Group Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Title: 6.2.5 - Coupled DEM-CFD simulations of a Wurster-granulator
ALL VOLUMES CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00102023/00151
 Material Information
Title: 6.2.5 - Coupled DEM-CFD simulations of a Wurster-granulator Particle Bubble and Drop Dynamics
Series Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Physical Description: Conference Papers
Creator: Fries, L.
Antonyuk, S.
Heinrich, S.
Palzer, S.
Jacob, M.
Publisher: International Conference on Multiphase Flow (ICMF)
Publication Date: June 4, 2010
 Subjects
Subject: Discrete-Element-Method
Wurster-granulator
Fluidized bed spray granulation
CFD
 Notes
Abstract: Coupled DEM-CFD simulations have been performed to study the fluid and particle dynamics of a Wurster-granulator on the scale of individual particles. The effect of process parameters like air flow rate and geometry details like the Wurster position has been studied. The collision behaviour of dry g-Al2O3 particles was identified experimentally and incorporated into the model. Based on a physical description of the material properties, an effective tool for design and scale-up of a Wurster granulator is obtained. The homogeneity of wetting and mixing in a Wurster granulator is evaluated based on the residence time distribution of the particals inside a conical spray zone, which allows an estimation of the wetting intensity of the individual particle surfaces.
General Note: The International Conference on Multiphase Flow (ICMF) first was held in Tsukuba, Japan in 1991 and the second ICMF took place in Kyoto, Japan in 1995. During this conference, it was decided to establish an International Governing Board which oversees the major aspects of the conference and makes decisions about future conference locations. Due to the great importance of the field, it was furthermore decided to hold the conference every three years successively in Asia including Australia, Europe including Africa, Russia and the Near East and America. Hence, ICMF 1998 was held in Lyon, France, ICMF 2001 in New Orleans, USA, ICMF 2004 in Yokohama, Japan, and ICMF 2007 in Leipzig, Germany. ICMF-2010 is devoted to all aspects of Multiphase Flow. Researchers from all over the world gathered in order to introduce their recent advances in the field and thereby promote the exchange of new ideas, results and techniques. The conference is a key event in Multiphase Flow and supports the advancement of science in this very important field. The major research topics relevant for the conference are as follows: Bio-Fluid Dynamics; Boiling; Bubbly Flows; Cavitation; Colloidal and Suspension Dynamics; Collision, Agglomeration and Breakup; Computational Techniques for Multiphase Flows; Droplet Flows; Environmental and Geophysical Flows; Experimental Methods for Multiphase Flows; Fluidized and Circulating Fluidized Beds; Fluid Structure Interactions; Granular Media; Industrial Applications; Instabilities; Interfacial Flows; Micro and Nano-Scale Multiphase Flows; Microgravity in Two-Phase Flow; Multiphase Flows with Heat and Mass Transfer; Non-Newtonian Multiphase Flows; Particle-Laden Flows; Particle, Bubble and Drop Dynamics; Reactive Multiphase Flows
 Record Information
Bibliographic ID: UF00102023
Volume ID: VID00151
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: 625-Fries-ICMF2010.pdf

Full Text


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010



Coupled DEM-CFD simulations of a Wurster-granulator


L. Fries*, S. Antonyuk*, S. Heinrich*, S. Palzert and M.Jacob#


Institute of Solids Process Engineering, Hamburg University of Technology, Hamburg, Germany
SNestle Research Center Lausanne, Vers-Chez-Les-Blanc, Switzerland
# Glatt Ingenieurtechnik GmbH, Weimar, Germany
lennart.fries@tuhh.de, stefan.heinrich@tuhh.de and stefan.palzer@rdls.nestle.com


Keywords: Discrete-Element-Method, Wurster-granulator, Fluidized bed spray granulation, CFD





Abstract

Coupled DEM-CFD simulations have been performed to study the fluid and particle dynamics of a Wurster-granulator on the
scale of individual particles. The effect of process parameters like air flow rate and geometry details like the Wurster position
has been studied. The collision behaviour of dry y-A1203 particles was identified experimentally and incorporated into the
model. Based on a physical description of the material properties, an effective tool for design and scale-up of a Wurster
granulator is obtained. The homogeneity of wetting and mixing in a Wurster granulator is evaluated based on the residence
time distribution of the particals inside a conical spray zone, which allows an estimation of the wetting intensity of the
individual particle surfaces.


