doppler shifts, !1 and !2 .
the calibration of particle concentration probes Then, subtracting !2 from !1 , we obtain the fre-
Amos et al. (1996) did detailed analysis and experi- quency of the interference, fD , called the Doppler fre-
ments. quency, which can be expressed as a function of the
angle by
receiving systems, among which a dual-beam forward- strated in Fig. 19. The distance
between the two
scattering mode is the most popular and commercially bright peaks is expressed by
available. It is popular because the signal can be
A. Single Tip
Reflection type
a Single fiber df
dp p Hartge et al. (1986, 1988)
Bubble fraction Tsukada et al. (1994)
Two-fiber df > dp p Matsuno et al. (1983)
b
Bubble fraction Horio et al. (1985)
df dp Solid loading Cocco et al. (1995)
and velocity
B. Combined type
If a particle passes through the interference region A typical signal obtained from a particle passing
with velocity component vpx , we have a beat of through the measuring volume and having the dia-
scattered light with the following frequency f: meter less than the fringe distance d is shown in
Fig. 19b and in Fig. 20b (where it is shown as
vpx 2vpx
f ¼ ¼ sin ð20Þ signal A).
2
In the measurement of solid–fluid two-phase flow
This expression is the same as Eq. (18) obtained from we need to discriminate the signals from coarse parti-
Doppler shifts. Equation (20) is usually called the cles and from fines (i.e., tracer seed). As schematically
fringe model for the beat frequency in LDV. shown in Fig. 19b, a fine (seed) particle whose diameter
Figure 16 Examples of optical fiber probes. (a) Optical fiber probe for simultaneous measurements of solid and gas velocities
(solid: light reflection measurement; gas: transmission measurement with O3 tracer). (Horio et al., 1992a.) (b) Water cooled
optical fiber probe. (Werther and Hage, 1996.) (With permission of United Engineering Foundation)
The advantage of LDV is that the setup can detect
v p ¼ fD
ð21Þ
2 sinð=2Þ reversal flows for particles, bubbles, and droplets in a
f
single and a multiphase flow system. LDV also
dp ¼ H
ð22Þ enables to measure fluctuating velocity, size, and con-
2 l
centration of suspended particles. However, it is more
Notations in Eqs. (21) and (22) are the same as those suitable to apply to the risers of rectangular cross
for LDV except for f , , l, and H, which are the section than to circular ones. The LDV’s limitation
focal length of the transmitter lens, the phase shift in determining particle concentration (roughly up to
between two photomultipliers, the space between 1%) can be improved greatly by combining it with an
detectors (say e.g., #1 and #2 or #1 and #3), and the optical fiber probe. Wei et al. (1998) reported mea-
optical constant, respectively. surements of particle velocities for solid concentration
Figure 18 Calibration curves for FCC particles. (a) Without Larostat. (b) With 0.5 wt% of Larostat. (Herbert et al., 1994.)
(With permission of Elsevier Science)
up to 0.21. More information on LDV and its appli- diameter 3 mm, length 0.3 m; the bundle consists of
cations can be found e.g., in Watrasiewicz and Rudd 71,000 optical glass fibers, 10 m diameter each). The
(1976), Drain (1980), Arastopour and Yang (1992), light reflected from the moving particles returns
van de Wall and Soo (1994), and Onofri and through the same fiber and is detected by two linear
Tadrist (1999). image sensors via a lens and a half mirror.
The particle passage through the measuring space
2.4.7 Space Filters is shown in Fig. 22b,c. Each of the wave patterns (I)
through (IV) in Fig. 22c corresponds to each state (I)
Space filtering is a method of determining particle velo-
through (IV) in Fig. 22b. In Fig. 22c curve (I) shows
city with the spatial light distribution of a gratelike
the background timing signal to distinguish the out-
structure including interference fringes. Dual beam
put signal of image sensor 1 from sensor 2. When a
mode LDV is one type of space filter; the others
particle passes sensor 1, a line signal of n1 bits is
include a reticle type filter and a filter with gratinglike
detected and it ‘‘moves’’ from the n1 bit to the n2
detectors.
bit within time interval t. Thus the radial and ver-
In the case of reticle-type filters, the illuminated
tical components of particle velocity, vr and vz , are
particle passes through the measurement volume and
obtained as
reflects light to the grating, thus being a source of a
periodical signal (Fiedler et al., 1997). From the fre- Xr
vr ¼ ð24Þ
quency f of this signal, the particle velocity vp is deter- t
mined as W
vz ¼ ð25Þ
fg t
vp ¼ ð23Þ
where Xr is the distance between the bit positions n1
and n2 and W is the distance between the two image
where g is the grating constant and is the image scale, sensors.
i.e., the ratio of image size at grating to that at the A CCD (charge coupled detector) type space filter,
measuring point. shown in Fig. 23, was used by Fiedler et al. (1997) to
An example of a light-reflection-type space filter measure sand particle velocity in a CFB riser. The
system is shown in Fig. 22a. The light from halogen authors were able to measure the velocities between
lamps is focused by a lens and guided by an optical 0.12 and 40 m/s, at local solid volume concentration
image fiber bundle into the measuring space (bundle 0.1–4%. They reported that particles finer than 5 m
adhered to the optical window, but this did not where " is the voidage, Kf is dielectric constant of the
decrease the accuracy of measurement: it only caused fluid, and Ks is the dielectric constant of solids (i.e., the
an increase in the d.c. level by scattering and reflec- ratio of the permittivity of the system to the permittiv-
tion. ity of free space).
Some data of the relative dielectric constant, K, are
2.5 Capacitance and Conductivity Sensors/Probes shown in Table 6.
An example of a capacitance probe and its output
A capacitance sensor provides information on solids signals are shown in Fig. 24. To eliminate the influ-
concentration in the target space between its two elec- ence of ‘‘external’’ capacitance of various parts of the
trodes. When a voltage E is applied between the elec- probe (e.g., cables) that can have much larger capa-
trodes, an equilibrium charge Q is accumulated in citance than that of the probe itself, a third conduc-
them. The charge can be determined by integrating tor called a guard is necessary, as discussed by Acree
the measured electric current between the electrodes Riley and Louge (1989). More on capacitance mea-
over time. Capacitance is the ratio Q/E. For a set of surements can be found in Bottcher (1952), Bakker
two parallel plates of area A, separated by a homo- and Heertjes (1958), Meredith and Tobias (1960),
genous dielectric of thickness d, the capacitance is Lanneau (1960), Werther and Molerus (1973),
given by Yoshida et al. (1982), Brereton and Stromberg
(1986), Louge and Opie (1990), and Brereton and
kA Grace (1993). Capacitance tomography measure-
C¼ ð26Þ
d ments are described in Sec. 3.3.
where k is the dielectric constant or permittivity of the
media between the electrodes [F/m]. k is a function of 2.6 Acoustic sensors
frequency and varies with temperature. Usually it is
expressed in terms of the relative dielectric constant 2.6.1 Microphones
K of the medium as
Acoustic emission from fluidized beds also provides
k information on hydrodynamic conditions. Win et al.
K¼ ½ ð27Þ (1995) used a high-sensitivity microphone to detect
kvacuum
the carryover of coarse particles from a multisolid
In gas–solids systems the volume fraction of solids is circulating fluidized bed consisting of porous alumina
inferred from the measurements of the effective dielec- (coarse particles) and glass beads (the fines) They
tric constant (or effective permittivity), expressed by detected acoustic signals, which the particles gener-
ated on the cyclone surface. Since coarse particles
Keff ¼ "Kf þ ð1 "ÞKs ð28Þ had larger mass, they generated sounds of magnitude
Figure 23 CCD type space filters. Grating constant: 64 m; image scale (i.e., the size at grating divided by the size at measuring
point): 0.0621; sampling frequency: 156 kHz. (a) CCD type space filter probe. (b) Block diagram for CCD type space filter.
(Fiedler et al., 1997.) With permission of Elsevier Science
Figure 24 An example of a capacitance probe. (a) High-temperature two-needle capacitance probe of Wiesendorf et al. (1999)
(With permission of American Society of Mechanical Engineers). (b) Output signals from a 12 mW (Chalmers) and a 250 MW
(Gardanne) CFB boiler obtained with the probe shown in (a). (Johnsson et al., 1999.) With permission of DECHEMA e.v.
Figure 25 Faraday cage used by Fasso et al. (1982) (With permission of Elsevier Science) (a) and examples of electrostatic
probes of Soo et al. (1964) (Reprinted with permission from Ind. Eng. Chem. Fundam., 1964, 3, 98–106 Copyright # 1964
American Chemical Society) (b) and Boland and Geldart (1971) (With permission of Elsevier Science) (c).
surements (particularly for the particles having long 2. The activity of the source; high counting rate
half-life). More on radioactive particle tracking can can be measured more accurately than a low
be found in Seville et al. (1995), Stein et al. (1997), one.
and Larachi et al. (1997). More on radioactive tracers 3. The energy of -rays: according to the
is also given in Sec. 2.11. Lambert–Beer law, the attenuation is an expo-
nential function of the linear attenuation coeffi-
cient and the thickness of the material between
2.11 Tracers the source and the detector.
4. The half-life: it should be at least an order of
2.11.1 Solid Tracers magnitude longer than the duration of the
To obtain information on solids movement and mix- tracking study. Otherwise, the loss of source
ing, the solid tracers are fed into a fluidized bed and activity may be significant during the experi-
detected either in situ, as in the case of isotope tra- ments. On the other hand, too long a half-life
cers, or by sampling and analysis. It is assumed that time is not recommended, because it may bring
the tracers behave like the solids in the flow. Some about difficulties in the test equipment after the
examples of commonly used tracers are shown in experiment (e.g., owing to the residual activity
Table 8. of the tracer).
In choosing the radioactive tracer one should take
into consideration the following aspects: 2.11.2 Gas Tracers
1. The purity of radionuclide; to assure that - By injection of a tracer gas into the flow (continuous
rays are emitted from disintegration of the or batchwise form) and by measuring its concentra-
selected radionuclide and not from some of its tion and velocity one obtains information on the
impurities. residence time distribution (RTD), mixing or mass
transfer. Examples of applications of gas tracers to
fluidized beds are presented in Table 9. Figure 27
Table 7 Half-Life of Some Positron presents a model of gas flow in the fluidized bed
Emitting Isotopes and an example of recorded data. Gas tracking is
also a useful way to measure mass transfer in flui-
Nuclide Half-life Nuclide Half-life dized beds, as reported by Ebert et al. (1993) and de
82 62 62 Kok et al. (1986).
Rb 78 s Zn =Cu 9.2 h
O15 122 s Ga66 9.7 h
N13 10 min Cu64 12.7 h 2.12 Suction Probes for Solids Sampling
C11 20.3 min I124 4.2 days
Ga68 68 min Ge68 =Ga68 271 days
There are two kinds of solids suction: isokinetic and
F18 110 min Na22 2.6 years
Ti45 3.1 h nonisokinetic. In isokinetic sampling, the gas suction
velocity is adjusted to equal the local gas velocity.
Source: Benton and Parker, 1996. This method was mainly applied to aerosol or other
Chemical NaCl tracer FCC reactor, the tracer detected by electric Bader et al. (1988)
conductivity of sample–water solution
Limestone Diffusivity in a large-scale FB boiler Ito et al. (1999)
Phosphor Effect of tracer size on solids mixing in the Du et al. (1999)
0.14 m i.d., 10.4 m high FCC riser
Color Colored particles Semicircular fluidized bed, video recording Yang and Keairns (1982)
Color Colored particles Video recording of particle velocity in 2D Gbavcic et al. (1990)
fluidized bed
Fracture Spheres of different Kinetic forces acting on particles in a fluidized Kono et al. (1987)
strengths bed determined from analysis of tracer particles
Fluorescent Paint pigment attached Particle velocity in the cocurrent downward flow Brewster and Seader (1980)
to coal particles of coal and air
Fluorescent Tracer detected by LDV, 319 176 mm cross- Hamdullahpur et al. (1987)
section FB
FCC impregnated with Particles illuminated by UV light, reflected light Kojima et al. (1989)
a fluorescent material detected by optical fibers Nowak et al. (1991)
Magnetic Ferromagnetic Solids mixing in a bed of FCC, Dt ¼ 0:15 m i.d. Avidan and Yerushalmi
Bed operated in a bubbling, slugging, and (1985)
turbulent regime
Sand coated with ferrite Solids mixing studied by magnetic separation Horio et al. (1986a)
Char impregnated with Measurement of the flow of dark colored Yamaki et al. (1994)
magnetic powder particles
Thermal Hot particles Temperature measurements Westphalen and
Glicksman (1995)
Radioactive Co60 Observation of particle motion in a FB Kondulov et al. (1964)
Na22 Studies on particle motion in a 0.03 m i.d. gas- Borlai et al. (1967)
solid fluidized bed. Scintillation detector placed
below the fluidized column
Na24 Particle motion in a FB of glass beads Van Velzen et al. (1974)
Scintillation counter placed above the bed
Radioactive Na2 CO3 RTD measurement in CFBs of various sizes Helmrich et al. (1986)
Ga68 Particle tracking in a CFB Ambler et al. (1990)
Radioactive SiO2 Studies on RTD in a 82.8 mm i.d. CFB Patience et al. (1991)
Sc46 oxide Studies on particle motion and solids mixing in a Larachi et al. (1994)
3D liquid–solid fluidized bed Larachi et al. (1995)
Radioactive tracers of Measurement of local particle velocities in a cold Weinell et al. (1995)
various sizes and CFB
densities
F18 Particle tracking in a 0.15 m i.d. sand bed Seville et al. (1995)
Au198 Studies on the hydrodynamics of a CFB tracer Godfroy et al. (1996)
Positron emission Tracer velocity measured 250 times per second Stein et al. (1997)
Irradiated Cu Particle tracking, 3D resolution below 15 mm Stellema et al. (1998)
Mn56 and cordierite Studies on particles tracking in the 80 MWth CFB Lin et al. (1999a)
boiler. 16 NaI(T1) detectors used, data
collected with the frequency of 20 Hz
Ca137 tracer combined Solid concentration axial profile and Schlichthaerle and
with optical fiber set instantaneous local values measured in a 0.4 m Werther (1999)
i.d. 15.6 m CFB riser, Gs up to 50 kg/m2 s
Cankurt and Yerushalmi (1978) CH4 Radial gas diffusivity estimated from radial profiles of tracer concentration.
