1.

INTRODUCTION
Fluidized beds are used widely in chemical processing industries for separations, rapid mass and
heat transfer operations, and catalytic reactions. A typical uidized bed is a cylindrical column
that contains particles and through which uid, either gaseous or liquid, ows. In the case of
uidized bed reactors, the particles would contain a catalyst, and for separations, the particles
might be an absorbent or adsorbent. The velocity of the uid is suciently high to suspend, or
uidize, the particles within the column, providing a large surface area for the uid to contact,
which is the chief advantage of uidized beds. As shown in Figure 1, uidized beds range in size
from small laboratory-scale devices to very large industrial systems. Regardless of whether the
uidized bed is used for a separation or reaction, a key goal is to operate the bed at a ow rate
that optimizes the application. Accurate models would aid signicantly, but modeling, even at a
qualitative level, of the complex dynamics in uidized beds continues to challenge engineers
and scientists. The challenge arises from the necessity of considering both the solid and uid
phases and the interplay between them to form a complete picture for understanding the
properties of uidization. Figure 2 assists in understanding the inherent challenge: multiple ow
patterns within uidized beds can be observed depending upon the velocity of the uid. For
suciently low rates of ow, uid passes through the void space between particles without
disturbing them. This case where the bed of particles remains in place is referred to as a xed
bed. At higher rates of ow, the drag forces acting on the particles can exceed the
gravitational forces and lift particles. However when the bed of particles expands, the drag
force drops as uid velocity in the void spaces declines. The result is a highly dynamic state to
which we refer as uidization. Regimes of uidization which can be easily identied from
qualitative observations include bubbling and slugging patterns at relatively low ow rates and
turbulent ow patterns at higher ow rates. At very high rates of uid ow, the drag force can
exceed the net gravitational forces on individual particles, even when the particles are widely
separated. In this regime of pneumatic conveying, particles are carried through the container
and must be reintroduced externally. You will investigate some key parameters that govern
uidization, including the dependence of the minimum uidization velocity on particle shape,
size and density. This is the point at which the xed bed of particles transitions to the
particulate regime.


Fig.1.The regimes of uidization as a function of the uid velocity. At very low ow rates (left),
the particles behave as a porous media, or xed bed. After the gas velocity surpasses a critical
value, the particles become uidized. (Schematic is based upon a similar diagram appearing in
Perry et al.

2. Minimum Velocity of Fluidization
The minimum velocity at which a bed of particles uidizes is a crucial parameter needed for the
design of any uidization operation. The details of the minimum velocity depend upon a
number of factors, including the shape, size, density, and polydispersity of the particles. The
density, for example, directly alters the net gravitational force acting on the particle, and hence
the minimum drag force, or velocity, needed to lift a particle. The shape alters not only the
relationship between the drag force and velocity, but also the packing properties of the xed
bed and the associated void spaces and velocity of uid through them. To nd the minimum
uidizing velocity, Umf, experimental and theoretical approaches can be used. Methods for
calculating the ow rate at which uidization occurs are described rst, as a review of
fundamental ideas that govern the behavior of the bed of particles. Then, a procedure for
estimating the minimum velocity from experimental measurements is described.
2.1 Calculating the Minimum Velocity
The incipient point at which the uid, or gas, ow causes the bed of particles to expand and lift
into the vertical column is marked by a conceptually simple balance. At U
mf
, the hydrodynamic
drag force on the particles F
d
, due to the ow of gas through the packed bed of particles,
matches (or just exceeds), the net gravitational forces F
g
,
0 = F
g
+ F
d
, (1)
Where the balance had been made in the direction of gravity.1 The calculation has many
similarities to evaluating the terminal velocity of a single particle in a ow. Here however, the
balance must be performed on the entire bed of particles as shown in Figure 3.
2.1.1 Gravitational Forces
The net gravitational forces on the bed of particles must consider the weight W of the particles
and the buoyancy forces Fb,

F
g
= W F
b
= (
p

f
) gV
p
, (2)
Where
p
is the density of the particles,
f
is the density of the uid, g is the gravitational
acceleration constant, and V
p
is the total volume of particles within the uidized bed. For a low
density uid, such as gas, the buoyancy force represents a small correction to the net
gravitational force. Since the drag forces (see Section 2.1.2) are generally written in terms of
the bed voidage, expressing the gravity force in the same way proves convenient. If the weight
and density of the particles is known, then the particle volume can be calculated. Using the
denition of bed voidage
m
, the volume of the particles can be written as
V
p
= AH (1
m
), (3)
Where A is the cross sectional area of the uidized bed and H is the height of the bed of the
particles prior to the onset of uidization.