Introduction

Fluidized bed spray granulation plays an important role in
the manufacturing of powder granules in the food and
pharmaceutical industries as it allows producing dust-free
and free-flowing particles. Liquid binder is sprayed into a
bed of solids to achieve granule growth, therefore the
moisture distribution in the apparatus is a key parameter
affecting both particle size and structure of the product
(Palzer 2008). The Wurster-granulator (Wurster, 1966) is a
common technique in the pharmaceutical industry used to
coat tablets. A cylindrical draft tube is inserted vertically
into the granulator, as shown in Fig. 1.


expansion
chamber




Wurster tube
annulus
nozzle


segmented
distributor
plate:
zone 1
zone 2
zone 3 -


Figure 1: Scheme of a Wurster-granulator

Combined with a bottom-spray injection nozzle, the Wurster
geometry induces a circulating movement of the particles
and divides the geometry into two zones. In the central part
inside the Wurster tube the particles are transported upwards
in a spout. The particles decelerate in the expansion


chamber and fall down to the dense region of particles
outside the tube, while they are dried by the warm
fluidization air. From the dense region, the particles are
transported back into the Wurster tube and the circulating
motion in the bed is repeated.
The Wurster process is controlled by many parameters
which can influence the quality of the coating layer. Since
operating conditions and the equipment geometry have been
developed empirically the actual influence of the
fundamental mechanisms in the process is not well
understood. The drying rate is highly dependent on the flow
field of the gas and particle phases in the equipment
To describe the process in detail on the scale of individual
particles, the Discrete-Element-Method (DEM) offers large
potential. As each particle is tracked individually, the
method allows a complete representation of the
particle-particle and particle-wall interactions and their
influence on the process dynamics.
In this contribution a DEM model is coupled with a CFD
simulation to consider the interactions between fluid and
particle phase using the Euler-Lagrange approach. Two
simulation studies are presented: the first study investigates
the fluid- and particle dynamics in a Wurster granulator and
highlights the effect of process parameters like the spout
velocity and the gap distance below the Wurster tube. The
second study analyses the wetting behaviour of the particles
based on their residence time distribution in a conical spray
zone at the nozzle tip. Details of the models will be given in
the next section.

Mathematical model

Following the concept of the kinetic theory of granular flow
(KTGF), multiphase CFD models have been established to
describe the fluid dynamics of spout fluidized bed systems









(Gryzka et al. 2009), (Karlsson et al., 2009). The method
allows simulating a lab-scale fluidized bed on a PC
workstation within a reasonable period of time. Yet, the
method fundamentally lacks a description of the
particle-particle interactions. To be able to model a fluidized
bed system on the scale of individual particles, several
authors have set up discrete particle models (DPM)
(Hoomans et al., 1996), (Tsuji et al., 1993). Some attempts
have been made to model the fluidized bed spray
granulation process with DPM (Goldschmidt, Kuipers,
2003), (Gantt, Gatzke, 2005), (Kafui, Thornton, 2008). The
concept of the modelling technique is briefly introduced in
this section.
The motion of each particle i in the system is calculated
using Newton's second law:

mi, dv = -VVp+ VJ (u, -v,)+m,)+Fm,g ,c,(1)
dt 1-1
The forces on the right hand side of Eq. 1 are respectively
due to pressure gradient, drag, gravity and contact forces (i.e.
due to collisions). The interphase momentum transfer
coefficient p is modelled by combining the Ergun equation
(1952) for dense regimes (e<0.8) and the correlation
proposed by Wen&Yu (1966) for more dilute regimes
(E > 0.8).
The gas phase is considered as continuum. The geometry of
the apparatus is discretized in mesh cells and the motion of
the gas phase is calculated using volume-averaged
Navier-Stokes equations.

a(Epg)+ V (FpgUg) =0 (2)


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

= 2a'. -k 6 (5)

e, ife, f O (6)
a= 1 2+lne,
1 if e, = 0
Here, a( is a function of the coefficient of restitution e,
which is defined as the ratio of rebound velocity Vr to
impact velocity v.

e = 1 (7)
rl V
Ekn
It can be determined experimentally with a free-fall tester
(Antonyuk, Heinrich et al. 2010).