Adams (1988) CH4 Studies on gas tracking in a CFB riser
Khattab et al. (1988) O3 Studies on gas velocity in fluidized beds. The tracer was detected by optical
fibers through which UV light was transmitted
Horio et al. (1992a) O3 Optical fiber probe applied to measure particle and gas velocities
simultaneously; ozone injection upstream of two optical probes, tracer
gas velocity measured using UV light, optical fiber and photomultiplier
Ye et al. (1999) O3 Ozone decomposition within a CFB. Cast steel used as ozone decomposing
catalyst
Sauer and Wallen (1999) CO2 Studies on gas mixing in a 2D ABFBC
Yang et al. (1984) He Concentration profiles within the bed measured by continuous gas
injection
Bader et al. (1988) He Continuous injection of gas into an FCC bed
Li and Weinstein (1989) He Continuous He injection applied to study gas backmixing in a 0.152 m i.d.
CFB
Martin et al. (1992) He Gas radial velocity profiles and dispersions in a CFB studied by gas
chromatograph analysis of concentration of injected He
Shen et al. (1992) He Studies on hydrodynamics and mass transfer in a 2D FB, thermal
conductivity detector used to measure tracer concentration
Rivault et al. (1995) He Tracer injection into the leg of primary cyclone; tracer gas sampled by a
displaceable probe equipped with a filter and analyzed
chromatographically
Lin et al. (1999b) He Studies on gas RTD in a high density FCC unit. Solids mass flux up to
430 kg/m2 s
Krambeck et al. (1987) SF6 Studies on tracer adsorption in the FCC bed. Sulphur hexafluoride
injected below the bed. Tracer concentration determined 1.8 m above
the bed
Dry et al. (1995) Warm gas Studies on gas–solid contact efficiency in two CFBs of different scale,
injection, 102 mm i.d. and 600 mm i.d
liquid N2
injection
Li and Wu (1991) H2 pulses Studies on axial gas mixing and RTD in the CFB containing FCC
particles (dp ¼ 58 m)
White and Dry (1989) Ar pulses Studies on gas residence time, tracer detected by two rapid response mass
spectrometers, reactor 0.09 m i.d. CFB
ment point the suction should be conducted at both up inlet of the sampling system is recommended (e.g.,
and down positions of the tube in order to measure the Werther et al., 1990). Naruse et al. (1994) applied the
local upward and downward solids mass fluxes (one probe shown in Fig. 30a,b and were able to sample the
should remember that even in the core of the riser gas separately from the bubble phase and from the
the solids downflux always exists, as reported e.g., by dense phase. The bubble passage was detected from
Herb et al., 1992). pressure drop between two sensors, as shown in Fig.
30c.
2.13 Gas Sampling
3 DATA PROCESSING AND VISUALIZATION
Gas sampling is a basic tool providing information SYSTEMS
about the local/total gas composition and concentra-
tion. Examples of gas sampling studies are listed in 3.1 Data Processing
Table 11.
In order to separate the gas and solid phase during Signal processing after the measurements is required:
fluidized bed sampling, a ceramic porous plate at the (1) to reduce noises and (2) to obtain characteristic
values such as mean values, characteristic frequencies,
and phase delay between signals. More information can
be found in, e.g., Kawada and Minami (1994) and
Edgar et al. (1997). Table 12 shows the major algo-
rithms for data processing applied to the fluidized beds.
Some examples of data processing are shown in
Figs. 31–33. The simplest window function is unity
for all data (irrespective to the time they were taken)
and gives the arithmetic moving average. Figure 31
presents the ammonia IR spectrum processed by a
polynomial equation as a window function. The num-
ber of averaged samples 2m þ 1 should be chosen
appropriately to obtain the necessary information out
of the data set.
In the case of stationary signals, the averaging
Figure 28 Effect of gas suction velocity on solids sampling. enhances the S/N ratio. Such averaging is used in
(Leckner et al., 1991.) (With permission of American Society ordinary chemical analyzers, e.g., FTIR, ICP, EDS,
of Mechanical Engineers) Uo ¼ 3:4 m=s; Gs ¼ 13 kg=m2 s: EPMA. If the number of averaging is too large, the
Author(s) Remarks
averaging process becomes time-consuming. However, A schematic of signal processing by low-pass filtering is
by that averaging the S/N ratio can be improved with- shown in Fig. 33.
out knowing the frequency of noises. Figure 32 shows
a schematic of the change of the signal vs. the number
3.1.1 Autocorrelation
of averaging.
Filtering in the frequency domain is the most pop- The autocorrelation function, defined in Table 12,
ular method of signal denoising, where a conventional becomes periodic for periodic signals. Figure 34
resistor–capacitor circuit is used as an analog filter to shows examples of the autocorrelation functions
eliminate the noises, i.e., signal components having and power spectra for a bubbling bed and a slugging
frequencies much higher than that of the main signal. bed.
Figure 29 High temperature solids sampling probe of Mattisson and Lyngfelt (1995). (With permission of American Society of
Mechanical Engineers)
Author(s) Remarks
Swift et al. (1975) Flue gas sampling to measure the concentration of Hg, F, and the solids fraction
Yang et al. (1984) Continuous He injection into a FB. Gas concentration profiles measured
Bader et al. (1988) Detection of He injected into an FCC riser
Adams (1988) Application of CH4 tracer to a CFB riser
Atkinson and Clark (1988) Detection of bubbles in a FB. Bubble passage determined using a dual static pressure probe
Naruse et al. (1994) Gas sampling probe, shown in Fig. 30c, applied to sample gas from bubble and
Naruse et al. (1995) emulsion phases separately. Reactor: 0.1 m i.d. 0.7 m FBC; bed material: sand, fuel: coal
Kassman et al. (1995) Gas sampling system applied to measure the concentration of NH3 and HCN inside a
12 mW CFB boiler
Hansen et al. (1995) Particle probe,
-probe, gas probe and alkali metal sampling probe used to monitor a
20 MWe CFB combustor
Mann et al. (1995) Alkali-sampling probe applied to characterize sorbents for a PFBC.
Hayrinen et al. (1999) Alkali concentration measurements in a 10 MW PCFB using laser fluorescence and plasmas
spectroscopy sensors
Kassman et al. (1999) Gas sampling from a 12 MW CFB boiler
Lin et al. (1999c) Simultaneous measurement of N2 O and NOx concentration, as well as particle
temperature in the vicinity of a burning char particle
Figure 30 Gas sampling system (a), probe details (b), and bubble passage determined from pressure drop (c). (Naruse et al.,
1994.) (With permission of Combustion Institute)
Examples
Arithmetic 1 Xm
Pressure S/N improvement, Hatano et al.
yi ¼ xði þ jÞ
moving 2m þ 1 j¼m equivalent to simple (1999)
average average for random
noise
Averaging 1X N
Pressure S/N improvement, for
y¼ x
N i¼1 i signal only repeatable
measurement; frequency
response information for
signal and noise is not
necessary
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Ð x2 ffi
Root mean Pressure Phase transition Yerushalmi
x1 ðx xÞ dx
2
square (RMS) xrms ¼ fluctuation and Cankurt
x2 x1
(1979)
Probability 1 lim Xn
t1 Determination of threshold
pðxÞ ¼
distribution x T!1 i¼1 T value to distinguish:
function (PDF) Light Bubble and emulsion phases Horio et al. (1985)
LDV Gas and particle signals in Tsuji and
pneumatic transport Morikawa (1982)
Light Cluster and lean phase Horio et al.
(1988)
ðT
Autocorrelation lim Pressure Slugging condition Broadhurst and
RðÞ ¼ xðtÞxðt þ Þ dt
function T!1 0 Becker (1976)
Pressure Bubbling condition Fan et al. (1981)
Light Bubbling regularity Greon et al. (1997)
Light Freeboard turbulence Horio et al. (1980a)
Heat transfer Dynamics of CFB Li et al. (1993)
ðT
Cross-correlation lim Pressure Bubbling condition Fan et al. (1981)
Rxy ðÞ ¼ xðtÞyðt þ Þ
function T!1 0 Pressure Pressure wave velocity Roy and
Davidson (1990)
Light Particle velocity Horio et al. (1980a)
Examples
Fractal analysis Self similarity, evaluated by Image Cluster shape Ito et al. (1994)
fractal dimension Pressure Chaos analysis Fan et al. (1990)
Ross-Pence (1997)
Solid Chaos analysis Bai et al. (1996)
momentum
probe signal
Chaos analysis Evaluated by Kolmogorov Pressure, light, Quantification of the Schouten and van
entropy, correlation -ray porosity, dynamics and phase den Bleek (1992),
dimension, Lyapunov hot wire prove transition in the BFB, Marcocchella
exponent, presence of the signal, etc. CFB, mixing chamber, etc. et al. (1997),
chaotic attractor, Hurst Ji et al. (2000),
coefficient, etc. Huilin et al.
(1995)
Briens et al.
(1997)
Wavelet Analysis of time-varying cf. Table 15.13 cf. Table 15.13 cf. Table 15.13
transform signal simultaneously from
both time and frequency
perspective ð
1 1
Wx ð; aÞ ¼ pffiffiffi xðtÞððt Þ=aÞ dt
a 1
power spectrum in which different dominant frequen- signals having sharp spectrum peaks, the MEM gives
cies are observed. a sharper spectrum than FFT, as shown in Fig. 38.
Figure 32 Effect of averaging on the shape of signal: (a) first, (b) second, (c) Nth, (d) final. (Minami, 1986.)
Figure 34 Autocorrelation function and power spectrum. Upper row: original pressure signal; middle row: autocorrelation (no.
of lags 20 ¼ 0:45 s; sampling interval: 22.5 ms; data points: 1000; Dt ¼ 0:1 m, u0 ¼ 0:1 m/s; bed material: silica sand, dp ¼ 0:183
mm); bottom row: power spectrum. (a) Smoothly fluidized bubbling bed ðL=Dt ¼ 1); (b) slugging bed (L=Dt ¼ 2). (Broadhurst
and Becker, 1976.)
Figure 36 Power spectrum density of pressure fluctuations in a bubbling fluidized bed of sand (van den Schaaf et al., 1999a). Dt :
0.80 m; u0 : 0.44 m/s; dp : 390 m; settled bed height ¼ 2:19 m. (With permission of American Society of Mechanical Engineers)
ð1
ing the riser surface by a conductive material (Myler et
ðt Þ dt ¼ 0 ð32-1Þ
1 al., 1986), or by covering the inside wall with cellophane
ð1 tape (Horio and Kuroki, 1994; Tsukada et al., 1997),
2
ðt Þ dt ¼ 1 ð32-2Þ the adhesion trouble can be avoided.
1
Observation of a high-temperature fluidized bed can
be easily done in a lab scale gold mirror furnace con-
The second exponential function in Eq. (29) is used
sisting of a quartz tube, a gold-coated Pylex tube, and
in ordinary Fourier transforms, where no window is
a coil heater between them. If temperature difference is
applied. In the Gabor wavelet transform a window
created in the bed by some reactions it can be well
around t ¼ is superposed by the first part of Eq.
observed from the outside as, e.g., done by Horio et
(29) (i.e., by the error function). On the other hand,
al. (1986c). To observe the inside of a hot fluidized bed,
in the classical Gabor transform, not the Gabor wavelet
an endoscopic system can be used (Zevenhoven et al.,
transform, a windowed Fourier transform with a fixed
1999b). More information can be found in Horio et al.
width window is applied, as illustrated in Fig. 39b.
(1980b), Yang and Keairns (1982), Prins (1987), Yang
and Chitester (1988), Takeuchi and Hirama (1991),
3.2 Visual Observation Zou et al. (1994), and Hull and Agarwal (1995).
3.2.1 Photo and Video Imaging
3.2.2 Laser Imaging
For visual observation through a column wall of a cold
test rig, particle adhesion to the wall by static electricity At low solids concentration the laser sheet imaging
should be reduced. By putting charge absorbers into the shown in Fig. 40a is useful in observing fluidized
bed (Ilias et al., 1988; Chang and Louge, 1992), by coat- beds. However, the original images on the light source
side are brighter than those on the other side, owing to respectively, which are located at a distance r from the
the expansion of laser beam. This can be corrected in rod lens. L 0 ¼ Lr0 =r gives the corrected scale, where r0
the following manner: suppose L and L 0 are the gray is the distance of a reference pixel in the image (cf. Fig.
scales of a pixel on the original and modified picture, 40b).
Type of wavelet
transform Measurement Derived information Reference
Orthonormal wavelet Pressure Decomposition of the signals into long-term and short- Hee et al. (1997)
transform fluctuation term correlation components; one represented by the
self similarity Hurst’s parameter of fractional Brownian
motion originated by bubble, and another represented
by the intensity parameter of Gaussian white noise
originated by gas jetting, small bubble formation, and
turbulence of fluidized beds, etc.
Morlet wavelet Heat flux Self-similar bifurcation and trifurcation phenomena Ross and Pence
(1997)
Discrete wavelet Optical fiber Phase separation (identification of the transition from Ren and Li
transform based on probe signal the dense phase to the dilute phase) (1998)
Mallet and Zhang
Discrete analog of Pressure Signal denoising Roy et al.
wavelet transform fluctuation (1999)
(orthogonal wavelet
basis functions by
dilating and translating
in discrete steps
Mallat’s pyramidal Pressure Signal filtered by a wavelet; peak frequency and value Lu and Li
algorithm used for fluctuation corresponded to the bubble frequency and diameter, (1999)
computing one- respectively
dimensional orthogonal
wavelet transform
Gabor wavelet Signal from an Continuous transform with a Gabor wavelet [cf. Eq. (29)] Fukayama
acceleration applied to a commercial scale AFBC et al. (2000)
sensor
At higher solid fluxes, the core of the riser can be intensity I0 . According to the Lambert–Beer law, the
observed using a hood or an endoscope as shown in intensity of light after transmission through an object
Fig. 40c,d. Examples of laser sheet images are shown in of thickness L can be written as
Fig. 40e.
IL ¼ I0 eL ð33Þ
3.2.3 Particle Image Velocimetry (PIV) where is the mass attenuation coefficient, is the
medium density, and accordingly is the linear
If particle images are distinct in a time series of
attenuation coefficient of the medium the beam passes
visualized images, one can track the particle and
through. L is the thickness, and the term L is the
determine its velocity and trajectory. This is the
absorptance of the medium.
PIV technique. Rix et al. (1996) measured particle
Now, let us introduce cells from 1 to m in the objec-
movement in the freeboard using PIV. Reese and
tive cross section as shown in Fig. 42a. The cross sec-
Fan (1997) tracked local velocities of bubbles and
tion is scanned by number of prefix beams i n, and
particles in gas–liquid and gas–liquid–solid fluidized
the average passage length for cell j was Iij (for those
beds using a continuous mode argon ion laser as a
cells where beam i does not pass Iij ¼ 0). The first
light source, and a CCD camera for image record-
medium fraction in cell j expressed by "j is the main
ing. Assuming linear movement during the time
unknown variable. Then the total absorbance ILi =I0
interval between consecutive images, particles and
can be written as
bubbles were tracked and their velocities determined.
