CFD Modelling of Fluidized Bed Reactors

2Preface
Fluidization technology is employed in a wide range of industrial opera-
tions, ranging from the pharmaceutical and food industry, to processes such
as catalytic cracking of petroleum, combustion and biomass gasification. Flu-
idization is also recognized to be the most promising technology to optimize
the available potential of renewable energy. Small-scale pilot plants are often
used to demonstrate a process’ viability and to make estimates of a com-
mercial plants performance. The characteristics of a fluidized bed process
greatly depends on the hydrodynamic behaviour of the system. However,
achieving a fundamental understanding of the mechanisms governing the be-
haviour of fluidized beds particularly under industrial operating conditions
still represents a major scientific and engineering challenge. Reliable designs
of commercial scale plants requires not only a good understanding of the
highly complex flow phenomena, but also detailed knowledge of how the hy-
drodynamics are affected by both geometry changes and plant scale-up. De-
velopment of fluid-bed models capable of correctly predicting the fluidization
phenomena would be highly desirable and would allow for both the cheaper
9
Preface 10
and faster development of innovations in fluidized bed processes.
In the chemical industry, computational fluid dynamics (CFD) modelling
has been recognized as having the potential for increasing process efficiency
by reducing plant downtimes and for improving large-scale technology deliv-
ery by decreasing the number of scale-up steps. In recent years, substantial
mathematical effort is being directed toward the development and validation
of CFD codes for multiphase applications, including fluidization.
During this study, CFD modelling simulations of gas-fluidized beds will
be performed using a commercial code developed by AEA Technology, CFX-
4.4 The Eulerian Granular model, which is already available within CFX-
4.4, will be used during this work. The model is based on the two-fluid
model (TFM) that treats each phase as an interpenetrating continuum and
solve mass and momentum equations for each phase. The Eulerian Granular
model (GKM) uses kinetic theory to describe properties of the particle phase
where, resembling the gas kinetic theory, the particle-particle interactions are
described as binary collisions. According to this model approach, in addition
to mass and momentum, an energy balance is solved to determine the kinetic
energy associated to the particle fluctuations and a coefficient of restitution,
0 < e < 1, is introduced to account for the inelasticity of the solids and thus
the particle-particle interactions.
The purpose of this work is to investigate the predicting capability of
GKM in respect of different hydrodynamic patterns, by means of time-
dependent 2D simulations of two problems of industrial importance: transi-
tion from bubbling to slugging fluidization of a Geldart [22] Group B powder,
and flow field of volatile matter generated by a fuel particle immersed in bed
10
Preface 11
of incipient fluidization. Both qualitative and quantitative results will be
presented, including bubble formation and coalescence, pressure drop across
the bed, fluid-bed voidage, bed expansion and bubble size. Results obtained
from simulations will be compared with available experimental data.
Simulations will be performed using a 1.7GHz Dual Pentium 4 Dell 530
workstation, available in Chemical Engineering Department of University
College London1
1Torrington Place, London WC1E 7JE, UK
11
3Introduction
3.1 Fuidization
An always increasing number of industrial applications, spacing from the
field of petrochemistry to alimentary, pharmaceutical and textile technolo-
gies, are based on the processes of fluidization, designed for the first time in
1942 for aviation gasoline Catalytic Cracking production, by a group of oil
companies, which included Standard Oil Indiana 1, the M.W Kellogg, Shell
and the Standard Oil Development Company 2.
“Fluidization is the operation by which particles are transformed into a
fluid-like state through suspension in a gas or liquid” 3. Typically a fluidized
bed is a cylindrical vessel, filled with granular solids particles, through which
it is introduced an upward flow rate. Based on the superficial fluid velocity,
operational conditions, as temperature and pressure, and the physical prop-
1
later Amoco and now BP Amoco
2now Exxon
3by Kunii and Levenspiel (1991) [38]
12
Introduction 13
erties of the particles4 this two-phase system can display a large range of
hydrodynamical behaviours. For low fluid velocity the solid phase remains
globally static and the fluid passes through the interstitial spaces, according
to the well-studied flow model of the packed beds 3.1.a
Figure 3.1: Various forms of contacting [38]
4particle size, solid density and particle size distribution
13
Introduction 14
As the velocity of the fluidizing fluid is increased there comes a point where
the bed is lifted and a dynamical equilibrium condition is established. The
drag force exerted by the fluid on the particles is equal to the buoyant weight
of the bed. This value of the superficial fluid velocity is said minimum flu-
idization velocity umf 3.1.b.
