Professional Documents
Culture Documents
Outline/Contents
Introduction.
Fluidization Flow Regimes.
Overall Gas (Voidage) and solids Hold-up.
Radial and Axial Solids Hold-Up Profiles.
Radial and Axial voidage distribution.
Gas and Solid Mixing.
Scale-Up.
Reactor Modeling.
INTRODUCTION
Gas distributor
Advantages
It has the ability to
process large volumes
of fluid.
Excellent gas-solid
contacting.
Heat and mass transfer
rates between gas and
particles are high when
compared with other
modes of contacting.
No hot spot even with
highly
exothermal
reaction.
Ease of solids handling.
Disadvantages
Broad residence time
distribution of the gas due
to dispersion and bypass
in the form of bubbles.
Broad residence time
distribution of solids due
to intense solids mixing.
Erosion of internals.
Attrition
of
catalyst
particles.
Difficult Scale-up due to
complex hydrodynamics.
kg^3
Group C (Cohesive)
:- very fine
particle, particle size < 30 m, difficult
to fluidize because inter-particle
forces are relatively large, compared
to those resulting from the action of
gas.
Group D (Spoutable) :- (e.g., Roasting
coffee beans) large particle, stable
spouted beds can be easily formed in
this group of powders.
2
f u mf
p f g
3
s D p mf
1501 mf
s D p u mf f
1.75
Experimentally, the most common method of measurement requires that pressure drop
across the bed be recorded as the superficial velocity is increased stepwise through Umf
and beyond, Umf is then taken at the intersection of the straight lines corresponding to
the fixed bed and fluidized bed portions of the graph obtained when Pbed is plotted
against U on log-log coordinates.
Bubbling Fluidization
This type of fluidization has
been called aggregative
fluidization, and under
these conditions, the bed
appears to be divided into
two phases, the bubble
phase and the emulsion
phase.
Turbulent Fluidization
Turbulent regime has the following features: High solid hold-ups (typically 25-35 %
by volume).
Limited axial mixing of gas.
Suitable for exothermic and fast
reactions.
Good gas-solid contact and hence,
favors reactant conversion.
high gas flow-rates operation and good
for isothermal operation.
Favorable bed to surface heat transfer.
Canada et al. 1978
Particle classification
FCC regenerators
Group A
0.5-1.5
Acrylonitrile
Group A
~0.5
Maleic anhydride
Group A
~0.5
Phthalic anhydride
Group A
~0.5
Ethylene dichloride
Group A
~0.5
Group A
~1.5
Bi et al. 2000
Grace (1986a) summarized the effects of particles properties and operating conditions
on fluidization behavior and prepared a flow regime diagram. The flow regime diagram
was further modified by Kunii and Levenspiel (1997).
For given particles and operating velocity, the gas-solid contact pattern can be
determined using this diagram. Likewise, for a given flow regime, this diagram could
provide available combinations of particle properties and gas velocity.
Yang 2003
Fluidization diagram
Us
U Gs
av
Solid hold-up
Yerushalmi and Cankurt, 1970
Pressure
Temperature
The transition velocity was almost independent of the static bed height, which
varied from 0.4 to 1.0 m (Grace and Sun 1990). Similar results were reported
by Cai (1989) and Satija and Fan (1985) with (Hmf/Dt) > 2. On the other hand,
for a shallow fluidized bed of (Hmf/Dt) < 2 with Group B and D particles,
Canada et al. (1978) and Dunham et al. (1993) found that Uc increased with
static bed height. This could be related to the undeveloped bubble flow in
shallow beds before transition to turbulent fluidization can occur (Bi et al.
2000).
Uc increases with increasing mean particle size and density (Cai et al. 1989, Bi
et al. 2000).
Column Diameter
Internals
Cai (1989)
Particle size
Particle
density
Column
diameter
Temperature
Matsen 1996
Effect of pressure
Higher operating pressures reduced the bed expansion (H/Hmf)
(Miller et al., 1981) .
The increase of bed expansion with pressure (Chiba et al., 1986, and
Chitester et al., 1984) .
The physical properties of the fluidizing gas, density and viscosity did
not have any significant effect on bed expansion (Denloye, 1982),
and Knowlton,1977).