!
More information on this technique can be found in
ILi X
Chen and Fan (1992) and in Fan (1995). An experi- ¼ exp 1 1 "j þ 2 2 ð1 "j Þ lij ð34Þ
I0
mental setup used by Chen and Fan (1992) is shown j
in Fig. 41.
By rearranging the above expression we obtain
X
3.3 Computer Tomography (CT) and Image I
ð1 1 2 2 Þlij "j ¼ ln Li 2 2 Li ð35Þ
Reconstruction I0
j
Tomographic visualization is based on numerical Thus we have a set of linear algebraic equations to
solution of a set of linear algebraic equations corre- obtain "j :
sponding to the transmission intensity of high-
energy electromagnetic radiation such as x- or - ðaij Þð"j Þ ¼ ðbi Þ ð36Þ
rays or other beams (visible light, positrons, neu- where
trons, ultrasound, magnetic resonance), and/or field
parameters such as impedance, inductivity, or capa- aij ¼ ð1 1 2 2 Þlij ð37Þ
citance.
I
bi ¼ ln Li 2 2 Li ð38Þ
3.3.1 Image Reconstruction Principle I0
As for an example, let us suppose that for the construc- Now let us study a more general case in which the
tion of a tomographic image we use light of initial distribution of attenuation coefficients, described by a
ð 2 ð 1
function f ðx; y), is projected through a line L as 1 l
f ðx; yÞ ¼ 2
defined in Fig. 42b. The measured value Pðl; Þ ( is 4 1 l x cos y sin
the projection angle and l is the length, as in Fig. 42b)
0
ð40Þ
@Pðl; Þ
can be written
dl d
ð @l
Pðl; Þ ¼ f ðx; yÞ ds ð39Þ where ðl x cos y sin Þ is the perpendicular dis-
L
tance of the point ðx; yÞ from the line l.
When f ðx; yÞ is a continuous function with a con- In reality the reconstruction is always done as an
tinuous first derivative, Radon (1917) reported that approximation of Radon’s solution for a certain num-
f ðx; y) can be reconstructed from an infinite set of ber of projections. As shown in Fig. 42b, for each angle
line integrals Pðl; Þ by of projection , the value of Pðl; Þ depends on the
Figure 42 An algebraic approach to image reconstruction in CT (a) and the geometry for CT image reconstruction (b).
(Adopted from Kumar and Dudukovic, 1997.) With permission of Elsevier Science.
Weimer et al. (1985) -ray Monitoring the voidage along the height of a FB using Cs137 as
radiation source
Seville et al. (1986) -ray Studies on voidage distribution above the gas distributor in a
0.146 m i.d. fluidized bed of sand
Seo and Gidaspow (1987) -ray, x-ray Voidage measurement in a FB using simultaneous densitometry
Weinstein et al. (1984) x-ray Application of tomography in a CFB, using an optical densitometer
to analyze the images
Weinstein et al. (1992) x-ray Studies on the distribution of solid fraction in a fast fluidized bed
system (d ¼ 0:15 m)
Contractor et al. (1992) x-ray Observation of images from fluidized beds using x-ray source and
phototransistors
Kantzas (1994) x-ray Images of density distribution and gas holdup obtained for
fluidized bed of glass beads and polyethylene fluidized by
nitrogen
Holoboff et al. (1995) x-ray CT scanner used to study variability of gas voidage in a
polyolefin/air fluidized bed of 0.1 m i.d
Durand et al. (1995) x-ray Imaging technique combined with video camera to study the
hydrodynamics of a polyethylene FB
Fiorentino and Newton x-ray Identifying of scale-up issues for predicting large-scale BP reactor
(1998) performance
Grassler and Wirth x-ray Application of a 60 kV CT scanner to a cold 0.19 m i.d., 15 m
(1999) high CFB, results obtained with 0.2 mm resolution
Kai et al. (2000) x-ray Dynamic imaging of cross-sectional voidage distribution in every
4 ms by a fast scanning system consisting of 18 x-ray sources
and 122 detectors
Huang et al. (1989) Capacitance Studies on voidage distribution in the fluidized bed
Halow and Nicoletti Capacitance Studies on voidage distribution in the fluidized bed, bed diameter
(1992) 0.15 m, sensing electrodes installed round the riser in four rings,
each containing 32 electrodes
Dyakowski et al. (1997) Capacitance Application to CFB, good agreement with comparison of voidage
calculated from pressure transducers
Rhodes and Wang (1999) Capacitance Studies on distribution of solids volume fraction in a 0.09 m i.d
riser, four sets of sensors applied, each sensor contained 12
electrodes
tion (e.g., Tanino et al., 1993). There are three different real images taken by a CCD camera in front of the
methods of monitoring the granule size distribution. In probe. The measurement system consists of illumina-
the easiest method, the granules are sampled by com- tion, purging, and telescope devices, and the image
pressed air and captured on the adhesive tape. Then processing for determining the mean particle size can
the particle image on the adhesive tape is taken with a be done relatively fast. More information is given, e.g.,
CCD camera, and finally the sampling tube is cleaned in Fluidization Handbook (1999).
by air, the spent part of the adhesive tape is wound up,
and new samples are sucked into the tube. In another
system, laser scattering of sampling frequency up to 50 4 FLUIDIZED BED DIAGNOSTICS
Hz is applied for size determination. Since the light
scattered by the particles in the measurement volume Combined with the proper knowledge of fluidization
has a scatter angle unique for the particle diameters, systems, the instrumentation and measurement tech-
the particle size can be determined by a coaxial multi- niques can provide us with much meaningful informa-
circular detector. Another type of instrument provides tion. Since information in other chapters should be
lbi ¼ ubi Ti ð47Þ Since ðDb ; lb Þ has to satisfy the normalization condi-
tion, we can write
However, lbi does not immediately give the bubble dia- ð Db
meter. To obtain the real bubble diameter, Db , from a ðDb ; lb Þdlb ¼ 1 ð49Þ
series of data for lbi , let us first assume (as a simplest 0
case) that the bubbles are spherical and uniform in Thus the time averaged pierced length is given by
size and that they are uniformly distributed radially ð Db
across the bed. The probability, ðDb ; lb Þ dlb , of 2
lb ¼ ðDb ; lb Þlb dlb ¼ Db ð50Þ
having a pierced bubble length within the interval of 0 3
lb lb þ dlb for bubbles of diameter Db is 8rjdrj=D2b ,
Finally by rearranging the above expression, the bub-
where the radius, r, is defined by r2 ¼ ðD2b lb2 Þ=4
ble size is given by
(accordingly we have r dr ¼ lb dlb =4Þ. Then, since
8rjdrj=D2b ¼ 2lb dlb =D2b , and since lb is less than Db , PN 3 !1=3
3 i¼1 lbi
the probability distribution density function for bubble Db ¼ ð51Þ
2 N
size ðDb ; lb Þ is given by
where N is the number of bubble samples.
2l
ðDb ; lb Þ ¼ b2 for 0 < lb < D b ð48-1) For a more general case, where the bubble size is
Db distributed (Tsutsui and Miyauchi, 1979), let us
assume a distribution density function for bubble
and diameter, ðDb Þ. In this case, ðDb ; lb Þ introduced
above in Eq. (48) becomes a conditional probability
ðDb ; lb Þ ¼ 0 for lb < 0 or Db < lb to get a pierced length lb from a bubble of size Db .
ð48-2Þ What we need is the probability density distribution
Figure 47 Type of signal from capacitance probes. (Lanneau 1960.) With permission of the Institution of Chemical Engineers.
where ðlb Þ is measured and ðDb ; lb Þ can be calcu- 4.2.2 Solid Circulation Rate
lated for a bubble of particular shape. Then the aver- Overall solids mass flux is one of the key parameters to
age bubble size is given by determine the regime of circulating fluidization. It also
ð 1 1=3 has a great influence on the pressure profile along the
3
Db ¼ ðDb ÞDb dDb ð57Þ riser, as well as voidage and mixing. Table 15 sum-
0
marizes the previous methods of measuring overall
If bubbles are not spherical, the same procedure solids mass flux.
can be applied by changing ðDb ; lb Þ. Cluster size mea- Solids circulation rate can be measured directly at
surements can be done on the same principle, but they the downcomer by devices such as those shown in
have a more complicated structure. Fig. 48. Then the overall solid mass flux is obtained
by
4.1.5 Solid Circulation and Mixing ws
Gs ¼ kg=m2 s ð59Þ
Solid tracers and probes discussed in Secs. 2.11, 12 A
can be applied to determine solid circulation and mix- where ws is the solids mass flow rate and A is riser cross
ing. If real solids sampling is conducted, bed sectio- section.
nizing is a useful tool—after a certain fluidization The solids circulation rate can also be determined
period the gas is cut off with a sequencer and the from the observed local time-averaged solids velocity,
partitioning device is dropped, so that it can quickly vp , and solids concentration, (1 "), by the following
reach the bed bottom. Then the bed particles are integration over the bed cross section:
sampled section by section by a vacuum cleaner ðð
(Horio et al., 1986a). Ws ¼ p ð1 "Þvp dA ð60Þ
A
4.1.6 Gas Sampling from the Bed
If the local pressure is measured in parallel with
To sample gas from fluidized beds it is important to overall solids circulation, the cross-sectional average
know if the sample is taken from the bubble phase, solids concentration (1 ") can be determined know-
Installation of a butterfly Lasch et al. (1988) Measurements based on periodical opening/closing of the valve(s)
or slide valve(s) in the Bader et al. (1988) Difficult to apply to hot FB due to valve requirements and its
downcomer Hartge et al. (1988) damage possibility at high temperatures
Beaud and Louge (1995)
Disc valve Ye et al. (1999) Measurements of solids mass flux in a CFB of cast steel
Three-way valve Arena et al. (1988, 1993) Measurements based on switching over the three-way valve
technique (placed outside the hot reactor) and accumulating solids in a
vessel through the certain time period. No continuous
measurements possible
Two ring sensors Dybeck et al. (1995) Measurement method based on effect of electrostatic induction of
installed in the CFB charged particles on a surrounding metallic ring
recirculating line Qi and Farag (1995) Measurements of solids fluxes in a 0.14 m i.d. CFB riser of glass
beads and FCC. The solids were collected in each probe for a
given time period and then weighed. Local particle concentration
was measured using the capacitance probe
Study on five ways of Burkell et al. (1988) Study on application of closing valve in the return leg, observation
solids mass flux of an identifiable particle, using a device to record the force
measurement systems imparted by returning solids form the cyclone, measuring the
pressure drop across the constriction in the return loop, and
estimating solids mass flow from the heat balance on a
calorimetric section in the standpipe
Semicontinuous solids Horio et al. (1992b) Application of load cell. The solids were accumulated in the device
mass flux meter through a certain time period. Difficult application to hot units
due to load cell requirements
Patience et al. (1990) Idea based on measurements of pressure drop and gas velocity.
Assuming that gas–solid slip velocity is related to the terminal
velocity of individual particles, the mass flux was calculated.
Davies and Harries (1992) Idea based on application of weighing chamber and an
electronic balance
Kobylecki and Horio Scoop-like device measuring continuously the mass fluxes of
(2002) various solids including group C
Whether a bed is completely fluidized or not can be The agglomerate size of Geldart’s group C powders
examined from the bed pressure drop at the velocity can be roughly determined by sieve analysis. In such
slightly higher than the minimum fluidization velocity. cases, Ro-tap sieving may not be advisable because its
If it is, the pressure drop should equal the bed weight action also helps agglomeration or attrition. If direct
divided by the bed cross section when gas density is determination of agglomerate size is critical for opera-
sufficiently low that the buoyancy force acting on par- tion, as, e.g., in spray granulation, the in situ particle
ticles can be neglected (cf. P in Fig. 44). size determination methods already introduced in Sec.
If the particles have a wide size distribution, the 3.4 should be helpful.
pressure drop vs. gas velocity curve shows a rather
vague transition from a fluidized condition to the 4.3.3 Expansion Characteristics
fixed bed condition, as already shown in Fig. 44b.
High bed expansion is an important feature of powders
From this figure one can obtain gas velocities corre-
suitable for good gas–solid contacting. To quantify the
sponding to the beginning of fluidization ubf and to the
expansion characteristics, the bed contraction mea-
completion of fluidization ucf .
surement test is recommended as a standard method
in the laboratory. In this test the gas supply to the bed
4.3.2 Agglomerating Behavior of Cohesive Powders is stopped after a certain fluidization period and the
change of bed height is recorded. During the measure-
As shown in Fig. 49, channel formation and its col-
ments some attention has to be paid to the effects of
lapse and transition to the agglomerating fluidization
plenum chamber volume and the distributor resistance
can be detected from the cyclic response of the bed
on the result. Figure 50 shows the observed bed con-
pressure drop. However, it should be noted here that
traction curves from which the emulsion phase voidage
in the cohesive powder fluidization the wall effects are
can be obtained.
quite important, and the scale effect related to them
For Geldart’s group A powders, the bed expands
cannot be neglected.
homogeneously even above the minimum fluidization
Very fine particles are fluidized forming agglomer-
velocity, umf , until the minimum bubbling velocity umb .
ates. Minimum fluidization velocity of such agglomer-
The umb is another good index of high emulsion phase
ates ua;mf can be measured in the same manner as in the
expansion, as shown in Fig. 51. This figure also illus-
case of noncohesive particles, i.e., from the decreasing
trates how umb is determined.
velocity period (period after #7 in Fig. 49). A particle
recycling device, such as a cyclone or a filter, is neces-
4.3.4 Regime Transition Velocities
sary for fine powder experiments. From ua;mf , the
apparent agglomerate size da can be determined from In addition to umf and umb , other regime transition
minimum fluidization velocity correlations such as velocities such as uc , uk , and utr should be important
Wen and Yu’s (1966) by using the particle density for high-velocity operation. Figure 52 shows the pres-
separately determined by mercury porosimetry or sure fluctuation response to gas velocity change, from
other methods. which uc and uk are determined. These characteristic
transition velocities should be determined with a col- are recommended to check Horio (1997, pp 34–36). A
umn where slugging can be avoided. Furthermore, to careful evaluation of the regime transition shows that
avoid criticism as for the early studies (Geldart and uc , uk , and utr are independent and unique character-
Rhodes, 1986), the axial solid holdup should be istic velocities.
always monitored, so that the dense bed is kept in
the column, preventing the pressure tap from expo- 5 SUMMARY
sure to the lean phase, which causes the decrease in
pressure fluctuation at higher gas velocities and gives Instrumentation and measurement techniques for
fictitious values for uc and uk . For this issue readers fluidized bed processes are reviewed, starting from
Figure 50 Fluidized bed collapse of group A powder (glass ballotini, 26 m), reported by Geldart (1986) (# John Wiley & Sons.
