1

Simulation of Hydrodynamics in a Downflow Fluidized Bed Reactor

Archwit Aimdilokwong, Terdthai Vatanatham, Sunun Limtrakul


Department of Chemical Engineering, Faculty of Engineering, Kasetsart University


ABSTRACT

A downflow circulating fluidized bed (downer) is important in gas-solid reaction industry. Apart from the
conventional upflow circulating fluidized bed (riser), a downer has a special flow characteristic. The flows of gas and
solids in a downer resemble the ideal plug flow. This makes the residence time distribution in a downer narrower
compared with a riser. Hence a downer gives higher selectivity for many multiple reactions and becomes an
important reactor in industry. But there is still inadequate data about hydrodynamics in a downer. In this study, a 9.3
m. high and 0.1 m. I.D. downer was simulated with Fluent software to study hydrodynamics. The effect of varying
solid circulation rate (G
s
) is studied. The radial profiles of solids fraction and solids velocity are used to characterize
hydrodynamics in a downer. Simulation results can be used to design a proper downer for industrial uses.

Keywords: Circulating Fluidized Bed, Downer, Simulation, Fluent, Hydrodynamics

Corresponding Author
fengsul@ku.ac.th
1. Introduction
In a fluidized bed, solid particles are fluidized by
fluid (either gas or liquid) and behave themselves like
a fluid when the fluidizing fluid velocity reaches an
exact amount called minimum fluidized velocity or U
mf
.
A fluidized bed is widely used as a reactor in which a
fluid reactant is injected into the bed for fluidizing
solid particles. The advantages of a fluidized bed is
enhancing gas-solid contacting and heat and mass
transfer rate. In addition, a fluidized bed containing
small particles has lower pressure drop than a fixed
bed reactor. The solid particles can be catalysts or
solid reactants depending on the application.
Fluidized bed reactors are important for chemical
process, mineral process, pharmaceutical process,
fluid catalytic cracking process, and solid catalyzed
reaction process.
Circulating fluidized bed (CFB) reactors are
operated with higher superficial gas velocity
compared to conventional fluidized bed reactors. By
using high superficial gas velocity, particles can be
circulated through circulating line. After particles
leave the reactor, they are separated by one or more
cyclones and can be re-generated or cooled down
before they re-enter the CFB reactor section again.
CFBs can be separated to two types by
considering the flow of gas phase and particle phase,
i.e. co-current flow and counter-current flow. The co-
current flow reactor has the same flow directions of
particles and fluid while the counter-current flow has
the opposite flow directions of particles and fluid.
CFBs can also be specified by the flow direction
of the particle phase. CFB reactors with the up flow
of solids are called risers while CFB reactors with
down flow of solids are called downers.
The hydrodynamics in CFB is very important for
the reactor performance because it affects both mass
SM02-1
15
2
( ) ( ) 0 = +

i i i i i
u
t

= 1
i
( ) ( )
( ) g u u
u u u
t
i i k i i i
i i i i i i i


+
= +

and energy transport phenomena in reactors. The

studies of hydrodynamic behavior in CFB can be
carried out by experiment and numerical simulation.
A hydrodynamics modeling in CFB is complicate
because of the interaction between gas and solid
phases. The experiments for several conditions of
CFB for design purpose are not possible because of
high expenses and time-consuming processes.
Computer simulation is then considered to be a good
choice to study hydrodynamics behavior in CFB
reactors.
CFB is operated at higher superficial gas velocity
than conventional bubbling fluidized beds (BFB). In
this operating regime, the flow behavior enhances the
performances of the reactors by increasing gas-solid
contacting efficiency, eliminating gas bubbles,
decreasing back-mixing, and also allows catalyst to
be regenerated. The velocity profiles of both gas and
solid phases inside CFB are closer to that of plug flow
while the flow behavior of BFB is more like mixed
flow. This improving flow pattern of CFB makes it
easier to control the selectivity of desired products for
reactions that have multiple products like Fluid
Catalytic Cracking (FCC) process.
In the last decades, computer efficiency is
improved. Hence, they are available for the
complicated simulation of CFB that requires high
computational performances. Therefore, Computer
simulation of CFB becomes an alternative way that
costs less expenses and time comparing to
experiments. Numerical computation can be carried
out by both self modeling and programming or using
commercial programs
In this work, hydrodynamics in a CFB downer
reactor will be investigated by using Two-Fluid Model
available in Fluent software. Effects of operating
conditions such as solid circulation rate on the reactor
performance will be studied.