Numerical simulations

For numerical DPM simulations, a Glatt GF3 insert was
modelled. The original geometry was supplied by Glatt
Ingenieurtechnik GmbH (Weimar, Germany) and slightly
simplified for the simulations, as shown in Fig. 2a. A
representation of the mesh used for the fluid dynamics
simulations is given in Fig. 2b. The commercial software
packages EDEM and Fluent were used to perform
simulations with 150,000 spherical particles of a particle
diameter dp = 2 mm, which corresponds to a batch size of
0.94 kg at an average particle density of 1500 kg/m3.


-EVP -V(er,)-S,+ep,g


Due to their presence and the volume fraction they require,
the particles influence the velocity profile of the gas phase.
This effect is accounted for by adding a sink term Sp to the
momentum balance, which contains the interphase
momentum transfer coefficient p and closes the
two-way-coupling. Forces between the gas and particle
phase are of opposite direction and equal magnitude. The
flow around particles is not fully resolved, as the size of the
fluid mesh cells is larger than the particle diameter.
Contact forces between particles are calculated according to
a contact model based on the theory developed by Hertz
(1882) for the normal impact and a no-slip approximation of
the model by Mindlin and Deresiewicz (1953) for the
tangential component, as proposed by Tsuji (1992).

Fab,n = -kn, 9.- nab rlnvab,, (4)
The elastic part of the contact force is represented by a
non-linear spring, where the force is proportional to the
spring stiffness kn and the displacement S .
Additionally, to account for visco-elastic material properties
that cause energy dissipation, a damping factor r1n related to
the coefficient of restitution (Tsuji 1992) is included in the
model.


slice 2 .


;lice1
m ",'


Wurster
tube
Fixation


Figure 2: (a) Simplified Wurster geometry for DEM-CFD
simulation. (b) Fluid mesh of 12,000 tetrahedral cells
For y-A1203 particles a constant coefficient of restitution of
0.8 was set in this study, according to experimental results
by Antonyuk et al. (2010). Further simulation parameters
are summarized in Table 1.


Table 1: Overview on simulation parameters

Fluidization velocity Wurster (zone 1) 8 m/s = 10.1 umf
Fluidization velocity ring (zone 2) 3 m/s = 3.8-umf
Fluidization velocity edge (zone 3) 4 m/s = 5.1 umf
Nozzle injection velocity 20 160 m/s
DEM simulation time step le-6 s
CFD simulation time step le-4 s
Number of grid cells (CFD) 12,000


a (pug9)+V(ep, gug)
at








Air is introduced to the apparatus at the bottom via a
segmented distributor plate that consists of three zones.
Below the Wurster tube the fluidization air velocity is
higher than in the ring-zone due to a larger porosity of the
plate. In the third zone near the wall of the apparatus a
slightly higher porosity of the plate and therefore higher gas
velocity is applied to avoid dead zones Additionally, air is
injected by a nozzle which is situated at the center inside the
Wurster tube. This study focuses on the fluid-dynamic
behaviour and does not include the injection of liquid
binder.

Table 2: Simulation scenarios in this work
Case Uspout wurster Uring air flow ratio hgap
[m/s] [m/s] [m/s] wurster/ring [mm]
1 20 8 4 29,78% 10
2 100 8 4 29,78% 10
3 160 8 3 41,98% 10
4 160 8 3 41,98% 20

In a series of case studies, the velocity of the inlet air, the jet
injection velocity spout, the distribution of air between
Wurster and annulus and the gap distance hgap below the
Wurster tube was varied. Material parameters were kept
constant throughout the simulations as well as the height hw
and diameter dw of the Wurster tube. The operating
conditions in the case studies are summarized in Table 2.