Figure 3: The bed of particles and the force balance. When the weight of the particles (W)
exceeds the buoyancy forces (Fb) and the drag forces (Fd) due to the uid velocity U, the
particles remain xed in place. The velocity U is the minimum uidization velocity if a small
increase of velocity, U, causes the bed to expand by a small amount H over its original height,
H.



2.1.2 Hydrodynamic Drag Forces
The local pressure drop through a porous medium is a function of the bed voidage, the ow
velocity, and details of the particles,
P = f (,U,De,s), (4)
and other properties of the uid. The velocity U is the supercial velocity, or volumetric ow
rate of the uid normalized by the cross-sectional area of the column. The equivalent volume
diameter De and sphericity factor s account for the details of the particle size and shape; for a
spherical particle, the sphericity equals one and the equivalent diameter is simply the diameter
of the sphere.2 For a homogeneous bed of monodisperse particles where is equal to m
everywhere, the pressure gradient P can be integrated over the bed of particles to give the
pressure drop P over the height H of the bed, or P = R H 0 (P) dz = PH. The drag force on
the bed of particles can then be calculated by multiplying by the cross-sectional area of the
column,
Fd = PAH. (5)
The pressure gradient, or drag force, depends on the ow velocity in a non-simplistic manner.
However, dierent regimes of ow can be easily identied, much like the well- known case of
the drag force on a single particle. The regimes are dened in terms of the Reynolds number,
Rep = DU
f
/ (1
m
) , (6)
Where is the viscosity of the uid and the voidage enters the traditional denition to correct
for the use of the supercial velocity U. Figure 4 shows measurements of the packed bed
friction factor,
fp = D
3
P/(
f
U
2
(1 )), (7)
as a function of the Reynolds number. The relationship exhibited in Figure 4 holds only for the
pressure drop prior to the incipient point of uidization, when the particles are packed and the
pattern of ow is that of a uniform, porous medium. For Rep 10, the ow is inertia less and
the relationship between fp and Rep is linear. At very high values of the Reynolds number (ReP
1000), the ow is considered to be within the inviscid Newton region and fp is independent of
Rep. In the viscous, or inertia less regime, the relationship between the pressure drop and
owis linear. This is embodied within the commonly used Carman-Kozney equation [6; 7],
P = 180U (1 )
2
/D
2

3
. (8)
This relationship is equivalent to writing fp = 180/Rep, which corresponds to the variables used
in Figure 4. This equation is derived from an idealized model of a packed bed of particles. In the
model, the tortuous path followed by the uid passing through the particle bed is replaced by a
a set of parallel cylinders having the same ow resistance. For the inviscid case, the Burke-
Plummer equation applies [5],
fp = 1.75; (9)
This expression can be determined by a simple inspection of the data presented in Figure 4.
Within this regime, the pressure drop is proportional to the square of the ow velocity. To
bridge the gap between the Carman-Kozney and Burke-Plummer equations, Ergun [2] proposed
the correlation,
fp = 150 /Rep + 1.75, (10)
Which is a linear combination of the viscous and inviscid relations, albeit with a modied
constant of 150 instead of 180 as given by Kozney [3] and Carman [4]. This is perhaps the most
widely used equation for describing ow through porous media [6]. The reason is clear upon
examining the solid line (Erguns correlation) in Figure 4, which ts the experimental data with
delity over the entire range of Reynolds numbers.


Figure 4: Friction factor as a function of Reynolds number for a packed bed of particles. Symbols
are data from experiments, the dashed line is the Carman-Kozney equation, and Erguns
correlation is represented by the solid line. (Figure adapted from Ergun & Orning [2].)