Bed expansion increased significantly with pressure but this
influence, very strong at low pressures, seemed to reach a maximum
at approximately 800kPa and decreased thereafter up to 1200kPa
(Llop et al., 1995; 2000, and Olowson and Almstedt, 1990) .
Radial Profile
Although, overall gas holdup
has been traditionally used
for
characterization
of
hydrodynamics of fluidized
bed columns, it is a single
lumped parameter. Hence,
for detailed characterization,
one need to look at the way
solid is distributed across the
reactor.
U=0.53 m/s,
sand particle (250 microns, 2.5 g/cm^3)
Bubbling regime, Fiber optical needle probe
Axial Profile
The axial solid hold-up obtained by fiber optical needle probe and CARPT
shows a quasi linear profile (Mabrouk et al. 2005).
Du et al. 2002
Du et al. 2002
Gas Mixing
(a) Axial Gas Mixing
Model
Tracer
injection
dp (m)
D(m)
U (m/s)
Uc (m/s)
Dzg (m2/s)
Steady state
362
0.1
0.8
0.88
1.00
1.08
1.20
0.85
0.22
0.235
0.230
0.245
0.215
One dimensional
dispersion
Steady state
59
0.152
0.1
0.5
1.3
0.43
0.1
0.55
0.60
Li and Wu (1991)
(Air-H2)
1D pseudohomogeneous
diffusion
Non-ideal pulse
58
0.09
1.0
1.0
1.0
0.44
0.45
0.51
0.56
One dimensional
dispersion
Pulse
75
0.1
0.417
0.516
0.614
0.691
0.792
0.892
0.977
1.051
1.142
0.47
0.080
0.102
0.11
0.195
0.130
0.167
0.097
0.060
0.075
Two-phase model of
van
Deemter (1980)
75
0.1
0.21
0.4
0.5
0.6
0.7
0.8
0.94
0.55
0.09
0.16
0.19
0.175
0.14
0.13
0.14
Pseudo-homogeneous
model with axial and
radial dispersion
Steady state
77.6
0.19
0.392
0.588
0.784
1.078
0.5
0.374
0.514
0.619
0.783
One dimensional
dispersion
Steady state
58
5.76
1.26
1.41
0.41
3.05
3.4
Unsteady state
60
0.203
0.21-1.5
0.5
Plotted in Fig.
Du et al. 2002
Solids diffusivities
Scale-up criteria
Glicksman et al, 1993, 1998
Properties of the Silica Sand Bed Materials Used in the Similarity Experiments
Low velocity
High velocity
Selected References
1. Sanderson, John, and Rhodes, Martin, Bubbling Fluidized Bed Scaling Laws:
Evaluation at Large Scales, AIChE Journal, 2005;51 (10): 2686-2694.
2. Glicksman LR, Hyre M, Woloshun K. Simplified scaling relationships for
fluidized beds. Powder Technol. 1993;77:177-199.
3. Horio M, Nonaka A, Sawa Y, Muchi I. A new similarity rule for fluidized bed
scale-up. AIChE J. 1986;32:1466-1482.
4. Glicksman LR. Scaling relationships for fluidized beds. Chem Eng Sci.
1988;43:1419-1421.
5. van den Bleek CM, Schouten JC. Deterministic chaos: A new tool in fluidized
bed design and operation. Chem Eng J. 1993;53:75-87.
6. Schouten JC, van der Stappen MLM, van den Bleek CM. Scale-up of chaotic
fluidized bed hydrodynamics. Chem Eng Sci. 1996;51:1991- 2000.
7. Glicksman LR, Hyre MR, Farrell PA. Dynamic similarity in fluidization. Int J
Multiphase Flow Suppl. 1994;20:331-386.
8. Glicksman LR. Fluidized bed scale-up. In: Yang W-C, ed. Fluidization Solids
Handling and ProcessingIndustrial Applications. Park Ridge, NJ: Noyes;
1999.
Reactor Modeling
The model is initially developed in rectangular coordinates for simplicity, but can be
transformed to any other coordinate system (e.g. cylindrical curvilinear) using elementary
vector calculus theory of vector operators (Mahecha and Grace et al. 2006).