Simulation results

Influence ofgas velocity

The simulation results show that the velocity of the air
injected via the nozzle has a strong influence on the
fluid-dynamics of the whole granulator. In Fig. 3, snapshots
of the particle positions and their velocities are displayed.
The colour indicates the particle velocity magnitude: blue
particles move slowly (v < 0.5 m/s) and red particles are fast
(v> 1.5 m/s). Three simulation cases are depicted at the
same time step (t = 1.4 s). The visual representations of the
simulation results in Fig. 3 and Fig. 5 show the central
vertical cross section of the apparatus and the whole volume
behind that plane.
It can be seen from Fig. 3 that a higher spout velocity causes
higher particle velocities inside the Wurster tube. In the
acceleration zone near the nozzle, the particles are
concentrated to the center of the Wurster for high spout
velocities (case 3) whereas they are more evenly distributed
over the whole tube for low spout velocities (case 1). The
height of the particle fountain increases with higher spout
velocity from 430 mm above the nozzle tip in case 1 to
490 mm in case 3.
Fig. 4 shows the velocity field of the fluid in the central
vertical plane at time t = 1.4 s for the three simulation cases.
Jet velocities higher than 14 m/s are cut off and displayed in
white colour in the graphic. It can be seen that the nozzle jet
has a low injection depth. Even at the highest injection
velocity of 160 m/s at the nozzle tip (case 3), the fluid
velocity decays to 14 m/s within less than 30 mm from the
tip. The momentum introduced by the jet is immediately
transferred to the particles in a relatively small zone above
the nozzle tip. In case 1, the penetration depth of the jet


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

injected at 20 m/s is almost invisible. In contrast to the
particle velocity distribution, the gas flow field inside the
Wurster tube is hardly influenced by the jet velocity.
m/s (a) Case 1 (b) Case 2


(c) Case 3


Figure 3: Instantaneous particle positions and velocity
distributions inside the Wurster tube at time t = 1.4 s.
Colours indicate the velocity magnitude


m/s
140
12.6
1.05
8.40
6.30
4.20
S210
oo00


(a) Case 1


1%?1


(a) Case 2


(c) Case 3


Figure 4: Instantaneous fluid velocity distribution in the
central vertical plane of the Wurster-granulator at time
t = 1.4 s. Colours indicate the velocity magnitude


,1 F









This is due to the fact that even in case 3 less than 2 % of
the total gas flow is injected via the nozzle. More than 98 %
of the gas enters the system via the distributor plate at the
bottom of the granulator.


(b) Case 2


Fixation of the
Wurster tube to
the apparatus wall


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

faster than 1 m/s at this level (case 3). For low spout
velocities as in case 1, some particles are moving
downwards at the inside of the Wurster tube, which
indicates unwanted back-mixing.


(a) Case 1


(b) Case 3


Figure 6: Instantaneous horizontal distribution of particles
in slice 2 at time t = 1.4 s

Influence of gap distance below the Wurster tube

The gap distance between the distributor plate and the
Wurster tube is a parameter that strongly influences the
particle dynamics inside the granulator as it controls the
recirculation of particles into the spout. Gaps of 10 mm and
20 mm were compared.


-




Figure 5: Instantaneous horizontal distribution of particles
in slice 1 at time t = 1.4 s