Reproduced with permission) (a), and collapsing of fluidized bed reported by Tung and Kwauk (1982) (b).
fundamental knowledge on sensor elements, probe image analysis systems. Data to evaluate sensing sys-
configuration, and tracers to more sophisticated tems are included as much as possible for readers’ con-
venience. In the last part of the chapter a practical
approach to the diagnostics of fluidization systems as
well as powders are discussed.
ACKNOWLEDGMENTS
REFERENCES
Norman Epstein
University of British Columbia, Vancouver, British Columbia, Canada
operating modes are possible under the rubric of 2.2 Buoyancy and Drag
liquid–solid fluidization.
We shall also first consider nonporous fluidized par- Consider now a representative single particle of
ticles that are all uniform in density, size, and shape. In volume V and density rp in such a fluidized swarm of
their pioneer study of individual particle motion in monodispersed particles. Macroscopically steady-state
such a liquid-fluidized bed, Handley et al. (1966) fluidization requires that downward gravitational
found that, in the absence of gulf streaming, ‘‘uniform force ¼ upward buoyant force þ upward drag, i.e.,
fluidization’’ resulted and was ‘‘characterized by a
Vrp g ¼ B þ FD ð1Þ
homogeneous random motion of the particles and by
a constant mean fluid-velocity and momentum distri- There is a major disagreement in the literature on the
bution throughout the bed,’’ but with larger particle correct expression to use for the buoyancy B. The con-
velocity fluctuations (viz. axial/radial 2.3, as later ventional formula for B advocated, for example, by
verified by Carlos and Richardson, 1968a; Latif and Clift et al. (1987), Jean and Fan (1992), and Clift
Richardson, 1972; Kmiec, 1978) and displacements in (1993), is simply the buoyant force under static (no-
the vertical direction than in the horizontal, i.e., flow) conditions,
Dia
Dir . We shall therefore begin by assuming,
after Couderc (1985), that ‘‘liquid–solid fluidization B ¼ Vrg ð2Þ
results in stable operation and beds which are homo- The counterargument, presented among others by
geneous, with a spatially uniformly distributed concen- Gibilaro et al. (1984, 1987) and Astarita (1993), posits
tration of solid particles.’’ This idealized assumption of that
‘‘particulate fluidization’’ will be subsequently sub-
jected to significant qualification. B ¼ VrB g ð3Þ
Figure 2 Frictional pressure drop as a function of liquid superficial velocity for monodispersed particles.
ð15Þ
in which the product of sphericity and equivolume
sphere diameter dp has been substituted for the equiva-
lent surface-volume sphere diameter dsv. Algebraic
Figure 4 Bed voidage as a function of liquid superficial manipulation and rearrangement of this equation
velocity. results in
which, for particles of fixed shape, is the equivalent of 4 FLUIDIZED BED EXPANSION
the surface-volume mean diameter. Figure 6 would
also apply to particles of mixed shapes and densities 4.1 Introduction
(as well as mixed sizes), for which estimation of Ubf
and Utf would require that the particle properties in Crucial to the design of a fluidized bed is a quantitative
Eq. (17) be those of the particles with the lowest and knowledge of the bed expansion as a function of the
the highest individual values of Umf , respectively, while liquid superficial velocity, i.e., of either L or e [the two
estimation of Uamf would require that rp in Eq. (17) be being interrelated by Eq. (13)] as a function of U. The
the mean particle density, chaotic behavior of a fluidized bed, in contrast, say, to
viscous flow through an ordered (e.g., simple cubic)
X
N
array of immobilized spheres, renders measurements
Mi
from the former inherently less reproducible than
i¼1
rp ¼ ð23Þ from the latter.
X
N
ðMi =rpi Þ The equations that have been developed for predict-
i¼1 ing the frictional pressure drop due to flow through a
fixed, including an expanded, packed bed (of which
In the case of an upwardly diverging tapered or conical there are many in the literature) yield a relationship
bed (e.g., y ¼ 58308) of uniform particles, the bottom between dpf =dz, e, and U that, when combined with
of the bed is fluidized before the top because of its Eq. (8) to eliminate dpf =dz, give an equation relating
higher liquid velocity (Kwauk, 1992). The result is e to U. Because, as mentioned earlier, particles in a
that, on increasing the velocity through the initially fluidized bed, in contrast to those in a rigid array,
fixed packed bed, a very much higher and sharper are relatively free to move around and arrange them-
peak pressure drop is reached than in Fig. 2 before selves so as to minimize or at least decrease the resis-
the pressure drop falls again and full fluidization is tance to flow (Richardson and Meikle, 1961), therefore
achieved (Peng and Fan, 1997). The plot of frictional for a given value of e, U for a fluidized bed will usually
pressure drop vs. liquid velocity then resembles that of be larger than U for a fixed bed under the same pres-
a gas-spouted bed (Mathur and Epstein, 1974), as do sure gradient. Consequently, the use of a packed bed
the corresponding fluid–particle mechanics, and, as in equation will in most cases underestimate U for a given
the case of both fluidization and spouting, reducing the e, or overestimate e and L for a given U (Happel and
velocity produces a hysteresis effect whereby there is no Epstein, 1951). If, for example, we remove the mini-
longer (or at most barely) a discernible maximum in mum fluidization (subscript mf) restriction from Eqs.
the plot of pf vs. U. Peng and Fan (1997) identified (15)–(19), then Eq. (17) becomes
for Re0 < 35, as the value given by Eq. (36), i.e., unity ¼ en ¼ ð1 cÞn ð42Þ
U0
in the absence of wall effect. For Re0 > 1000, it is not
known whether the trend of k with Re0 shown by Eq. which for dilute solutions is well approximated by
(37) continues, but Eq. (36) is still presumably applic-
U
able to account approximately for any wall effect. ffi 1 n0 c ð42aÞ
U0
It is possible in many cases to reduce the uncertainty
associated with an assumed value of k by integrating Equation (42a) is of the same form as Eq. (25) above
Eq. (30) with the boundary condition U ¼ Umf at derived by Batchelor (1972) for dilute suspensions in
e ¼ emf, so that viscous flow (Re0 0:2). Note that as Re0 decreases,
n
U e the value of n 0 in Eq. (41) appears to be approaching
¼ ð39Þ the 6.55 of Eq. (25), and a value of n 0 6:5 was actu-
Umf emf
ally found by Di Felice and Parodi (1996) in some of
Dividing Eq. (31) by Eq. (38) in order to solve for k, we their viscous sedimentation experiments. In the pre-
find sence of a measurable wall effect (dp =Dc > 0:01), the
Umf intercept of the dilute suspension line is U00 , which for
k¼ ð40Þ viscous flow (Di Felice and Parodi, 1996) is related to
U0 enmf
U0 by either of the two empirical equations proposed
If emf and Umf are known or determined by the by Francis (1933) to describe the settling velocity of a
methods described in the previous section, and n is ball falling axially through a liquid in a cylindrical
determined from Eq. (34), then k can be estimated vessel. The simpler of the two equations, which give
from Eq. (40). Agreement, say, within 5% of a pre- almost identical results up to dp/Dc ¼ 0.3, is
vious estimate should bolster confidence in that
estimate. Disagreement, however, even assuming that U00 dp 2:25
emf, Umf , and n are correct, could be due to an unsa- ¼ 1 ð43Þ
U0 Dc
tisfactory original estimate of k, but it could also be
due to the fact that the primary straight line relation- Equation (43) yields answers very close to those
ship between log U and log e sometimes starts at a derived theoretically by Ladenburg (1907) and Faxén
value of e somewhat in excess of emf (Wilhelm and (Emersleben and Faxén, 1923) up to dp =Dc ¼ 0.1 and
Kwauk, 1948; Richardson and Zaki, 1954). 0.2, respectively, and agrees fortuitously also with
The critical voidage, ec, above which there is a the value of kð¼ Ue =U0 Þ given by Eq. (35) up to
change of slope in the plot of log U vs. log e for con- dp =Dc ¼ 0.2. The agreement of U00 by these equations
centrated suspensions, varies between 0.85 and unity with the dilute suspension intercept at e ¼ 1 has been
when the slope change is positive, e.g., at Re0 ¼ 50– demonstrated by Garside and Al-Dibouni (1977) for
1000, according to Rapagnà et al. (1989), and between dp =D up to 0.06 and Re0 up to 1200, i.e., well beyond
0.85 (or even 0.9, Chong et al., 1979) and 0.7 when the viscous flow. The corresponding equation of Munroe
change of slope is negative, which occurs at Ar 1600 (1988–89), obtained by fluidizing single spheres in a
(Re0 35) according to Fan et al. (1985). Most liquid cylindrical vessel in the range of Re0 ¼ 150–3400,
fluidization applications occur at e < ec. shows a much smaller wall effect. An empirical equa-
Referring to Fig. 7 it can be seen that, discounting tion for U00 =U0 proposed by Di Felice (1996), which
those situations in which the slope of log U vs. log e spans the whole range of Re0 from the Stokes to the
remains unchanged all the way to U ¼ Ue ¼ U00 (i.e., Newtonian region, including the data of Fidleris and
ec 1), the region in which e exceeds ec bifurcates on Whitmore (1961) in the intermediate region, is
testing, qualification and elaboration. Note that a high From Eq. (34), in the Newton regime (85,000
value of n in Eq. (32) is accompanied by a high value of < Ar < 1:4 1010 , 500 < Re0 < 200,000), n ¼ 2:4.
K in Eq. (46), in which the value of n is that from Eq. Therefore, again assuming k ¼ 1 in Eq. (32), it follows
(34) for unflocculated spheres, i.e., the lowest possible that
for the given Ar or Re0.
The increase of n as dp decreases and hence as spe- FDS =FD0 ¼ ðe2:4 Þ2 ¼ e4:8 ð53Þ
cific surface increases could in part be due to the elec- for the high Reynolds number regime. Since
trokinetic Dorn effect, which retards sedimentation in
ionic liquids (Tackie et al., 1983; Masliyah, 1994), and FD ¼ Vðrp rÞg ð6Þ
is most certainly associated with the flocculation of it follows that FD ¼ FD0, the drag force at the terminal
colloidal particles (Dixon, 1977). Equation (46) is velocity being balanced by the same gravitational force
again applicable to colloidal suspensions, with K now modified by the same buoyancy as given by Eq. (6).
denoting the volume of the flocs divided by the volume Therefore
of the particles contained by the flocs, and U0 the
terminal free settling velocity of an average floc rather FD =FDS ¼ f ðeÞ ¼ e4:8 ð54Þ
than of the particles within the floc (Michaels and at both ends of the Reynolds number spectrum. Wen
Bolger, 1962). The values of K in Eq. (46) and of n and Yu (1966) assumed that the same voidage func-
in Eq. (32) are now typically an order of magnitude tion, which they took as e4:7 instead of e4:8 , would
greater than for larger nonagglomerated particles also apply to intermediate values of Re0. Therefore
(Dixon, 1977).
FD C
Another approach to bed expansion is based on a ¼ D ¼ e4:7 ð55Þ
postulate by Wen and Yu (1966), anticipated by the FDS CDS
drag force measurements of Rowe and Henwood But for Re < 1000 (Schiller and Naumann, 1933),
(1961), Rowe (1961), and Richardson and Meikle
24
(1961b) on immobilized beds of spheres at various voi- CDS ¼ ð1 þ 0:15 Re0:687 Þ ð56Þ
dages and Reynolds numbers. This postulate states Re
that the drag force FD on a given sphere in a fluidized Substituting Eq. (56) into Eq. (55) and rearranging,
swarm of monosized spheres to the drag force FDS on a
single isolated sphere past which the same fluid flows at CD Re2 e4:7 ¼ 24Re þ 3:6Re1:687 ð57Þ
where eS and eL are the voidages displayed by the 1. dL/dS, i.e., in size difference, as would be
respective monocomponent beds when they are each expected.
separately fluidized by the same liquid at the same 2. "L,, i.e., in bed expansion, as has been fre-
superficial velocity, U, as the binary bed. These voi- quently observed (Wakeman and Stopp, 1976;
dages can be separately measured or they can be pre- Juma and Richardson, 1979; Epstein and
dicted by the bed expansion equations elaborated Pruden, 1999), though not by Garside and
earlier. Assuming no major disturbances due to bulk Al-Dibouni (1974), who found this trend to
circulation or instability effects, the value of g can be interrupted by strong mixing in the voidage
function as a crude qualitative measure of the degree range of e ¼ 0.65–0.80.
of particle segregation as follows (Epstein and Pruden, 3. ð3 mÞ=mn, which is roughly equivalent to
1999): g < 0:015 0:005: little segregation, good par- ð3 b=nÞ=b, where n is given by Eq. (34) and
ticle mixing (Fig. 8b). 0:015 0:005 < g < 0:045 b by Eq. (65), and varies from 0.43 to 0.22 as
0:015: partial segregation with no interface between one moves from the Stokes to the Newton
layers (Fig. 8d). 0:045 0:015 < g 0:1: segregation regime. Thus as the extent first of inertial
with fuzzy interface and some intermixing (Fig. 8c). effects and then of turbulence increases, the
g > 0:1: clean-cut segregation with sharp interface degree of segregation decreases (Epstein and
and little intermixing (Fig. 8a). The last criterion is Pruden, 1999), even without complicating effects
consistent with the generalization of Di Felice (1995) such as bulk circulation and hydrodynamic
that ‘‘complete segregation is to be expected when the instability.
size ratio is greater than about two.’’ The fuzzy inter- Another approach to the problem of binary particle
face that occurs with the third criterion (Fig. 8c) is a stratification is based on the more traditional concept
transition zone, the depth of which extends over the of axial dispersion according to Fick’s law of diffusion
whole bed length for the second criterion (Fig. 8d). in competition with segregation, as formulated by
The implications of Eq. (79) can be made manifest Kennedy and Bretton (1966). Assuming semibatch
by writing the Richardson–Zaki (1954) equation, Eq. liquid fluidization that results in partial segregation
(32) above, for the two monocomponent beds repre- of the type represented by Fig. 8c or 8d, then at any
sented by Eq. (79), so that bed level within the partially mixed region, the volu-
U ¼ kS U0S enSS ¼ kL U0L enLL ð80Þ metric particle mixing flux for species i must equal the
particle segregation flux for that species, i.e.,
But in general (Pruden and Epstein, 1964),
Di dci
U0 / dpð3mÞ=m ¼ a0 dpð3mÞ=m ð81Þ ¼ ci Upi ð85Þ
dz
rp ¼ constant
pffiffiffi dp ¼ constant pffiffiffi
dL =dS ¼ 2 ðrph rÞ=ðrpl rÞ ¼ 2
eS eL g el eh gh
2 !1=mn 3
rpl r rpl r rpl r respectively. For the larger, less dense particles (species
gh ¼ 1 el 4 5 lL) subjected alone to liquid fluidization,
rph r rph r rph r
rBlL ¼ rplL ð1 eÞ þ re ð112Þ
ð111Þ
The voidage e can be expressed in terms of the super-
Since mn equals 4.8 both in the Stokes and the Newton ficial liquid velocity U, explicitly by means of the
regimes and is in any case always in excess of unity, it Richardson–Zaki relation, Eq. (32), or by Eq. (64)
follows that the expression in square brackets is always with (65), and implicitly (requiring iteration) by
positive, so that gh (in contrast to g) always increases means of Eq. (62) with (63), applied to the given
as the voidage decreases. Thus we can explain the solid–liquid system. A plot of rBlL vs. U from UmflL
empirical rule, long known in the ore dressing industry to U0lL can thus be generated. Similarly, for the
(Gaudin, 1939), that classification by size, i.e., sizing, smaller, denser particles (species hS) subjected alone
is best performed under dilute (‘‘free settling’’) con- to fluidization by the same liquid at the same temp-
ditions, while classification by density, i.e., sorting, erature and pressure,
prevails under concentrated (‘‘hindered settling’’ or
‘‘teeter bed’’) conditions. We shall see below how this rBhS ¼ rphS ð1 eÞ þ re ð113Þ
rule manifests itself in the bed inversion phenomenon.