This model includes most existing fluid bed reactor models as special cases, allowing
clear connections to be established among the models and showing the significance and
implications of each simplifying assumption. This will lead to a more systematic approach
to fluidized-bed reactor modeling, facilitating what has been called the optimum degree of
sophistication (Aris, 1961).
Once the more general model has been developed and debugged, we will be in a position
to apply it to important and potentially viable industrial processes such as partial oxidation
reactions and hydrogen production processes (Mahecha and Grace et al. 2006).
Generalization of Models
The dynamic equations take into consideration in a rigorous manner the heat and mass
capacities of the gases and solids in each pseudo-phase (Elnashaie, Elshishini, 1993).
2)
3)
4)
The model is not restricted to a single flow regime. Its hydrodynamic parameters can be
calculated as proposed by (Abba, Grace et al., 2003) for several adjacent flow regimes.
5)
Both mass and heat dispersion are included along all coordinate axes (Bird, Stewart et al., 2002).
6)
The model deals with anisotropic mass diffusion and heat conduction.
7)
The model takes into consideration three-dimensional convective velocities (Bird, Stewart et al.,
2002).
8)
The convective velocities can be calculated using any function (e.g. accounting for changes in the
number of moles and gas volumetric flow (Abba, Grace et al., 2002)). Changes with time,
temperature, pressure and chemical reaction are also covered.
The model accounts for catalyst chemisorption (Elnashaie, Elshishini, 1993) and
solid capture of any species.
10)
11)
The model accounts for deactivation of catalyst (Chen, Yan et al., 2004).
12)
The model considers the use of membranes to remove certain products (i.e. to break
the thermodynamic barrier) or to supply certain reactants (i.e. to improve the system
selectivity to a desired product). Membrane deactivation fuctions can also be
included (Raich & Foley, 1995).
13)
The catalyst effectiveness factor may differ from 1 (Elnashaie, Elshishini, 1993).
14)
In the energy balance, different expressions for calculating the internal energy
(Smith, Van Ness et al., 1996) can be used including, where appropriate, sensible
and latent heats (in case of change of phase).
15)
The reactor cross-sectional area can vary along the height of the reactor. The model
does not need to be modified when using different geometries.
Pseudo-phase approach
Control volumes for the conservation balances include both gas and solid
phases, without ignoring the effect of the solids on the system dynamics (Gas
carried inside the solids and the heat and mass capacitances of the solids are
included in the mole and energy balances).
Mole Balance
The molar rate balance over a differential element for phase (p) is given by:
Energy Balance
The differential energy balance for phase (p) is given by:
Energy dissipation due to viscous effects is neglected. The number of energy balance
equations is N(P) where N(P) is the number of pseudo-phases. The generalized energy
balance for phase (p) is as follows:-
Pressure Balance
A simplified differential pressure balance in the z direction for phase (p) is given by:
The density of phase (p) can be calculated using the void fraction as:
The differential control volume of pseudo-phase (p) has no external exchange with the surroundings.
The interaction of the pseudo-phase with its surroundings should thus be included in the boundary
conditions.
The boundary conditions should be specified according to the geometric arrangement of the system,
and may vary from case to case.
The boundary conditions (i.e. for the simplest single-phase case) may assume axial symmetry, zero
flux at the walls and Danckwerts criteria when the diffusion in the fore and aft sections is negligible
(Danckwerts, 1953). A base set of boundary conditions is displayed in Table 1. Other details of the
model can be found elsewhere (Mahecha-Botero, Grace et al., 2005).
Reactor parameters
Results
Predicted steady-state ETY molar flows in the highand low-density pseudo-phases vs height in the reactor.
Predicted steady-state HCl molar flows in the highand low-density pseudo-phases vs height.
Results (Contd)
(Mahecha and Grace et al. 2006).
Remarks
(Mahecha and Grace et al. 2006).
The model is able to describe fluidized bed reactor systems relying on fewer
assumptions than other models in the literature. When different combinations
of assumptions are incorporated in the model, it simplifies to a number of fluid
bed reactor models previously presented in the literature.