To assess the radial distribution of particles in the granulator,
horizontal slices were cut out of the simulated geometry.
This was done at two different heights, as indicated in
Fig. 2a. The first slice is situated in the lower part of the
granulator, just below the tip of the injection nozzle. The
second slice contains the upper border of the Wurster tube
and allows visualizing the flow conditions of particles
entering the expansion. The thickness of each of the slices is
10 mm.
Fig. 5 and 6 show instantaneous particle positions in the two
slices, seen from the top. Colours indicate the magnitude of
the vertical component of the particle velocity. Red particles
are transported upwards, blue particles fall downwards. A
comparison can be drawn between the first 3 case studies to
evaluate the influence of the spout velocity on the horizontal
distribution of the particles. Comparing case 2 and 3 in
slice 1 it can be seen that the gas velocity in the outer ring
has a strong influence on the bed expansion. For case 3, a
dense bed is seen in slice 1 where there is a high porosity in
case 2. Looking at the particles inside the Wurster tube it
can be observed that the high spout velocity in case 3 tears
the particles towards the center of the tube. A comparison
between case 1 and 2 in Fig. 5 reveals that at identical flow
conditions in the annulus, the gas injection affects the
bubble formation in the annulus. At higher spout velocity
(case 2) larger bubbles appear.
In slice 2 at the upper end of the Wurster tube, only few
particles are present (Fig. 6). Their movement is directed
upwards inside the tube and downwards in the annulus at all
three cases, indicating that the circulating regime is intact in
a wide fluidization range. Along the border of the Wurster
tube, a deceleration of the particles can be observed in
case 1, whereas at high spout velocity, all particles mover


Case 3


Case 4


10
Figure 7: Instantaneous particle position and velocity
distributions in the Wurster zone for different gap distances
at time t = 1.4 s

It can be seen in Fig. 7 that a larger gap distance below the
Wurster increases the number of particles that are
transported into the tube. In case 3, in the lower half of the
Wurster particles are only present in the center of the tube.
All particles rise at a velocity magnitude above 1 m/s which
is indicated by green and red colour. Contrary to that, in
case 4 particles can be found spread over the whole
diameter of the Wurster tube. Near the wall of the tube they
move at low velocity, indicated by blue colour.

Residence time in the spray zone

In the Wurster-granulator, the residence time of the particles
inside the spray zone is primarily determined by the upflow
velocity of gas and particles in the draft tube. Once a stable
circulating regime is established in the apparatus, the
particles are homogeneously wetted. For the given process
conditions the particles spend on average 25 ms per cycle
inside the spray zone, as illustrated in Fig 8. The residence


(a) Case 1


(c) Case 3









time distribution obtained after 5 s simulation time shows a
clear peak at 25 ms and smaller peaks multiples of this
period. These are related to particles, which have passed
through the spray zone more than once. If the simulation is
continued, a narrow discrete Gaussian distribution can be
expected which will continuously shift to the right with
time.
The average particle concentration inside the spray zone is
low very low for this case, indicating the gap distance below
the Wurster tube (10 mm) should be increased.


20000


2000


. 1600


S1200
0
E 800
c


0 iI II#III "#III 1111J II 1iii 111111!IN Ii1111 1 0 i.
0.000 0.010 0.020 0.030 0.040 0.050
residence time [s]
Figure. 8: Residence time distribution of particles in the
spray zone for the Wurster-granulator

In the Wurster-granulator all particles are transported
through the spray zone at constantly high velocity (driven
by the spout) which is not the case for a standard top spray
fluid bed granulator. For this reason, the Wurster-granulator
achieves especially homogeneous wetting of the particles
which can well be approved by the numerical simulations
performed in this study.

Table 3: Motion of particles in the spray zone


Mean particle fraction inside spray zone
Average particle velocity magnitude
Avg. ratio of residence time to simulation time


1%
5 m/s
1%


Conclusions

Process parameters like the spout velocity and geometrical
settings like the Wurster gap height influence the fluid- and
particle dynamics of a Wurster-granulator, which was
analysed with the help of coupled DEM-CFD-simulations.
Based on experimental data of the impact behaviour of
visco-elastic particles, the fluidization regime and based on
that the functionality of a Wurster granulator can be
predicted. The homogeneity of the liquid distribution among
the particle phase was investigated numerically. The
residence time distribution of the particles in a conical spray
zone at the tip of the injection nozzle allows estimating the
wetting intensity of the individual particles. The results
show that the Wurster-granulator is characterized by a


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

narrow residence time distribution. Due to the directed flow
regime in the Wurster-granulator, the wetting of the
particles in this type of apparatus is particularly
homogeneous
Future measurements in a real Wurster granulator will
deliver results to validate the model. Granulation
experiments will underline the importance of a
homogeneous liquid distribution in the granulator for the
size distribution, shape and structure of the products.