Note also in Table 2 [or in Eqs. (84) and (111)] that, Again, e can by the same means as for Eq. (112) be
in contrast to segregation by size, which is attenuated expressed in terms of U, so that a plot of rBhS vs. U
as one moves from the Stokes to the Newton regime, from UmfhS to U0hS can then be obtained. If the two
i.e., as Re0 increases, there is no such effect in the case curves do not intersect between UmflL and U0hS, then no
of segregation by density. bed inversion will occur, and the component that has a
Axial dispersivities for particle species of different consistently higher bulk density will always tend to seg-
density, whether or not the sizes are the same, can be regate toward the bottom
of the bed. The reduced bulk
determined by methods similar to those for fixed den- density difference, rB =ðrplL rÞ, then, provides an
sity, mixed size binaries, but their experimental deter- indication of the degree of segregation in accord with
mination has received much less attention in the g ¼ ðrBL rBS Þ=ðrp rÞ for pure sizing (Epstein and
literature. Equation (103), to the extent that it is Pruden, 1999). If, however, the two curves do intersect
applicable at all, is restricted to binaries of equal between U ¼ Umf1L (> UmfhS Þ and U ¼ U0hS (< U01L Þ,
density particles. as in Fig. 9, then bed inversion will occur.
The intriguing subject of bed inversion has received
much attention in the literature, both experimental
5.5 Binary Mixtures of Particles Differing in Size (Hancock, 1936; Cleasby and Woods, 1975; Van
and Density: Bed Inversion Duijn and Rietema, 1982; Moritomi et al., 1982;
Epstein and LeClair, 1985; Gibilaro et al., 1986a;
Consider a binary mixture of spheres of diameters dlL Jean and Fan, 1986; Di Felice et al., 1988; Qian et
and dhS (<dlL) and densities rplL and rphS (>rplL), al., 1993; Funamizu and Takakuwa, 1996) and theore-
tical (Van Duijn and Rietema, 1982; Patwardhan and followed, respectively, by the two additional segrega-
Tien, 1985; Epstein and LeClair, 1985; Moritomi et al., tion patterns (b) and (d) of Fig. 10.
1986; Gibilaro et al., 1986a; Jean and Fan, 1986; To gain further insight into this phenomenon, apply
Syamlal and O’Brien, 1988; Funamizu and Eq. (62) with (63) or, alternately, (64) with (65) to var-
Takakuwa, 1996; Richardson and Afiatin, 1997; Asif, ious binary mixtures of the two solids species fluidized
1997, 1998; Epstein and Pruden, 1999; Howley and by the given liquid, assuming after Wen and Yu (1966)
Glasser, 1999). The earlier work incorrectly assumed and Gibilaro et al. (1986a) that dp is the Sauter mean
that the inversion velocity was entirely independent of particle diameter given by Eq. (75), evaluating rp as the
the overall solids composition of the binary and that volume-average density by Eq. (74), and noting that
therefore inversion occurred at the intersection point C
on Fig. 9 (Hancock, 1936; Van Duijn and Rietema, e ¼ 1 c1 c2 ¼ 1 clL chS ð114Þ
1982; Epstein and LeClair, 1985). According to this
view, an idealized inversion phenomenon, i.e., one Assuming knowledge of r, m, dlL, rlL, dhS, and rhS,
not subject to flow disturbances that induce mixing, then for any given value of U, Eq. (62) or Eq. (64)
would show only the three segregation patterns (a), becomes a relationship between clL and chS. Thus
(c), and (e) of Fig. 10. In the light of further careful plots of chS vs. clL can be generated with U as a para-
experimentation (e.g., Moritomi et al., 1982; Jean and meter. From the families of chS and clL curves, one can
Fan, 1986), backed by theory (Gibilaro et al., 1986a), it determine plots of rB vs. clL (or rB vs. chS) by substi-
was found that the relative proportions of the two tuting Eqs. (74) and (114) into Eq. (4), so that
solids species in the binary influenced the inversion
velocity, and that this fluid velocity was preceded and rB ¼ clL rplL þ chS rphS þ ð1 clL chS Þr ð115Þ
For the lower values of U between UmflL and U0hS the bulk densities of the mixed bed at the inversion point,
plots of rB vs. clL rise monotonically over the entire depending on the overall solids composition, vlL. For
range of clL, while for the higher values of U between example, assuming an overall solids composition,
the same limits, they fall monotonically. However, for clL =ðclL þ chS Þ ¼ vlL ¼ 0:4, given by point B on Fig.
a range of intermediate values of U, the plots of rB vs. 9, then on increasing the liquid superficial velocity
clL show maxima (Gibilaro et al., 1986a), the locus of from UmflL to U0hS, we can observe five stages to the
which is represented by curve ABD on Fig. 9. (For inversion process:
consistency the monocomponent species lines are best 1. From U ¼ U mflL to U ¼ U A, the bottom layer
generated by the same equation, either (62) or (64), as contains pure solids species hS and the top pure species
used to generate ABD.) This curve represents the range lL, both layers expanding with increasing U.
of possible inversion velocities and the corresponding 2. From U ¼ U A to U ¼ U B, the expanding
bulk densities over which the bottom layer is a bottom layer composition is a mixture of the two
binary mixture of solids, the solids composition, species, with clL =ðclL þ chS Þ increasing from 0 to 0.4
clL =ðclL þ chS Þ, of which is given by the nonlinear and rB decreasing from rBA to rBB as U increases,
scale at the top of Fig. 9. It also represents the range while the contracting top layer continues to contain
of possible inversion velocities and the corresponding pure lL.
merge with the single-phase fluid flow line at U U0 . Nu / e0:79n1:66 ð1 eÞ0:12 ð173Þ
The increase in h for a liquid-fluidized bed relative to
single-phase liquid flow at the same value of U can be from which it follows that Nu is a maximum at e ¼
by as much as a factor of seven (Richardson et al., ð0:79n 1:66Þ/(0.79n 1.54). Thus for spheres in the
1976; Klaren and Halberg, 1980; Bremford et al., Stokes regime (Re0 < 0.2), n ¼ 4:8 and emax ¼ 0.95,
1996). while in the Newton regime (Re0 > 500), n ¼ 2:4
Using 2665 data points from 43 publications, and emax ¼ 0.66. Experimentally measured values of
different weighting procedures and different numbers emax generally fall between 0.62 and 0.82, with the lar-
of dimensionless parameters, Haid et al. (1994) and gest cluster in the vicinity of 0.73 (Jamalahmadi et al.,
Haid (1997) arrived at six empirical equations, of 1997). Note that Eq. (172) is inapplicable at e ¼ 1.
which the one with the lowest standard deviation Combination of Eqs. (172) and (32) is also equi-
(11.5%) and the smallest average relative error valent to
(32.0%), considerably smaller than that obtained for
the same data using 38 other proposed correlations h / dp1 Re0:72
0 dp
0:17
ð174Þ
from the literature, was
For spheres in the Stokes regime, Re0 ¼ Ar/18 / d3p ,
hdp d 1:33
Nu ¼ ¼ 0:0413Re0:79 Pr0:67 ð1 eÞ0:12 e1:66 ffiffiffiffiffiffiffiffi h /3=2
whence from Eq. (174) it followspthat p , while
l in the Newton regime, Re0 ¼ 3Ar / dp , whence
0:10 h / d 0:25
dp rp r 0:04 p . Thus the larger the particles the larger the
ð172Þ value of h, but the smaller the influence of dp on h.
Dh r
Both effects are in agreement with experimental results
where Dh is the hydraulic diameter of the column (or (Haid et al., 1994; Jamialahmadi et al., 1996).
annulus, in the case of a vertically immersed cylindrical Equation (172) also implicitly indicates that, at any
heater or cooler). The range of parameters (some inter- given bed voidage, h increases with particle density,
related, e.g., e depends upon Re for any given value of primarily because the value of Re required for expan-
Ar) covered by Eq. (172) was Re ¼ 0.020–9400, sion to the given voidage will increase with particle
Pr ¼ 1.65–7700, e ¼ 0.40–0.99, Ar ¼ 3.85–67,000,000 density and secondarily (and very slightly) through
and dp/Dh ¼ 0.0013–0.210. If Eq. (32) is combined the term (rp r=rÞ0:04 . This effect too is in agreement
with Eq. (172), then with experiment (Jamialahmadi et al., 1995).
Figure 14 Schematic diagram of a fluidized bed bioreactor. (After Grady Jr et al., 1999.)
K1
expanded freeboard section at the top of the conven-
cpp ¼ specific heat capacity of solid particles,
tionally operated column (Wright and Raper, 1996) or J
m
K1
at the bottom of the inverse bed (Nikolov and ct ¼ local or overall volumetric concentration
Karamanev, 1987). The elutriation problem is much in liquid of particle species i, m3
m3
attenuated or even eliminated by use of a tapered c ¼ saturation concentration in Fig. 12, wt%
fluidized bed (Scott et al., 1978). c c ¼ bulk liquid undersaturation for crystal
dissolution points in Fig. 12, wt%
c c ¼ bulk liquid supersaturation for crystal
ACKNOWLEDGMENTS growth points in Fig. 12, wt%
c20 ¼ volumetric concentration of smaller
Thanks are due to John Grace, Arturo Macchi, Dusko particles in pseudofluid through which
larger particles move, m3
m3
Posarac, and Yunbi Zhang for help in locating refer-
CA ¼ bulk concentration of solute A in liquid
ences, and to the Natural Sciences and Engineering solution at given bed level, kg
m3
Research Council of Canada for continuing research CAS ¼ concentration of solute A in liquid
support. solution at particle surfaces, kg
m3
CA1 ¼ value of CA at z ¼ 0, kg
m3
CA2 ¼ value of CA at z ¼ L, kg
m3
NOMENCLATURE ðCA Þ1:m: ¼ logarithmic mean of CAS–CA1 and CAS–
CA2, kg
m3
a ¼ exponent in Eq. (44), dimensionless CD ¼ drag coefficient of particle in bed fluidized
a; b; c ¼ exponents in Eq. (167), dimensionless at superficial velocity U, dimensionless
ai ¼ constant for particle species i in Eq. (89), CDS ¼ drag coefficient of isolated particle at
m
s1 velocity U, dimensionless
a0 ¼ coefficient in Eq. (81), mð2m3Þ=m
s1 CD0 ¼ drag coefficient of isolated particle at
a 00 ¼ constant in Eq. (176), dimensionless velocity U0, dimensionless
A ¼ cross-sectional area of column, m2 C 1 , C2 ¼ constants in Eq. (17), dimensionless
Ac ¼ portion of submerged surface not affected d ¼ diameter of spherical particle, mm or m
by particle contact, m2 davg ¼ average diameter of spherical particles as
Ap ¼ portion of submerged surface affected by given by Eq. (136), mm or m
particle contact, m2 deff ¼ effective particle diameter ¼ dp K 1=3 , mm
Ar ¼ Archimedes number ¼ dp3 ðrp rÞrg=m2 , or m
dimensionless di ¼ spherical particle diameter of species i
b ¼ coefficient in Eq. (108), mðmþ3Þ=m
kg1=m
particles, mm or m
s1 dp ¼ diameter of sphere having same volume as
bi ¼ constant for particle species i in Eq. (89), particle, mm or m
s1 dpi ¼ equivolume sphere diameter of species i
0
b ¼ exponent in Eq. (168), dimensionless particles, mm or m
b 00 ¼ exponent in Eq. (176), dimensionless dp ¼ local mass average or mass median
B ¼ buoyant force on single particle in swarm particle diameter, mm or m
of particles, N ds ¼ diameter of sphere having same surface
B1 ¼ parameter defined by Eq. (29b), area as particle, mm or m
dimensionless dsv ¼ diameter of sphere having same surface-to-
B2 ¼ parameter defined by Eq. (29c), volume ratio as particle, mm or m
dimensionless Da ¼ axial dispersion coefficient of liquid,
B 0
¼ surface mass flux/mainstream mass flux, cm
s1 or m
s1
dimensionless DA ¼ molecular diffusivity of solute A in liquid,
c ¼ volumetric particle concentration of cm2
s1 or m2
s1
monodisperse solids, m3
m3 Dc ¼ column diameter, m
ci ¼ local or overall volumetric concentration Dh ¼ hydraulic diameter ¼ 4 cross-sectional
in liquid of particle species i, m3
m3 area/wetted perimeter, m
mixture
K1
mf ¼ minimum fluidization
lB ¼ parameter defined by Eq. (29d),
S ¼ smaller particles of fixed density binary
dimensionless
mixture
leddy ¼ eddy conductivity of heat, W
m1
K1
1; 2; 3 ¼ particle species 1, 2, and 3, respectively
lp ¼ thermal conductivity of particles,
W
m1
K1
m ¼ liquid viscosity, Pa
s
m2 ¼ pseudofluid viscosity for determining
motion of larger particles through REFERENCES
suspension of smaller particles, Pa
s
r ¼ liquid density, kg
m3 Adler IL, Happel J. The fluidization of smooth spheres in
rB ¼ bulk density of liquid-fluidized bed, liquid media. Chem Eng Prog Symp Ser 58(38):98–105,
kg
m3 1962.