2. Mathematical model
CFBs have a complex flow behavior such as the
non-uniform spatial distribution of particles, large slip
velocities between the phases and the sensitivity of
the hydrodynamics to the operating parameters.
These effects are important to the performance of the
reactors. Empirical correlations have been used but
they are limited by databases used and neglect the
radial gradient of the basic parameters. However,
some fundamental based models can be used to
predict how various parameters vary with the system
conditions.
In the numerical modeling parts, the flow behavior
of gas and solid can be classified as either
Lagrangian or Eulerian according to the framework in
which they are developed. In the Lagrangian
approach, each particle has been tracked and has its
own equation of motion. In dense flows, the
computational requirements for Lagrangian approach
are extremely high. In the Eulerian approach which is
used in this simulation, each phase has only one
equation of motion and can be applied to the
interesting case with relatively small computational
efforts but these equations contain some terms that
must be chosen carefully to get the accurate results.

2.1 Governing equations
The continuity equations:


with the constraint

where
i
is volume fraction of phase i; u
i
is
velocity vector of phase i.

The conservation of momentum of phase i (i= gas,
solid; k i)




where
i
is stress tensor of phase I; is
interphase drag coefficient.
(1)
(2)
(3)
SM02-2
15
3
( )
s s s s b s s s
S 2 I u P + + =
( ) | | ( ) I u
3
1
u u
2
1
S
S
T
s s s
+ =
( )
s s s s s s s
g e P + + =
0
2
1 2
1
3 / 1
max, , s
s
0
1
5
3
g

(
(

|
|
.
|

\
|

=
96
5
,

s p s
dil s
d
=
( )
2 / 1
0
1
3
4
|
.
|

\
|
+ =

s
p s s b
e g d
s

( )
s s s s s s s
u
t
=

3 :
2
3
( )
( )
( )
2 / 1
0
2
0
0
,
1
5
4
1
5
4
1
1
2
|
.
|

\
|
+ +
(

+ +
+
=

s
p s s
s
dil s
s
g e d
e g
g e

The conservation of solids fluctuating energy, called
the granular temperature ( ), may be written as:


where
s
is collisional dissipation of granular
temperature described by Lun et al (1984).
The last term derived by Gidaspow (1992)
represents the transfer of the kinetic energy of
random fluctuations in particle velocity from solid
phase to gas phase.

2.2 Constitutive equations
Solid phase stress:


Deformation rate:


Solid pressure is derived by Lun et al. (1984) and
composed of a kinetic term and a second term due to
particle collisions:


where e is the restitution coefficient for particle
collisions and the value of 0.9 is used in this
simulation.

Radial distribution function (g
0
) is a correction factor
that modifies the collisional probability for dense
particles flow derived by Ogawa et al. (1980):




Solid phase shear viscosity is derived by Gidaspow
(1992):
(
1
5
)
where
dil s,
is solid phase dilute viscosity and
described as:



Solid bulk viscosity is derived by Lun et al. (1984):



3. Simulation conditions
The downer reactor used in this simulation (Fig. 1)
is 0.1 m. in diameter and 9.3 m. in height. At the top
of the downer, gas and particles are injected together
into the downer column. Fluid Catalytic Cracking
(FCC) catalyst particles used in this simulation is 67
m in diameter and has a density of 1500 kg/m
3
.
The grid size used is 2.5 mm. which creates
148,800 square cells to be calculated for the
simulation.








Fig. 1: Downer column

3.1 Numerical methods
The finite volume method is used to solve the
transport equation together with the first order upwind
differencing scheme. The pressure variable is solved
by SIMPLE (Semi-Implicit Method for Pressure-Linked
Equations) developed by Patankar (1980). The
solution is solved by time-iterative method with the
time interval of 10
-3
seconds.