2.1.3 Solving for U
mf

Balancing the forces as indicated in Equation 1, using Equations 2-3 for Fg and equations given
in Section 2.1.2 to determine Fd, results in expressions that can be solved to determine Umf,
the velocity U at which uidization occurs. All other quantities must be either measurable or
known from another source; if the latter, be sure to explicitly indicate the source and the value
used. Note that the error in the measurements and values from other sources propagate into
an error for the calculation of the minimum velocity of uidization. This error must be
considered when assessing the comparison to the experimentally determined value of Umf as
described in Section 2.2. Rather than balancing the forces as in Equation 1, some methods of
predicting Umf rely upon a direct correlation. One such result was developed by Wen & Yu [7],
Re
mf
= (33.7
2
+ 0.0408Ar)
1/2
33.7, (11)
Where Ar and Remf are the Archimedes number and a modied Reynolds number. The
Archimedes number is a ratio of gravity and viscous forces,
Ar =D
e
3

f
(
p

f
) g/
2
, (12)
Where on must use De instead of D = sDe. The modied Reynolds number is given by
Re
mf
=DU
mf

f
/ . (13)
Equation 11 can be solved directly to give a predicted value of Umf after the denitions of Ar
and Re
mf
are substituted into the expression.


Figure 5: Diagram of the pressure drop P as a function of the supercial velocity. The negative
sign is indicative of the fact that the drag force acts in opposition to the velocity U. The point E
marks the velocity U
mf
at which uidization occurs.


2.2 Experimental Evaluation of U
mf

Measurements of the pressure drop across the bed of particles can be used to identify the
minimum velocity of uidization. As diagrammed in Figure 5, the pressure drop increases with
ow rate until the bed expands and increases the porosity (point A). Note that the velocity and
pressure drop relationship is not necessarily linear as shown, depending upon the range of ReP
covered. Upon further increasing the velocity, the pressure drop attains a maximum value. Be-
tween points A and B, the frictional drag force causes the particles to rearrange, which can alter
the voidage. Upon rearrangement, the pressure decreases and point B lies above point C as a
result. As U is increased beyond point C, the pressure drop remains approximately constant
until some point D where the velocity is not signicantly greater than at point C. If the process
is reversed by steadily lowering the velocity U, point E will be found instead of point B due to
the dierent voidage resulting from the rearrangement of the particles, and line EF is the
process for reforming the xed bed of particles. This conceptual diagram provides the basis for
the experimental determination of Umf. To identify point E, the uid velocity is increased until
the pressure goes through a maximum and then ceases to change; this method denes the line
CD. The rate of uid ow is then reduced to get the line EF. The minimum uidization velocity is
the velocity at which these two lines intercept. The increments in velocity must be small to
resolve point E. Also, the curve represented in Figure 5 is a pseudo-steady one: after each
increment in velocity U, sucient time must be given for the pressure drop to equilibrate.
Finally, the experimentally determined value of Umf is subject to error, which should be
considered before comparing to any calculated values from Section 2.1.
2.3 Laboratory Objectives
Using the laboratory scale column that is described in Section 4, you will measure the minimum
velocity of uidization, using the ideas presented in Section 2.2. The measurements should be
compared to predicted values to determine the ecacy of the various approximations made in
Section 2.1. Aside from testing the case of spherical particles, you can
investigate the minimum velocity of uidization for cylindrical particles and particles that are
akes, and
perform a study to examine the variation in the minimum uidization velocity from cycle to
cycle.
Limits on your time in the laboratory will prevent you from pursuing all of the possible
objectives; keep in mind that thoroughness is preferred to completing all the possible tasks.
Also, your instructors may ask you to perform some variations on the stated objectives.
2.4 Experimental Methods
The experimental apparatus shown in Figure 6 will be used in your investigation of the
minimum velocity of uidization. You will utilize only gas ow and can experiment with three
types of particles: spherical particles, cylindrical particles, and akes. In each case, the particles
are monodisperse. The equipment is suciently instrumented to provide you with data that
can be used to determine the minimum velocity of uidization using the concepts presented in
Section 2.2. Appendix B provides operating procedures for starting-up and shutting-down the
equipment, as well as information about the particle sizes and equipment. You should come to
the laboratory prepared to run the equipment and perform your experiments, as time will be
limited. To do so, you should be familiar with the operating procedures in Appendix B and you
should have a plan of the steps you will need to take in order to meet the goals of the
experiment. This includes a list of the parameters that will need to be set or measured during
your time in the laboratory. When you report on your experiment, note any deviations you
make from the listed procedures and comment on any suggestions for improvements that you
may have. Please, do not simply list the procedures in your report; rather reference this
manual.