With fine particles5, fluidized by liquid, an increase of flow rate beyond
umf will produce the uniform bed expansion. In this new regime the void
fraction and height of the bed are increased, this behaviour is called homoge-
neous fluidization or smooth fluidization Figure 3.1.c. In gas–solid systems
this kind of fluidization is observable only in special conditions, as high pres-
sure, and with solid particles very fine and light.
If instead the fluid is a gas and the particles have a greater mean diameter6
all gas in excess that needed to just fluidize the bed, is subdivided in a series of
bubbles. These are generated at the bed bottom and rise up through the solid
phase, reaching the bed surface. Coalescence and splitting phenomena govern
the complex dynamic of such a system, this regime is known as bubbling or
aggregative fluidization Figure 3.1.d. The superficial fluid velocity to which
the first bubble appears is called minimum bubbling velocity umb.
The bubbles can grow up to assume dimensions comparable with vessel
diameter, especially in narrow columns, this regime is indicated like slugging
Figure 3.1.e, and the superficial gas velocity at which this happens is called
minimum slugging velocity ums. According to particles size, two kind of slugs
can be distinguished, respectively axial slugs 3.1.e and fat slugs 3.1.f.
In the first one the solid phase tends to move close to vessel walls, turning
5Group A of Geldart classification [22]
6Group B of Geldart classification [22]
14
Introduction 15
around the gas bubbles, in the last one, instead, slugs grow so much that exert
on the overlooking suspension a piston like pressure, pushing it upward, and
inducing an oscillatory motion, periodically repeated because of continuous
formation, rising up and eruption of slugs, to which corresponds the lifting
and the relapse of the solid particles on the vessel bottom.
From the bubbling regime a further increase of the superficial flow rate
causes a return of great homogeneity conditions, characterized by the forma-
tion of particles clusters, which replace the bubbles and improve the degree
of mixing, this kind of state is know as turbulent regime.
Finally, for ulterior superficial gas velocity increase, the gravitational at-
traction does not succeed to keep back the solid particles.These subjected to
fluid–solid interactions, are carried away by pneumatic conveying, determin-
ing, without special equipment, the bed disintegration.
3.1.1 Powder Classification
Fluidization regime prediction for various type of solid particles with air
at ambient conditions is easily possible by empirical powder classification
proposed by Geldart [22] in 1973, Figure 3.2:
• Group A, Aeratable particles of mean diameter between 30 < dp < 150µm
and density of ρs < 1500 Kg/m3 undergo an homogeneous fluidization
process, with uniform bed expansion, and are characterized by different
values for umf and umb. Cracking catalyst belong to this group.
• Group B, Bubbling particles of mean diameter between 40 < dp < 500µm
and density between 1500 < ρs < 4000 Kg/m3, which show aggrega-
tive fluidization with umf ≈ umb, coalescence is the predominant phe-
15
Introduction 16
Figure 3.2: Diagram for classifying powders into groups [23]
nomenon and a max stable bubble size in never achieved. Example of
this category is coarse sand.
• Group C, Cohesive particles very fine and light such as talk, flour,
cement, difficult to fluidize. Their behaviour is strongly influenced by
interparticle forces, an attempt to fludize this kind of powder produces
channels or a discrete plug.
• Group D, Different particle very large and massive, like lead shot which,
with an upward fluid flow exhibit a spouting fluidization process.
For an overall resume of fluidization conditions is useful to note the scheme
Figure 3.3.
16
Introduction 17
Figure 3.3: Gas-solid flow classification [18]
3.1.2 Liquid Like Behaviour
The fluidized bed shows liquid-like properties such as Figure 3.4
Figure 3.4: Liquid-like behaviour [38]
1. flowing easily following hydrostatic pressure law
2. maintaining an horizontal level when tilted
17
Introduction 18
3. allowing for low-density objects to float on the bed surface
It is noticeable to cite on that point an experiment designed for the first time
in the Chemical Engineering Department of UCL7 in which starting from two
duck–shape toys one of plastic collocated on the bed bottom and the other
of brass posed on the top, and blowing an air flow from down as soon as it
is reached the minimum fluidization velocity umf the brass duck sinks to the
bottom and the plastic one goes up and floats on the bed surface in the same
way shown on the water.
3.1.3 Advantages
Fluidized state offers several advantages on engineering application field:
1. ease of operations
2. rapid mixing of solids, which leads to isothermal conditions throughout
the bed
3. high heat and mass transfer rates between gas and particles
3.1.4 Industrial Applications
To give some insight into the workings of contacting scheme, consider a
catalytic cracking reactor as shown in Figure 3.5: The high temperatures
and carbon like materials developing, deactivate periodically catalytic parti-
cles, so necessarily regeneration happens in a separate reactor, where these
7University College London, London UK
18
Introduction 19
Figure 3.5: Operating principle for stable circulation [38]
19