3.2 Initial and boundary conditions
Initially, the velocity of both gas and solids in the
downer column are set to be zero with no solids
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(4)
U
s
,
s

9.3 m. high
Zero gradient
condition
Non-slip wall
u
g
=0, u
s
=0
U
g
SM02-3
15
4
present in the column. The granular temperature in
downer is initially 10
-4
which is the same as the inlet
granular temperature.
At the downer inlet the velocity and volume
fraction of both phases are specified. For both cases
of simulation, the superficial gas velocity is 10 m/s.
Solids enter the downer column at the velocity of 3.7
m/s. At outlet, the zero gradient condition is used. At
the downer wall, the no slip condition are applied to
both gas and solid phase which will set the normal
and tangential velocities to be zero at the wall.

4. Results and discussions
The hydrodynamic behavior in the downer column
such as solid fraction and velocity are illustrated
throughout the column. In addition, the effect of solid
circulating rate (G
s
) on hydrodynamic behavior is
observed.

4.1 Axial profiles of particle velocity and volume
fraction
Axial solid holdup profile is plotted in Fig. 2. The
solid holdup drops dramatically in the first 2 metres
below the entrance and becomes stable down the
column. Solid holdup might be affected by the axial
velocity profile of solids that is shown in Fig. 3. The
increasing solid velocity along the downer decrease
solid holdup to maintain constant flux at different
downer section.
For both cases, solids enter the downer with small
velocity magnitude at the value of 3.7 m/s. The
superficial gas velocity is set to be 10 m/s. The
difference between the gas velocity and the solid
velocity creates enormous drag force. At the top of
downer, solid velocity increases rapidly because the
big drag force and the gravity work in the same
direction. Further down the column below 1.5 m.,
solid velocity increases slower because of very small
or negative drag force. After the distance of 3 m.
from downer top, solid velocity becomes constant
because the negative drag force increases until it
balance with the gravity force which remains constant
throughout the downer column.
To see the effects of increasing solids circulation
rate (G
s
) clearly, gas superficial velocity and solid
feeding velocity are fixed to compare case 1 (G
s
=100
kg/m
2
s) and case 2 (G
s
=200 kg/m
2
s). The solid
volume fraction is double from case 2 at the entrance
of downer to obtain 200 kg/m
2
s of solid entrance flux.
In Fig. 2 , the solid holdup of case 2 is approximately
two times greater than solid holdup of case 1 due to
the increasing solid circulation rate.
0.006
0.008
0.010
0.012
0.014
0.016
0.018
0 1 2 3 4 5 6 7 8 9
Distance from Downer Top (m.)
S
o
l
i
d

H
o
l
d
u
p
Case 1 (Gs=100 kg/m2s)
Case 2 (Gs=200 kg/m2s)
Fig. 2: Axial solid holdup profile at G
s
equal to 100
and 200 kg/m
2
s

In Fig.3 the solid velocity drops about 0.1 m/s as a
result of double solid circulation rate. Normally, at
this high solid circulation rate, there might be particle
clusters form in the wall region. Because clustering
of solid particles reduce the effective drag force, the
final solid velocity will be higher than the case of
lower solid circulation rate if clustering exists. The
lower final solid velocity by increasing solid circulation
rate to 200 kg/m
2
s implys that there is no clustering of
solid particles in the downer even at this high solid
circulation rate which is better than the upflow riser.
SM02-4
15
5
8.5
9.0
9.5
10.0
10.5
0 1 2 3 4 5 6 7 8 9
Distance from Downer Top (m.)
S
o
l
i
d

V
e
l
o
c
i
t
y

(
m
/
s
)
Case 1 (Gs=100 kg/m2s)
Case 2 (Gs=200 kg/m2s)
Fig. 3: Axial solids velocity profile at G
s
equal to 100
and 200 kg/m
2
s