Acknowledgements

We would like to thank Nestec S.A., Switzerland for the
financial support and Dipl.-Ing. Michael Jacob from Glatt
Ingenieurtechnik GmbH, Weimar, Germany for the supply
of geometry data.

Nomenclature


Diameter of Wurster tube (m)
Coefficient of restitution (-)
Force (N)
Gravitational acceleration (ms-2)
Height of Wurster tube (m)
Spring stiffness (Nm-1)
Mass (kg)
Normal unit vector (-)
Pressure (Pa)
Particle position (m)
Particle drag sink term (Nm3)
Time (s)
Gas velocity (ms-1)
Particle velocity (ms-1)
Relative velocity at the contact point (ms-1)
Volume (m3)


Greek letters
p Interphase momentum transfer coef. (kgm3s-1)
8 Displacement (m)
e Porosity (-)
1 Damping coefficient (Nsm 1)
p Density (kgm 3)
T Gas phase stress tensor (Pa)

Subscripts
f Fluid
g gas
i Particle index
n Normal direction
r Rebound
t Tangential direction

References

Antonyuk, S., Heinrich, S., Tomas, J., Deen, N.G., van
Buijtenen, M.S. and J.A.M. Kuipers: Energy absorption
during compression and impact of dry elastic-plastic
spherical granules, in Proof in Granular Matter, December
2010.

Gantt, J.A., Gatzke, E.P.: High-shear granulation modeling
using a discrete element simulation approach. Powder
Technology 156 (2005) pp. 195-212.






7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010


Goldschmidt, M.J.V., Kuipers, J.A.M.: Discrete element
modelling of fluidised bed spray granulation. Powder
Technology 138, 2003, 39-45.

Gryczka O., Heinrich S., Deen N.G., van Sint Annaland, M.,
Kuipers, J.A.M., Jacob, M., Morl, L.: Characterization and
CFD-modeling of the hydrodynamics of a prismatic spouted
bed apparatus. Chemical Engineering Science 64 (14), 2009,
3352-3375.

Hertz, H. Uber die Beriihrung fester elastischer Korper.
Journal fur die Reine undAngewandte Mathematik 92, 1882,
156-171.

Hoomans, B.P.B., Kuipers, J.A.M., Briels, W.J., Van Swaaij,
W.P.M.: Discrete particle simulation of bubble and slug
formation in a two-dimensional gas-fluidised bed: a
hard-sphere approach. Chemical Engineering Science 51 (1),
1996,99-118.

Kafui, D.K., Thornton, C.: Fully-3D DEM simulation of
fluidised bed spray granulation using an exploratory
surface-energy based spray zone concept. Powder
Technology 184, 2008, 177-188.

Karlsson, S., Rasmuson, A., van Wachem, B, Niklasson
Bjorn, I.: CFD Modeling of the Wurster Bed Coater. AIChE
Journal 55 (10), 2009, pp. 2578-2590.

Mindlin, R.D., Deresiewicz, H.: Elastic spheres in contact
under varying oblique forces. Transactions ofASME, Series
E. Journal of applied Mechanics 20, 1953, 327-344.

Palzer, S.: Influence of material properties on the
agglomeration of water-soluble amorphous particles.
Powder Technology 189 (2), 2009, pp. 318-326.

Tsuji, Y., Tanaka, T., Ishida, T.: Lagrangian numerical
simulation of plug flow of cohesionless particles in a
horizontal pipe. Powder Technology 71, 1992, 239-250.

Tsuji, Y., Kawaguchi, T., Tanaka, T.: Discrete particle
simulation of two-dimensional fluidized bed. Powder
Technology 77 (1), 1993, 79-87.

Wurster, D. E.: Particle Coating Process. U.S. Pat. No.
3,253,944 (Wisconsin Alumni Research Foundation, May
31, 1966)




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.7 - mvs