rBA, rBB, Agarwal GP, Hudson JL, Jackson R. Fluid mechanical
rBD ¼ bulk densities, (rB)max, of bottom layer at description of fluidized beds. Experimental investigation
points A, B, and D, respectively, in Fig. 8, of convective instabilities in bounded beds. Ind Eng Chem
kg
m3 Fundam 19:59–66, 1980.
d h
p 3 i
FI;g r d u Same as expansion stage
dt g 6 b b
ðt
3 2 pffiffiffiffiffiffiffiffiffiffiffi du=dt
FBA Not applicable d prl ml pffiffiffiffiffiffiffiffiffiffi dt
2 b 0 tt
p 2 p 2
FC D ð1 þ eÞes rs u2e d e r u2
4 o 4 b s s
ð ð ð
d h
p 3 i
d rm um dV
FI;m ¼z r d u Same as expansion stage
dt dt m 6 b b
Source: Luo et al. (1998a).
rffiffiffiffiffiffiffi
Also, the volume equivalent bubble diameter, db , is s
Dmax 2:53 ðfor a ¼ 0:21Þ ð23aÞ
related to a and a by grg
db in liquids, and
a¼p
3
ffiffiffiffiffiffi ð20Þ
8a rffiffiffiffiffiffiffi
s
Note that the centrifugal force is affected significantly Dmax 3:27 ðfor a ¼ 0:3Þ ð23bÞ
grg
by the gas density, the aspect ratio of the bubble, the
bubble size, and the bubble rise velocity. The bubble is in liquid–solid suspensions. Further, based on the
not stable if Fx is larger than Fs , that is, Davies–Taylor equation, the rise velocity of the max-
pffiffiffiffiffiffiffiffiffiffiffiffiffi imum stable bubble is
8a4=3 Eð 1 a2 Þ s
2
ub db ð21Þ !1=4
0:312 rg 1:6sg
umax ¼ ð24Þ
rg
When the centrifugal force is larger than the surface
tension force, the bubble would be stretched in the
The comparison of experimental maximum bubble
x-direction. During the stretching, the aspect ratio, a,
sizes and the predictions by various instability theories
becomes smaller while db and ub can be assumed to
is shown in Fig. 11. The internal circulation model can
remain constant. As a result, the centrifugal force
reasonably predict the observed pressure effect on the
increases, the surface tension force decreases, and the
maximum bubble size, indicating that the internal cir-
bubble stretching becomes an irreversible process.
culation model captures the intrinsic physics of bubble
Using the Davies–Taylor equation (Davies and
breakup at high pressures. The comparison of the pre-
Taylor, 1950) for the bubble rise velocity, the maxi-
dictions by different models further indicates that bub-
mum stable bubble size is expressed by
ble breakup is governed by the internal circulation
pffiffiffiffiffiffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffi
s mechanism at high pressures over 1.0 MPa, whereas
Dmax 7:16a2=3 Eð 1 a2 Þ1=2 ð22Þ the Rayleigh–Taylor instability or the Kelvin–
grg
Helmholtz instability is the dominant mechanism at
The simplified forms of Eq. (22) are: low pressure.
3 TRANSPORT PHENOMENA condition specific, and the flow regime map shown in
Fig. 12 is obtained in an air–water–1.5 mm glass bead
Bubble dynamics and characteristics discussed above three-phase fluidized bed. The fluidized bed is 8.26 cm
determine the hydrodynamic and heat and mass trans- in diameter and 2 m in height. This map puts the com-
fer behaviors in three-phase fluidization systems, which mon three-phase fluidization operating conditions
is important for better design and operation of three- represented by the dispersed bubble and coalesced
phase fluidized beds. In this section, various hydrody- bubble flow regimes in proper perspective.
namic variables and transfer properties in three-phase
systems are discussed. Specifically, areas discussed in
the hydrodynamics section are minimum fluidization, 3.1.1 Minimum Fluidization
bed contraction and moving packed bed phenomenon, The first topic of study for the hydrodynamics of three-
flow regime transition, overall gas holdup and hydro- phase fluidization is minimum fluidization. For a given
dynamic similarity, and bubble size distribution and gas velocity, the minimum liquid flow rate required to
the dominant role of larger bubbles. Later in this sec- fluidize a bed of particles (Ulmf) may generally be deter-
tion, important topics covering transport phenomena mined from the change in the bed dynamic pressure
will be discussed, which include heat and mass transfer drop behavior that occurs as the bed changes from a
and phase mixing. fixed bed to a fluidized bed. There are considerable
variations on minimum fluidization phenomena
3.1 Hydrodynamics among small/light, large/heavy, and mixed particle sys-
tems. However, Ulmf in general can be evaluated
Flow behavior of cocurrent upward gas–liquid flow to mechanistically by considering an intrinsic condition
a bed of particles is characterized by a complex inter- for minimum fluidization where the total pressure
action among gas, liquid, and solid phases. A typical drop over a bed of particles at the fixed state is equal
flow regime map for three-phase fluidized beds is to the total bed weight per unit bed cross-sectional area
shown in Fig. 12 (Zhang et al., 1997). Seven different as formulated by Song et al. (1989). In this evaluation,
flow regimes are identified in this map, which vary with the pressure drop in the fixed bed can be described by a
gas and liquid velocities. The regime map is operating flow model developed by Chern et al. (1983, 1984).
grm4l grd 2p
Mo ¼ Eo0 ¼
r2l s3 s
rl d p U l rp Ug
Re0 ¼ bd ¼ bU ¼ ð32Þ
ml rl Ul
Photograph C shows the flow pattern near the The effects of gas velocity and pressure on the
maximum in the heat transfer signal. The heat trans- surface-to-bed heat transfer coefficient in a three-
fer surface experiences high shear flow due to the phase fluidized bed are shown in Fig. 18 (Luo et al.,
primary-wake-induced upward liquid and solid flow 1997a). With an increase in pressure, the heat transfer
toward the wake central axis, causing enhancement coefficient increases, reaches a maximum at pressures
in the heat transfer. The velocity of gel particles of 6 to 8 MPa, and then decreases. The following cor-
near the heat transfer surface is roughly estimated relation can be used to predict the heat transfer coeffi-
from the streak of particles. The upward particle velo- cient in three-phase fluidized beds at high pressures:
city is in the range of 20 cm/s, which is close to the !
bubble rise velocity (24 cm/s), confirming that the par- 0 0:45 0:396 0:6768
h ¼ h eg þ ð36Þ
ticle is in the central wake region. Thus the maximum Ug0:45 upt;0
heat transfer rate is obtained along the wake central
axis in the primary wake region. Figure 17 clearly where h 0 is the heat transfer coefficient of a liquid–solid
demonstrates the importance of bubble wake beha- fluidized bed with the same solids holdup, and upt;0 is
vior on the heat transfer characteristics of three- the particle terminal velocity in the fluidizing liquid at
phase fluidized beds. ambient pressure. The units for Ug and upt,0 in the
equation are in m/s. The heat transfer coefficient, h 0 ,
can be calculated by the correlation (Richardson et al.,
3.2.3 Pressure Effect 1976):
It is known that pressure has a significant effect on the e0:38
hydrodynamics of three-phase fluidized beds, and it Nu 0 ¼ 0:67 Re0:62 Pr0:33 s
ð37Þ
1 es
would also affect the heat transfer characteristics,
because heat transfer behavior strongly depends on The heat transfer behavior between an immersed
the hydrodynamics of the system. The effect of pres- solid surface and surrounding bulk fluids in a slurry
sure on heat transfer is mainly through the variations bubble column (nitrogen-Paratherm NF heat transfer
in liquid properties and hydrodynamic parameters, as fluid—53 mm glass beads) at elevated pressures is
summarized in Table 2. The overall effect of pressure shown in Fig. 19 (Yang et al., 2000b). It is found
on heat transfer behavior depends on the outcome of that pressure has a significant effect on the heat trans-
the counteracting effects of each individual factor. fer characteristics in a slurry bubble column. The heat
Parametric effect
Effect of the Effect of pressure on heat transfer
parameter on heat increase on coefficient with
Parameter transfer coefficient parametric value increase in pressure
ml þ
rl þ þ (small) þ (small)
kl þ þ (small) þ (small)
Cpl þ þ (small) þ (small)
s No direct effect No direct effect
db þ
eg þ þ þ
þ: Increase; decrease.
Source: Luo et al. (1997a).
transfer coefficient decreases appreciably with increas- heat transfer rate in slurry bubble columns. The addi-
ing pressure. The variation in heat transfer coefficient tion of fine particles to the liquid phase enhances heat
with pressure is attributed to the counteracting effects transfer substantially, and the effect of temperature on
of the increased liquid viscosity, decreased bubble size, the heat transfer behavior is mainly determined by the
and increased gas holdup or frequency of bubble pas- change in liquid viscosity.
sage over the heating surface as the pressure increases. It is noted that the pressure effects on the heat
In slurry bubble columns, it is observed that the bubble transfer coefficient are different between large-particle
size reduces significantly with an increase in pressure, and small-particle systems (Luo et al., 1997a; Yang
especially under low pressures, which would result in a et al., 2000b). Similar observations were also found
decrease in the heat transfer coefficient. Therefore the for hydrodynamics and bubble characteristics
bubble size is the most important factor affecting the between large-particle and small-particle systems.
Luewisutthichat et al. (1997) photographically stud-
ied bubble characteristics in multiphase flow systems.
They found that large-particle systems (i.e., three-
phase fluidized beds) exhibit appreciably different
Figure 18 Heat transfer coefficients as a function of gas Figure 19 Effect of pressure and gas velocity on heat trans-
velocity at different pressures in a three-phase fluidized bed fer coefficients in a slurry bubble column (nitrogen-
(nitrogen-Paratherm NF heat transfer fluid–2.1 mm glass Paratherm NF heat transfer fluid–53 mm glass beads).
beads. (From Luo et al., 1997a.) (From Yang et al., 2000b.)
rg ðatmosphericÞ
decreases the bubble size. Therefore a significant
increase in the interfacial area at high pressures is
Equation (44) can be used to estimate the interfacial
expected, which would result in an overall increase of
area at high-pressure conditions based on the atmo-
kl a.
spheric data. However, this procedure needs to be
Typical results regarding the influence of pressure
further verified.
on the specific gas–liquid interfacial area in a bubble
Studies of the interfacial area in a packed-bed col-
column are shown in Fig. 20 (Oyevaar et al., 1991).
umn at elevated pressures indicate the insignificant
The interfacial area in the figure was determined by
effect of pressure for three-phase systems (Molga and
means of CO2 absorption into aqueous solutions of
Westerterp, 1997). The possible explanation for the
diethanolamine (DEA) in a column of 8.1 cm ID.
lack of pressure effect in three-phase fluidized beds is
The pressure has a positive effect on the interfacial
that the column packing controls the bubble coales-
cence and breakup process. Therefore the variation
of bubble size with pressure is relatively smaller with
packing than without. A comparison of interfacial
areas obtained in bubble columns with packing and
without packing indicates higher interfacial areas in
the packed-bed bubble column, which implies that
packing tends to break up the bubbles.
Paratherm liquids is shown in Fig. 26. In the simula- rising path, and induces a complicated wake flow field,
tions, an 80 mm (width) 50 mm (height) domain is which, in turn, affects the bubble formation behavior.
used with two orifices at the bottom, an exit at the top, From the above examples, the computational fluid
and two sidewalls as boundary conditions. The liquid dynamics simulation is able to capture the dynamic
initially fills the domain to a certain height. At the behavior of each individual phase in three-phase
beginning of the simulation, gas is injected into the flows, and it also demonstrates its great potential in
liquid through two orifices. Paratherm NF heat transfer other fields.
fluid and nitrogen are used as liquid and gas phases,
respectively. The gas inlet velocity is 14.0 cm/s. The
diameter of each orifice is 6 mm. To study the interac- 5 SUMMARY
tion between bubbles formed from different orifices, the
orifice separation distance is set as 20 mm. As can be Gas–liquid–solid three-phase fluidized bed systems
seen from the figure, due to the closeness of the orifices, have been widely used in many industrial applications.
the fluctuations of flow field induced by the bubble Successful application of three-phase fluidization sys-
forming from one orifice have an effect on the bubble tems lies in a comprehensive understanding of bubble
forming from the other orifice. This interactive effect dynamics, hydrodynamics, and transport phenomena,
yields varied bubble shapes and an alternate detach- particularly under high pressure and temperature con-
ment pattern of bubbles from two orifices. Initially ditions. High-pressure and high-temperature operation
the bubble from each orifice rises rectilinearly (Fig. of three-phase fluidized beds is commonly encountered
26a); after detachment, however, the bubbles tend to in most industrial applications of commercial interest.
move to the centerline, rise in zigzag fashion, and break The flow characteristics of reactors at high pressure
up at the free surface (Figs. 26d–f). Based on simulation and temperature are distinct from ambient conditions;
results, it is found that owing to the closeness of the for example, elevated pressure leads to higher gas
orifices, two forming bubbles induce a high liquid velo- holdup and smaller bubble size in the system, and
city in the area near the centerline. The high velocity thus dramatically affects transport phenomena includ-
thus results in a lower pressure in this area and draws ing heat and mass transfer, and phase mixing. The
bubbles toward the centerline. In other words, bubble effects of pressure and temperature on fluid dynamics
and bubble wake interaction leads to a zigzag bubble and transport properties are mainly in turn due to
variations in bubble characteristics and changes in the To date, the design of three-phase fluidized beds still
physical properties of fluid phases. General flow char- relies heavily on experimental observations, empirical
acteristics in a three-phase fluidized bed are described correlations, and engineering models. With increasing
in this chapter with specific attention given to high- computer power, the employment of the computa-
pressure phenomena. tional fluid dynamics approach has gained consider-
1.2 Industrial Applications Most of the liquid filtration operations follow the
mechanism of cake filtration. As the filtration pro-
As mentioned earlier, liquid–solids separation technol- ceeds, the particles retained on the filter medium will
ogy is basic to many manufacturing industries (chemi- form a growing cake with porous structure. The small
cal, mineral, food, beverages, etc.) as well as to particles that are able to pass through the pores initi-
pollution abatement and environmental control. It is ally will be trapped at a greater depth as they traverse
difficult to find any important engineering enterprises through this porous cake. This cake becomes the true
in which liquid–solids separation does not play an filter medium and hence plays a very important part in
important part. Major industrial and commercial the entire filtration operation. The mechanism of flow
applications of four key unit operations for liquid– within the cake and the external conditions imposed on
solids separation are summarized here: the cake are the basis for modeling a filtration process.