4.2 Radial profiles of particle velocity and volume
fraction
The radial profile of solid holdup shown in Fig.4
present a peak at r/R = 0.6-0.8 as reported by many
researchers (Yang et al., 1991; Bai et al., 1991; Wang
et al., 1992). The solid first accumulate at the wall
and the peak of solid holdup moves toward center
with more distance from the top entrance. In case 1
with solid circulation rate of 100 kg/m
2
s, the flat core
has a uniform solid holdup at 7*10
-3
but the peak
solid holdup has a value more than 8 *10
-3
which is
15 % more. But in case 2, the peak holdup is almost
30 % more than the core holdup. The solid hold up
in the annulus is about 2-3 times lower than the core
holdup in both cases. The low solid holdup in the
annulus decreases the effective volume of the
reactor. Solid velocity also has a small value in the
annulus as well as gas velocity (not shown here).
Since there is not much solid presented in this
annulus region, the effect of this phenomenon on the
solid residence time distribution is small.
The radial solid velocity profiles of both cases in
Fig.5 have the same shape. Solid velocity in the core
is uniform and decreases toward the wall until the
value reaches zero at the wall according to the non-
slip boundary condition of the solid phase.
The solid velocity profile in a downer separates to
core and annulus zones because of low gas velocity
in the wall region and high gas velocity near the
center. This solid velocity profile can be explained as
a result of the mixed profile of relatively uniform
profile of solids velocity and the parabolic profile of
gas velocity if they are simulated separately.
Increasing solid circulation rate, solid velocity profile
of case 2 provides wider uniform core with lower
value than that of case 1. The reason is that more
content of particles induce more plug flow behavior of
particle velocity in the higher solid circulation rate
condition. The same result has also been reported by
Zhang et al. (2000).

5. Conclusions
In this simulation, the two-fluid model is used to
characterize gas-solid flow systems through the
application of solid vibrating energy (granular
temperature). The model can predict the solid holdup
peak in downer reactor which was also reported by
many researchers for their experiments. The flow in
a downer is separated into core and annulus zone.
There is relatively high solid concentrantion and
velocity with uniform profiles in the core. On the
contrary, solid phase in the annulus is dilute and has
small velocity. Axial solid holdup is doubled when
increase solid circulation rate twice, hence assumed
to be linearly affected by increasing solid circulation
rate. The decreasing axial solid velocity when
increasing solid circulation rate is due to more
interaction among solid particles. The higher solid
concentration also makes solid velocity profile a wider
core because more content of particles induce more
plug flow behavior of particle velocity. However, this
is not seen clearly here. This effect may be more
obvious if a new boundary condition is set to allow
solids to slip at the wall. These facts agree that a
downer reactor is suitable for processes that require
uniform RTD of solid particles.


SM02-5
15
6
H=9.155 m.
H=2.112 m.
H=4.398 m.
H=9.155 m.
Fig. 4: Radial solid holdup profile at different distances from downer top

Fig. 5: Radial solid velocity profile at different distances from downer top
H=0.020 m.
0
5
10
15
S
o
l
i
d

V
e
l
o
c
i
t
y

(
m
/
s
)
0
5
10
15
-0.05 -0.03 -0.01 0.01 0.03 0.05
Radial Position (m.)
S
o
l
i
d

V
e
l
o
c
i
t
y

(
m
/
s
)
0
5
10
15
S
o
l
i
d

V
e
l
o
c
i
t
y

(
m
/
s
)
0
5
10
15
-0.05 -0.03 -0.01 0.01 0.03 0.05
Radial Position (m.)
S
o
l
i
d

V
e
l
o
c
i
t
y

(
m
/
s
)
0.0111
0.0111
0.0112
0.0112
0.0113
0.0113
0.0114
S
o
l
i
d

H
o
l
d
u
p
0.000
0.002
0.004
0.006
0.008
0.010
S
o
l
i
d

H
o
l
d
u
p
0.000
0.002
0.004
0.006
0.008
0.010
-0.05 -0.03 -0.01 0.01 0.03 0.05
Radial Position (m.)
S
o
l
i
d

H
o
l
d
u
p
0.0223
0.0223
0.0224
0.0224
0.0225
0.0225
S
o
l
i
d

H
o
l
d
u
p
0.000
0.005
0.010
0.015
0.020
S
o
l
i
d

H
o
l
d
u
p
0.000
0.005
0.010
0.015
0.020
-0.05 -0.03 -0.01 0.01 0.03 0.05
Radial Position (m.)
S
o
l
i
d

H
o
l
d
u
p
Case I Case II
Case I Case II
SM02-6
15
7

6. Notations
d
p
Particle diameter
e Particleparticle restitution coefficient
g
0
Radial distribution function
k
s
Diffusion coefficient of granular temperature
P Pressure
P
s
Solids pressure
Re Particle Reynolds number

I
Stress tensor of phase i
U
i
Velocity vector of phase i
Interphase drag coefficient

i
Volume fraction of phase i

s
Collisional dissipation of granular temperature

i
Viscosity of phase i

i
Density of phase i

s
Granular temperature

7. Acknowledgements
This research was supported by the Thailand
Research Fund (TRF) under Research Career
Development Project, the Kasetsart University
Research and Development Institute (KURDI), CHE-
ADB Graduate Research and Education Development
Program in Chemical Engineering at Kasetsart
University.