Figure 6: The experimental apparatus in the unit operations laboratory. The diagram shows the
major elements of the experimental equipment and Table 1 contains the key for the various
parts.
1 Air Compressor
2 Pressure Gauge 1 for Air Compressor
3 Valve 1 controlling air pressure entering Fluidized Bed
4 Pressure Gauge 2 for air entering Fluidized Bed
5 Water Bath
6 Valve 2 controlling volumetric air ow rate
7 Air Flow Meter
8 Thermometer
9 Pressure Gauge 3 for uidized bed
10 Measuring Tape
11 Fluidized Bed

DEMONSTRATION CAPABILITIES
The apparatus is designed to facilitate the study of flow through fixed and fluidized beds of
solid particles. Provision is made for the simultaneous study of air and water systems in order
to demonstrate the difference between aggregative and particulate fluidized bed
characteristics. Students can readily observe the important phenomenon of bubbling.
Pressure drop through packed and fluidized beds, for air and water systems
Verification of Carman-Kozeny equation
Onset of fluidization
Observation of differences between particulate and aggregative fluidization
DESCRIPTION
Upward flow of fluid through a bed of particles is a naturally occurring phenomenon, for
example the movement of ground water. Industrial applications include ion-exchange,
extraction of soluble components from raw materials and other chemical processes. The
Armfield CEL apparatus facilitates the study of flow through fixed and fluidized particle beds.
The simultaneous study of air and water systems and the phenomenon of bubbling can be
observed. The apparatus comprises a laminate backboard, mounted on a tubular steel frame,
which supports the two clear acrylic cylindrical columns for air and water. Each column is
supplied with sintered bronze bed plates and wall tappings. The columns are demountable in
order to remove the particle bed. Separate manometers are provided for measurement of bed
pressure drop in the columns. Water is circulated from a sump tank through a control valve and
variable area flow meter to the appropriate column by a pump. An overflow returns the water
to the sump tank. Air supply to the second column is by diaphragm pump via a bypass control
valve and variable area flow meter. Air is discharged to atmosphere. The apparatus is designed
to be used on the laboratory bench.
REQUIREMENTS
Electrical supply:
CEL-A: 220-240V/1ph/50Hz@2A
CEL-G: 220V/1ph/60Hz@5A
OVERALL DIMENSIONS
Height: 1.00m
Length: 0.75m
Depth: 0.60m
SPECIFICATIONS
A bench top apparatus for the study of fixed and fluidised beds of solid particles
Two cylindrical columns, one for air and one for water, each column diameter 50mm and
height 550mm, with sintered bronze bed plates
T wo sizes of bed material (Ballotini) are supllied, 200/300 and 500/750 micron ranges
Each column has tapping points and a manometer for measurement of the bed pressure drop
(manometer fluid supplied)
Pumped water and air supplies with control valves and variable area flowmeters Flow
ranges: Air: 2-25 litres/min Water: 0.2-2.0 litres/min.

References
[1] R.H. Perry, D.W. Green, and J.O. Maloney. Perrys Chemical Engineers Handbook. McGraw-
Hill, New York, 7th edition, 1997.
[2] S. Ergun and A.A. Orning. Fluid ow through packed columns. Chemical Engineering
Progress, 48:8994, 1952.
[3] J. Kozney. Ueber kapillare leitung des wassers im boden. Sitzungsber Akad. Wiss., 136:271
306, 1927.
[4] P.C. Carman. Fluid ow through granular beds. Transactions, Institution of Chemical
Engineers, 15:150166, 1937.
[5] W.E. McCabe, J.C. Smith, and P. Harriott. Unit Operations of Chemical Engineering. McGraw
Hill, 2001.
[6] W.C. Yang. Flow through xed beds. In Wen-Ching Yang, editor, Handbook of Fluidization
and Fluid-Particle Systems. CRC Press, 2003.
[7] C.Y. Wen and Y.H. Yu. Mechanics of uidization. Chemical Engineering Progress Symposium
Series, 62:100111, 1966.















Length of column= 63cm
Thickness of column sheet=0.3cm
Diameter of the column=
Length of flowmeter=43cm
Length of manometers=65cm
Dimensions of switch board= 8 cm x 8 cm