Under a pressure gradient, if the particle packing filter medium. The specific cake resistance, a, can be
arrangement in the cake can sustain the drag force expressed as
without deformation, the cake is regarded as incom-
pressible. However, stresses developed in the parti- 1
a¼ ð2Þ
culate structure usually lead to deformation and krs ð1 eÞ
compression with substantial changes in the porosity
and permeability as a load (fluid drag) is applied. This where k is the permeability, rs the density of the solid
kind of filter cake is known as a compressible cake. particles, and e the porosity of the filter cake, defined
as the ratio of the volume of voids in the cake to the
total volume of filter cake.
2.1.1 Cake Filtration Equation—Two Resistance Equation (1), known as the two-resistance filtration
Model model, is a simple expression for describing the filtra-
tion of incompressible cakes, the specific cake resis-
The structure of the deposited cake resembles the
tance, a, can be regarded as a constant. In the case
structure of a packed bed. Based on the principles
of compressible cakes, the effect of variation in cake
for flow through a packed bed with the consideration
porosity on specific cake resistance must be considered
of resistances offered by both the filter medium and the
(Tiller and Shirato, 1964). Filtration tests should be
filter cake, an equation of cake filtration can be
performed under different pressure drops to establish
obtained (Akers and Ward, 1977; Cheremisinoff and
an empirical relation between the specific cake resis-
Azbel, 1989; Tiller et al., 1987a):
tance, a, and the pressure drop across the filter cake,
dV AðPÞ Pc (McCabe et al., 1993):
¼ ð1Þ
dt m½ðacV=AÞ þ Rm a ¼ a0 ðPc Þn ð3Þ
where V is the filtrate volume, t the filtration time, m where a0 and n are empirical constants. The constant n
the fluid viscosity, c the mass of solid per unit volume represents the compressibility coefficient of the filter
of filtrate, A the cross-sectional area of filter, P the cake. For an incompressible cake, n equals zero. The
pressure drop across the filter, and Rm the resistance of value of n ranges from 0.25 for moderately compress-
2.3.3 Ultrafiltration
Ultrafiltration (UF) is a membrane filtration process
used for separating or collecting submicron-size par-
ticles (0.001 to 0.02 mm) from a suspension (Cheryan,
1986). It is usually employed to concentrate or frac-
tionate a solution containing macromolecules, colloids,
salts, or sugars. The UF membrane functions as a sieve
with the pore size of molecular dimensions.
A UF process can be operated in either batch or
continuous mode. The determination of membrane
surface area ðAÞ for the UF process requires three
parameters: (1) flux, J, which is the measure of the
membrane productivity; (2) the volume of permeate,
Vp , passed through the membrane, and (3) the volume
of retentate, Vr , retained on the membrane surface.
The average flux, Jav , can be estimated from the
equation
Jav ¼ Jf þ 0:33ðJi Jf Þ ð13Þ
where Jf is the final flux at the highest concentration
and Ji the initial flux. The material balance gives
Vf ¼ Vr þ Vp ð14Þ
where Vf is the volume of feed. The volume concen-
tration ratio (VCR) is defined as
Vf
VCR ¼ ð15Þ
Vr
Eqs. (13) through (15) can be used to estimate the
membrane surface area (Chiang and He, 1995),
which can be expressed as
Figure 2 Membrane filtration process configurations. V f Vp
A¼ ð16Þ
(From Zeman and Zydney, 1996.) Jav
The configuration of UF is usually designed as poly-
meric and asymmetric modules for high productivity
dead-end passages and variable channel dimensions, and resistance to plugging. Membrane modules used in
the Carman–Kozeny equation may be used. There- the UF process design are similar to those adopted in
fore, the flux across the microfiltration membrane ðJÞ MF process design (see previous section).
can be calculated by
dV e3 P 2.4 Filter Media
J¼ ¼ ð12Þ
Adt Kð1 eÞ2 Sp2 mLm
After specifying the filter type and the optimum oper-
where K is the Kozeny coefficient, depending only on ating conditions, the remaining issue for a filter design
the pore configuration, Sp the surface area of a single is the selection of the most suitable filter medium. In a
particle; and Lm the effective membrane thickness. All filtration process, the fundamental role of a filter med-
other parameters are as defined in the previous section. ium is to separate effectively the particulates from
For more detailed information regarding the per- a flowing fluid to provide a sufficiently clean filtrate
formance of microfiltration, readers are referred to without clogging and damaging the medium (i.e., low
energy consumption and long operating period). Thus ity of the media is, therefore, used as an indication of
proper selection of the filter medium is often the most high porosity and in turn a low particle retentivity. The
crucial step for assuring efficient filtration operation. ideal medium would provide the maximum open (free)
The major considerations include the permeability of area for flow while it meets the required retentivity. A
the medium relative to a pure liquid, its retention comparison of the free area for commonly used filter
capacity relative to particulates of known size, and media is presented in Table 2.
the medium pore size distribution. An alternative classification of filter media is based
The resistance (or permeability) of a filter medium on the minimum size of the trapped particles (see Table
directly affects the capital and operating costs. Most 3). Obviously, this classification provides only a gen-
manufacturers also employ the permeability of filter eral guideline to the types of media available. In fact,
media as a measure of particle retention, which is many process parameters (such as the particulate
related to its pore size and porosity. A high permeabil- concentration, etc.) affect the retention behavior of
Ultrafiltration
Flat plate Dorr-Oliver USA Ioplate US1 Polysulphone 5.0–12.2 2.54 0.37 0.067 1:894:25 104 21.84 48.0–108.0
US2 Polyvinylidene fluoride 5.0–12.2 2.54 0.37 0.067 1.80–1:42 104 21.84 30.0–36.0
US3 VC-AN copolymer 5.0–12.2 2.54 0.37 0.067 0.95–1:18 104 21.84 24.0–30.0
Flat plate DDS Lab Unit 35 US4 Polysulphone 0.68–1.36b 5.90b 0.22 0.150 0.69–6:94 105 6.02 4.0–41.0
Denmark
Lab Unit 20 US5 Polysulphone 0.10–0.20b 5.19b 0.076 0.018 0.16–8:10 105 0.78 6.0–42.0
Tube Abcor 24.0 HFA UT1 Cellulose acetate 7.60b 2.54 3.05 0.204 9:43 105 19.30
USA
HFM UT2 Polyvinylidene fluoride 11.36–22.7b 2.54 3.05 0.204 1.41–2:82 104 43.26 36.0–72.0
Tube Romicon- PM UT3 Polysulphone 3.32–6.48 1.10 0.635 0.022 0.116–1:62 104 5.39 1.3–18.0
Amicon
USA
XM UT4 Modacryl polymer 2.84–6.02 1.10 0.635 0.0021 1.39–1:74 104 4.87 15.0–19.0
Tube Berghof BM UT5 Polyamidimide 0.011–0.266 0.60 0.30 5:65 104 0.116–6:94 105 0.083 0.07–4.20
Germany
Microfiltration
Tube Enka/ Dyna-Sep MT1 Polypropylene 4.21–7.97 5.50 1.83 0.032 0.365–1:62 104 33.44 20.0–89.0
Membrana Sampler
Germany
Microdyne MT2 Polypropylene 1.0–3.0 1.80 0.56 1.2 0.556–2:78 104 3.60 10.0–50.0
a
US ¼ ultrafiltration slit, UT ¼ ultrafiltration tube, MT ¼ microfiltration tube.
b
Estimate
c
v ¼ 106 m2 =s for water at 258C
d
Rew ¼ ðhVw =vÞ or ðdVw =vÞ
Source: Belfort G. In: Muralidhara HS, ed. Advances in Solid–Liquid Separation. Columbus: Battelle Press, 1986, pp 182–183, Table 2.
particulates within the filter media. The detailed infor- 2.5.1 Requirements for Filter Aid Selection
mation regarding the filter media properties and their
Filter aid selection should be based on laboratory
selection can be found in the Handbook of Filter
tests. The requirements for preliminary evaluation
Media (Purchas, 1996) and the Filters and Filtration
of the selected filter aids may be summarized as
Handbook (Dickenson, 1994).
follows:
A good filter medium should have the following
features: 1. Form a thin and rigid lattice layer with high
porosity
1. A wide size distribution of particles from the
2. Have low specific surface or coarse size
slurry, producing a clean filtrate
3. Have a narrow fractional size distribution by
2. An economic filtration time, i.e., minimum
removing the finer size fractions
filtrate flow resistance
4. Create a rapid particle bridging and settling or
3. Easy discharge of the filter cake from the
a uniform filter aid layer
medium
5. Be chemically inert and able to prevent
4. Sufficient strength to withstand filtering pres-
medium cracking and clogging
sure and mechanical wear
5. Avoidance wedging of particles into its pores These requirements are all found in the two most com-
and an adequate medium lifetime mon filter aids. The first one is the diatomaceous silica
6. Low cost type filter aid (also called diatomite, kieselguhr or dia-
tomaceous earth), which contains 90% or almost pure
2.5 Filter Aids silica and particle size mostly smaller than 50 mm. Its
bulk density ranges from 128 to 320 kg/m3 . Calcinated
In order to improve the filtration characteristics of diatomaceous additives display their high retention
hard-to-filter suspensions, such as those with slow ability with relatively low hydraulic resistance. the rela-
filtration rate, rapid medium binding, or unsatisfac- tive permeability of the calcinated diatomite increases
tory clarity, addition of another particulate solid up to 3–20 times that of natural diatomite. The disad-
material is one of the best alternative measures. vantage of diatomite is that it may foul filtering liquids
Such solid material is termed a filter aid. Filter aids by dissolved salts and colloidal clays. The second most
added to the suspensions build up a porous, perme- common filter aid is expended perlite, which is a
able, and rigid lattice structure for retaining solid glasslike volcanic rock. The porosity of this filter
particles and allowing the liquid to pass through. aid is in the range of 0.85–0.9. Its bulk density ranges
Smallest particle
Main type Subdivisions retained, microns*
from 48–96 kg/m3 . The key advantage of perlite over 2.5.2 Filtration with Filter Aids
diatomite is its relative purity.
In general, filter aid filtration should be used only for
Cellulosic fiber (ground wool pulp) is applied to
systems that meet two key requirements. First, the
cover metallic cloths. This filter aid forms a much
desired product is the filtrate, not the cake. Second,
more compressible cake with good permeability but
the filter aid is acceptable in the filter cake or the filter
displays a smaller particle retentivity than diatomite
cake can be easily repulped and refiltered to remove the
or perlite. The cost of cellulose is higher than that of
filter aid. In addition, the particle-settling rate should
diatomite or perlite. Thus this and other filter aids are
be less than 0.012 m/min and the particle concentration
only applied to special cases (such as precoat stabiliza-
lower than 0.1 wt% (<0.3% for rotary drum precoat).
tion or chemical resistance).
simple belt vacuum filter is illustrated in Fig. 6. Filtrate 2.6.2 Deep-Bed Filtration Units
permeates through the filter medium and is directed to
centrally situated drainage pipes. In this filter, the sus- Commercial deep-bed filters consist of a cylindrical or
pension feed, the wash liquor feed, the cake discharge, rectangular packed bed through which the suspension
and cloth cleaning are continuous. The cake thickness to be filtered is passed. The common types of deep-bed
generally ranges from 1 to 25 mm. Cakes up to 100– filter include the slow sand filter, the rapid filter, and
150 mm thick are possible with some fast-draining direct filtration with a flocculated mixture. Typical
materials. The chamber and collector are divided into deep-bed filters are 0.5–0.3 m in height and 1 m in dia-
sections from which filtrate and washing liquid can be meter. Smaller packing material (filter media) provide
discharged. Efficient cake discharge can be accom- a greater surface area and result in a more effective
plished by separating the belt from the filter cloth capture of suspended solids, but the bed pressure
and directing the latter over a set of discharge rollers. drop and clogging tendency also increase. Often, the
The simplicity in design is an advantage of belt filters. design of the deep-bed filter is to employ mixed size
In addition, the countercurrent cake washing and media packed in multilayers, as shown in Fig. 7
removal of thin cake can be easily achieved in such (Cheremisinoff, 1998).
filters. The shortcomings of this type of filter are In water filtration, the dual media filters are usually
large area requirements, inefficient use of the available designed using coarse anthracite coal on the top of fine
filter area, and ineffective washing at the belt edges. silica sand. The coarse anthracite layer serves to prevent
Other types of filters, such as the rotating table the formation of surface deposits on the sand bed,
filter, the candle filter, filter presses, etc., are described resulting in the formation of a compressible cake
in detail in the recent filtration literature (Dickenson, along the bed depth. Thus the depth removal of parti-
1994; Schweitzer, 1997; Svarovsky, 1990). culates throughout the bed would be the key feature.
Kdp2 ðrs rÞ dp ut r
Particulate settling ut ¼ Rep ¼ < 0:1
m m
particle size range of 1 to 200 mm
4ðrs rÞgdp 0:5
ut ¼
3rCD Rep 0:2
ut ¼ terminal settling velocity; dp ¼ the particle size;
rs ¼ the solid particle density; r the liquid density;
m ¼ the liquid viscosity; K (Kozeny constant) ¼ 0:002 for ut in
m/h, dp in micrometer, and m in cp;
g ¼ gravitational constant; Cd ¼ drag coefficient
Hindered setting u ¼ ut en non-flocculated suspension
(Richardson and Zaki, 1954a,b)
n is a function of the particle Reynolds number (see Table 5b)
d =D
u ¼ ut ½1 kf fS ð4:70þ17:8 Þ
flocculated suspension
(Scott, 1984)
f ¼ kf f s
f ¼ settling units
fs ¼ volumetric dry concentration
kf ¼ the factor allowing for the liquid closely associated with the solids
d ¼ the mean volume-surface-length diameter relevant to sedimentation
not greater than 20% of the design level to avoid shut- ture (Osborne, 1990; Perry et al., 1997; Schweitzer,
down due to inevitable process upsets or overloads. 1997).