8. References
Bai, D.-R., Jin, Y.,Yu, Z.-Q., and Gan, N.-J. 1991.
Radial profiles of local solid concentration and velocity
in a concurrent downflow fast fluidized bed. Circulating
fluidized bed technology III. Pergamon Press,
Toronto.
Deng, R., Wei, F., Liu, T., and Jin, Y. 2002.
Radial behavior in riser and downer during the FCC
process. Chemical Engineering and Processing. 41:
259-266.
Ergun, S. 1952. Fluid flow through packed columns.
Chemical Engineering Progress. 48: 89-94.
Gidaspow D., R. Bezburuah, and J. Ding. 1992.
Hydrodynamics of Circulating Fluidized Beds, Kinetic
Theory Approach. In Fluidization VII, Proceedings of
the 7th Engineering Foundation Conference on
Fluidization, pages 75-82
Grace, J.R., Avidan, A.A., and Knowlton,
T.M.1997. Circulating Fluidized Beds. Chapman &
Hall.
Herbert, P.M., Gauthier, T.A., Briens C.L.,
Bergougnou, M.A. 1995. Application of fiber optic
reflection probes to the measurement of local particle
velocity and concentration in gas-solid flow. Powder
Technology. 80: 243-252.
Kawaguchi, T., Sakamoto, M., Tanaka, T., and
Tsuji, Y. 2000. Quasi-three-dimension of spouted
beds in cylinder. Powder Technology. 109: 3-12.
Kimm, N.K., Berruti, F., and Pugsley, T.S. 1996.
Modeling the hydrodynamics of downflow gas-solids
reactors. Chemical Engineering Science. 51(11):
2661-2666.
Kunii, D., and Levenspiel, O. 1997. Circulating
fluidized-bed reactors. Chemical Engineering
Science. 52(15): 2471-2482.
Lun C. K. K. , S. B. Savage, D. J. Jeffrey, and N.
Chepurniy. 1984. Kinetic Theories for Granular Flow:
Inelastic Particles in Couette Flow and Slightly Inelastic
Particles in a General Flow Field. J. Fluid Mech.,
140:223-256,
Ogawa S., A. Umemura, and N. Oshima. 1980.
On the Equation of Fully Fluidized Granular Materials.
J. Appl. Math. Phys., 31:483,
Wang Z., Bai, D. and Jin, Y. 1992.
Hydrodynamics of cocurrent downflow circulating
fluidized bed (CDCFB). Powder Technology. 70:
271-275.
Zhang, H., and Zhu, J-X. 2000. Hydrodynamics
in downflow fluidized beds (2): Particle velocity and
solids flux profiles. Chemical Engineering Science.
55: 4367-4377.
Zhang, H., Zhu, J-X., and Bergougnou, M.A.
1999. Hydrodynamics in downflow fluidized beds (1):
solids concentration profiles and pressure gradient
SM02-7
15
8
distributions. Chemical Engineering Science. 54:
5461-5470.
Zhang, M., Qian, Z., Yu, H., and Wei, F. 1999.
The near wall dense ring in a large-scale down-flow
circulating fluidixzed bed. Chemical Engineering
Science. 54: 5461-5470.
Zhang, M., Qian, Z., Yu, H., and Wei, F. 2003.
The solid flow structure in a circulating fluidized bed
riser/downer of 0.42-m. diameter. Powder
Technology. 129: 46-52.
Zhu, J.-X., and Manyele, S.V. 1998. Radial non-
uniformity index (RNI) in fluidized beds and multiphase
flow systems. 48
th
Canadian Chemical Engineering
Conference. London.

SM02-8