Rake speed requirements depend on the type of solids
entering the thickener. Rake speed ranges used are 3
3.2.2 Thickener Types and Selection
to 8 m/min for slow-settling solids, 8 to 12 m/min for
fast settling solids, and 12 to 30 m/min for coarse A thickener consists of several basic components: a
solids or crystalline materials. tank to contain the slurry, a feed well for feed supply
The detailed description of the major mechanical (to minimize the turbulence effect), a rotating rake
components used for the thickeners, such as tank, mechanism, an underflow solids withdrawal system,
feed well, drive support structure, drive and lifting and an overflow launder. Recirculation of the under-
devices, rake structure, underflow withdrawal, and flow back to the thickener feed line is a common prac-
overflow collection systems can be found in the litera- tice, but care must be exercised in the design to avoid
Duty classification
the overload and underflow line plugging. There are normally increase the solid settling rate 2–10 times
two basic types of continuous thickeners: conventional over that obtained from conventional thickeners.
thickeners and high-rate thickeners. Obviously, the required unit area will reduce by the
The conventional thickeners can be operated with same factor. As a result, the capital cost will be reduced
or without flocculants. They may be divided into three significantly. However, care must be exercised in adding
classes based on their specific drive-support configura- flocculants, since overflocculation will produce a gellike
tions: bridge supported, center column supported, and structure that is hard to remove from the thickener.
traction driven. The design configuration of bridge- Numerous combination and component design var-
supported thickener is illustrated in Fig. 11. The dia- iations are possible. Selection of the type of thickener
meter of this type of thickener can be selected up to should primarily be based on installation and operat-
45 m. The key advantages of this thickener include (1) ing costs. The selection of the final design is made after
ability to produce a denser and more consistent under- an initial conceptual approach is defined. To specify a
flow concentration; (2) use of a simplified lifting; and thickening system, one should consider the following:
(3) fewer parts subject to mud accumulation. The cen- (1) the characteristics of feed suspension: flow rate,
ter-column-supported thickeners are usually designed solid loading, particle-size distribution, and pH value;
for large diameter units (>20 m). The mechanism is (2) the model experimental results for underflow con-
supported by a stationary center column. The traction centration and overflow clarity; (3) a tank size esti-
thickeners are an economical configuration of the mate; (4) the required power or torque; (5) the
center column support. They are mostly adaptable to capability of the selected rake; (6) the materials of con-
the larger tanks over 60 m in diameter. For example, struction; (7) a control scheme; and (8) a cost estimate.
the Superthickener or Caisson thickener is a very big Table 7 lists typical design sizing criteria and operating
center-column-supported unit over 120 m in diameter. conditions for a number of commercial thickeners and
In this thickener, the underflow is withdrawn into the clarifiers. The information presented in the table can be
column and pumped back to the circumference by the used as a guide for selecting the gravity sedimentation
underflow pump. Its maintenance generally is easier units and preliminary cost estimate.
than other types of thickeners, but its installed cost
may be higher.
High-rate thickeners or high-capacity thickeners are 3.3 Clarifiers
designed to take advantage of maximizing the floccula-
tion efficiency. They have specially designed feed wells Continuous clarifiers arre generally employed to treat
and flocculation systems as illustrated in Fig. 12. The dilute suspensions (industrial process streams and
feed well design is aimed at providing a good dispersion domestic wastewater) with a large percentage of
of the flocculants in the feed and transporting the floc- relatively fine (usually smaller than 10 mm) solids.
culated feed suspension into the settling zone of the The clarified liquid is the main product. Clarifiers typi-
thickener without destruction of the newly formed cally are designed for lighter duty operation than
flocs. The increase in flocculation efficiency will thickeners.
3.3.1 Clarifier Design ket levels at a minimum of 1.6 m below the clean water
surface. The design criteria and operating conditions
As mentioned earlier, clarifiers and thickeners are very for commercial clarifiers are given in Table 7.
similar in their overall designs and layouts. The design Clarifier design has traditionally been based on the
considerations of the thickeners can also be used as a principle of dilute sedimentation. The design calcula-
basis for clarifier design. The design configuration of tions of a clarifier include feed well design, sedimenta-
clarifiers may have the following features: (1) rapid and tion basin design (providing enough residence time for
good mixing of chemical additives with feed slurry and the separation), and solid flux (G). The other design
adjustable recirculator speed with at least a range of parameters, such as the torque requirement and clari-
3 : 1; (2) control of the mixer speed (not exceeding 2 m/ fier area, are similar to those of the thickeners, which
s) and the scraper tip speed (less than 3 m/min with have been discussed in the previous section (3.2.1).
speed variation of 3 : 1); (3) a discharge system for The design configurations of the feed well should
easy automation and variation of amount discharged; provide: (1) a decreasing velocity gradient inside and
(4) devices for measuring and varying the slurry con- outside the feed well; (2) gentle mixing to promote
centration in the contacting zone; (5) suspension blan- ideal conditions for dispersing flocculant or coagulant
Table 7 Design Criteria and Operating Conditions for Commercial Thickeners and
Clarifiers
Solid %
Unit area, Overflow rate,
Slurry Feed Underflow m2 day/mg m3 /(m2 h)
kept in a fluidized state. Typical operating rates are The major advantage of the tilted-plate clarifier is its
around 5 to 10 m3 /(m2 h). The concentration of sus- increased capacity per unit of plate area compared to
pended solids in the overflow ranges from 10 to the conventional clarifier. There are two key shortcom-
40 mg/L. The operating and maintenance costs are ings existing in this type of clarifier. First, it produces a
low compared to other types of solid-contact clarifiers. varying underflow solids concentration lower than
These clarifiers are well suited to the case where the other type of gravity clarifiers. Secondly, it is difficult
flow rate varies only slightly. to clean the scale formed within the channel. The appli-
In recent years, tilted-plate clarifiers (see Fig. 14) cations of tilted-plate clarifiers include clarification of
have been widely used. This type of clarifier is also plating and pickling wastes, paper mill effluent, and
known as lamella or tube settler. It contains a multiple tertiary wastewater treatment. For the clarifier selec-
plate assembly inclined at 308 to 608 from the horizon- tion one can follow the same guidelines as for thickener
tal. Clarified liquid and settled solids may flow counter- selection, as discussed in the last section.
currently or cocurrently along each channel. Generally,
a plate angle of 308 to 408 is a suitable inclination when
operating cocurrently, whereas an increase of up to 558 NOMENCLATURE
to 608 may be necessary with the countercurrent
arrangement. The suspended solids settle only a short A ¼ clarifier surface area
distance in the channel before falling into the base. Ao ¼ unit area
Settled solids are collected in a lower compartment C ¼ solid concentration in the feed
and discharged by pumping. The selection of the plate Co ¼ test or feed solid concentration
spacing should be based on two considerations. First, Cu ¼ underflow solids concentration
the plate spacing must be large enough to accommo- D ¼ thickener diameter
date the opposite flows. Secondly, the channel space G ¼ solid flux into the clarifier
Ho ¼ initial height of suspension in the test
arrangement should be able to limit the interference
kT ¼ constant dependent upon the application of
and plugging as well as to provide enough residence
thickener
time for the solids to settle in a short distance. The Q ¼ effluent flow rate
tilted-plate clarifiers usually use 10 to 50 plates with Qu ¼ underflow rate
spacing from 0.05 to 0.1 m. The channel lengths range T ¼ torque
from 1 to 3 m with widths of about 1.2 m. Operating tC ¼ compression time for a required particular
capacities vary from 0.5 to 3 m3 =ðm2 hÞ. underflow concentration
Figure 20 Peeler centrifuge: (a) feed and separation phase, (b) solids discharge phase. (From Jacobs and Penney, 1987.)
the stroke zone of the pusher plate. A layer of solid greatly increased as the solid concentration is
cake 2.5 to 7.5 cm thick is formed. This solid layer is increased. They are capable of handling fragile crys-
moved over the filtering surface by the reciprocating tals, which may be damaged in other types of centri-
pusher plate. Each stroke of the pusher plate moves the fuges. The small pusher centrifuges can handle a few
solid layer a few inches toward the lip of the rotor; on grams per second, while the large ones have capacities
the return stroke a space is opened on the filtering up to 55,000 kg/h. Based on the number of rotors,
surface in which more solid cake can be deposited. pusher centrifuges are divided into two types: single
When the cake reaches the lip of the rotor, it is thrown stage pusher centrifuges with only one rotor and multi-
into the large casing by the centrifugal acceleration. stage pusher centrifuges with two or more rotors. The
The liquor passes the filtering screen and leaves the multistage pusher centrifuge is particularly suitable for
machine by ducts well separated from the solids dis- particles forming a soft cake or having a high frictional
charge. resistance to sliding on the filter medium. A schematic
The pusher centrifuges are normally used when the of a two-stage pusher centrifuge is also shown in
feed can be concentrated above 60% by volume, since Fig. 21. The purpose of the multistage design is to
the operation capacity of the pusher centrifuge is keep the length of the rotor short.
Disc Conveyor
Tubular (nozzle type) bowl
Source: Cheremisinoff PN. Solid/Liquids Separation. Lancaster PA: Technomic, 1995, p 240, Table 6.3.
NOMENCLATURE
5 HYDROCYCLONES
A L ¼ arithmetic mean cake area
A a ¼ logarithmic mean cake area
A2 ¼ area of filter medium The hydrocyclone is based on the principle of centri-
b ¼ height of the basket in a filter centrifuge fugal force causing the separation of solids from a
Dpc ¼ cut diameter liquid by the differences in density and particle size.
G ¼ gravity acceleration A typical hydrocyclone consists of a cylindrical section
mc ¼ total mass of the solids in the filter and a conical section, as shown in Fig. 23. It does not
Nc ¼ centrifugal number have any internal rotating parts. An external pump is
qc ¼ feed volumetric flow rate used to transport the liquid suspension to the hydro-
r1 ¼ distance from the surface of liquid to the axis cyclone through a tangential inlet at high velocity,
of rotation
which in turn generates the fluid rotation and the
r2 ¼ distance from the wall of the bowl to the axis
necessary centrifugal force. The outlet for the bulk of
of rotation
rA ; r B ¼ distance from the particle to the axis of the liquid is connected to a vortex finder located on the
rotation axis of the cylindrical section of the vessel. The under-
ri ¼ inner radius of the cake flow, which carries most of the solids, leaves through
Rc ¼ centrifugal force an opening (apex) at the bottom of the conical section.
Rm ¼ filter medium resistance The principle and basic design of the hydrocyclone
has been known for more than a century (Bretney,
Greek Symbols 1891), but it did not find significant application in
industry until the late 1940s. These separation devices
a ¼ specific cake resistance were first used in mining and mineral processing, but in
r ¼ density of liquid (filtrate) recent years their applications have spread to many
r ¼ density difference between solids and liquid other industries, including chemical manufacturing,
m ¼ liquid viscosity power generation, and environmental cleanup.
o ¼ angular velocity
5.1 Separation Efficiencies
UCu ET Rf
ET ¼ ð26Þ ET0 ¼ ð27Þ
QCf 1 Rf
where U is the underflow volumetric flow rate, Q the
where Rf is the underflow-to-throughput ratio defined
feed volumetric flow rate, Cu the solid concentration in
as
the underflow, volume, or mass fraction, and Cf
the solid concentration in the feed, volume, or mass U
fraction. Rf ¼ ð28Þ
Q
Because an overall mass balance must apply, the
total efficiency ET can be determined by measuring
any two of the three streams (feed, underflow, and
overflow) for total solid amount, assuming no accumu- 5.1.3 Grade Efficiency
lation of solids in the hydrocyclone.
In place of the total efficiency, a grade efficiency corre-
sponding to a particular particle size is used, since
5.1.2 Reduced Total Efficiency
a hydrocyclone is a size-dependent separator. A
There are two problems associated with total effi- graphical representation of the relationship between
ciency. First, if a hydrocyclone delivers both liquid the grade efficiency and the particle size is called the
and solid to the underflow and nothing to the overflow, grade efficiency curve, as shown in Fig. 24 (Svarovsky,
an ideal total efficiency of 1 will result. Second, without 1985).
any separation, a hydrocyclone, by simply splitting the The grade efficiency curve can be determined by
feed to an overflow and an underflow, will result in measuring the total efficiency and the particle size
certain ‘‘guaranteed’’ total efficiency. In order to over- distribution of any two of the three streams (feed,
come these weaknesses, several alternative definitions underflow, and overflow). One of the following
of efficiency are used. One of the definitions is the equations can be used for calculating the grade
reduced total efficiency ET0 : efficiency:
5.2.3 Velocity and Pressure Distributions Equilibrium Orbit Theory. The general concept
that particles of a given size reach an equilibrium
The velocity of the fluid flow in a hydrocyclone can be
radial orbit position in the hydrocyclone forms the
resolved into three components: tangential, axial, and
basis of equilibrium orbit theory. The fine particles
radial. The most useful and significant of these three
reach equilibrium at small radii where the flow is
components is the tangential velocity.
moving upwards and transports fines to the overflow,
The tangential velocity of the fluid in the hydro-
while the coarse particles find equilibrium position at
cyclones increases as the radius decreases, which is
large radii where the flow is moving downwards and
expressed by the empirical relationship
carries these particles to the underflow outlet (apex).
VRn ¼ const ð33Þ The dividing surface is the locus of zero vertical velo-
city (LZVV). The size of the particles that find equili-
where V is the tangential velocity, R the radius, and n brium radius on LZVV will be the cut size that has
the empirical exponent, usually between 0.5 and 0.9. an equal chance to finish in either overflow or under-
For the outer regions of a free vortex by definition n flow.
will be 1. From Eq. (33) one can conclude that as the In developing the equilibrium orbit theory, a key
radius R approaches zero the tangential velocity V will assumption made by Bradley and Pulling (1959) is
approach infinity. In practice, this cannot happen since the existence of a ‘‘mantel’’ in the hydrocyclone,
Eq. (33) only holds true until small values of the radius which precludes inward radial velocity in the region
are reached, when the tangential velocity starts to fall immediately below the vortex finder. Furthermore,
with a further decrease in radius. The static pressure the LZVV is assumed to be in the form of an imaginary
increases radially outward because of the centrifugal cone whose apex coincides with the apex of the hydro-
field induced by vortex flow in the hydrocyclone. cyclone and whose base is at the bottom of the mantle.
This static pressure head is primarily determined by Based on these assumptions, the equilibrium orbit
the distribution of the tangential velocities within the theory has led to the development of empirical correla-
flow. Therefore the tangential velocity distribution can tions for determining the cut size and pressure drop in
be estimated from the simple measurements of the hydrocyclone operation.
radial static pressure distribution. The major deficiency of the equilibrium orbit theory
Since the outer and the inner layers move in oppo- lies in its lack of consideration of the effect of turbu-
site vertical directions (i.e., the flow in the outer vortex lence flow on particle separation and the residence time
moving down and the inner vortex moving up), there is of the particles in the hydrocyclone (as not all particles
a well-defined locus of zero vertical velocity between are able to find equilibrium orbits within their resi-
the two vortices. This locus forms an invisible bound- dence time). In spite of such weaknesses, it proves to
ary, which plays an important role in particle separa- be a reasonable approach for determining the hydro-
Running cost
Geometrical proportions Scale-up constants criteriona
Type and size of
hydrocyclone Di =D D0 =D l=D L=D Cone angle, y, degrees Stk50 Eu Kp np (Stk50 Þ4=3 Eu