Epstein N., Grace J.R. Spouted and Spout-Fluid Beds: Fundamentals and Applications

Подождите немного. Документ загружается.

Hydrodynamics of spout-ﬂuid beds 127

72. K. B. Park, J. L. Plawsky, H. Littman, and J. D. Paccione. Mortar properties obtained by dry

premixing of cementitious materials and sand in a spout-ﬂuid bed mixer. Cem. Concr. Res.,

36 (2006), 728–734.

73. S. Rakic, D. Povrenovic, V. Tesevic, M. Simic, and R. Maletic. Oak acorn, polyphenols and

antioxidant activity in functional food. J. Food Eng., 74 (2006), 416–423.

7 Spouted and spout-ﬂuid beds

with draft tubes

Zeljko B. Grbav

c, Ho ward Littman, Morris H. Morgan III, and

John D. Paccione

The draft tube spout-ﬂuid bed (DTSFB) is an extremely versatile ﬂuid–particle system

that can be widely applied industrially. This conﬁguration aids performance because

control of the ﬂuid residence time and solid cycle time distributions can be accomplished

easily. Thus understanding the solid and ﬂuid phase hydrodynamics is essential to

optimizing DTSFB operation.

A typical DTSFB system consists of four interconnected zones, as shown in Figure 7.1:

a spout-ﬂuid bed feeder, a moving bed annulus, a draft tube (pneumatic conveyor), and a

freeboard fountain. For a given geometry and particles, ﬁve control quantities determine

the bed hydrodynamics: the inlet jet ﬂuid mass ﬂowrate, F

; the inlet auxiliary ﬂuid mass

ﬂowrate, F

ax0

; the internal annulus ﬂuid mass ﬂowrate, F

; the inlet section length, L

;

and the annulus height, H

. The draft tube inlet length,L

, controls both the ﬂuid leakage

and solids crossﬂow. The auxiliary ﬂuid mass ﬂow, F

ax0

, controls the pressure drop across

the bed and alters the internal solids ﬂow. To describe the overall and local dynamics of

such systems, appropriate ﬂuid and particle models must be devised for each separate

region of the DTSFB. The discussion in this chapter addresses operating characteristics,

applications, and design concepts. Initially we focus on the basic hydrodynamic features

of these systems, and then we provide an overview of applications involving the basic

DTSFB and novel hybrid conﬁgurations.

7.1 Operation of draft tube spout-ﬂuid beds

Grbav

cetal.

provided one of the ﬁrst comprehensive examinations of the basic

properties – minimum ﬂuid ﬂow rate, pressure drop and solids circulation rate – of a

DTSFB. Figure 7.2 summarizes some of their initial ﬁndings. One might expect that

all the quantities plotted in that ﬁgure should approach their spouted bed values as

→ 0 and L

/H → 1. For a given bed, however, two lower limits of L

/H are encoun-

tered: (1) a minimum inlet section length below which no particle circulation occurs,

and (2) a maximum value, less than bed height H, at which vertical pneumatic transport

cannot be achieved in the draft tube because air short-circuits through the annulus.

Grbav

cetal.

also examined the ﬂuid ﬂow pattern and solids circulation rate of

liquid–solid spout-ﬂuid beds equipped with draft tubes. They employed a 196-mm

Spouted and Spout-Fluid Beds: Fundamentals and Applications, eds. Norman Epstein and John R. Grace.

Published by Cambridge University Press.



Cambridge University Press 2011.

Spouted and spout-ﬂuid beds with draft tubes 129

ax0

Figure 7.1. Schematic of draft tube spout-ﬂuid bed.

1.0

0.8

0.6

0.4

0.2

0 0.2 0.4 0.6

D = 110 mm, D

= 25 mm

= 10.8 mm, d

=1.84 mm

SPOUTED BED

0.8 1.0

/DP

/(G

)

Figure 7.2. Variation of basic spouting parameters with inlet s ection length (H = 0.23 m, U

0.95 m/s, P

= 3.47 kPa, (G

)

= 562 kg/h). Adapted from Grbav

cetal.

diameter semicircular column equipped with a semicircular draft tube of diameter

34.5 mm. Spherical glass particles of 1.20, 1.94, and 2.98 mm diameter were spouted

with water. For a ﬁxed annulus inlet volumetric ﬂowrate, V

, the transition to stable

particle circulation exhibited several key features with increasing spouting ﬂowrate,

. At low spouting ﬂowrates, a cavity formed just above the spout inlet nozzle, and

there were no particles in the draft tube. Further increases of V

caused particles to

be entrained from the annulus, and at a certain value of V

, particulate ﬂuidization of

the solids occurred in the draft tube. With an additional increase in V

, the ﬂuidized

medium in the draft tube expanded, b ut the voidage there remained nearly constant

130

Zeljko B. Grbav

c, Howard Littman, Morris H. Morgan III, and John D. Paccione

1.2

1.0

0.8

0.6

0.4

0.2

0.0

0.0 0.2 0.4 0.6

)

min

/ V

(mm)

/ V

0.8 1.0

100

SPOUT-FLUID BED

WITHOUT DRAFT

TUBE

Figure 7.3. Minimum spouting ﬂowrates required to initiate draft tube particle circulation (d

1.94 mm, V

= 5.17 cm

/s, H = 585 mm). Adapted from Grbav

cetal.

because particles were continually entrained from the annulus into the draft tube. At a

critical value of V

, called the minimum turnover ﬂowrate, the expanded ﬂuidized bed

reached the top of the draft tube, and particles began to leave the top of the draft tube.

The draft tube voidage at this point, estimated from the pressure gradient there, slightly

exceeded the minimum ﬂuidization voidage. The minimum ﬂow rate necessary to pro-

duce particle ﬂow through the draft tube increased as L

increased, as shown in Figure 7.3.

for minimum turnover decreased with increasing V

and the Littman et al.

spout-

ﬂuid line, shown by a dashed line, provides an upper bound for this quantity.

Figure 7.3 shows that the solids mass ﬂux in the draft tube increases rapidly with L

for

ﬁxed V

and V

. The solid circulation in a DTSFB can be established without spouting

if V

is sufﬁciently large, as shown in Figure 7.4. In addition, if the annulus height is

slightly below the top of the draft tube, the annulus region cannot be ﬂuidized. When

increasing V

= 0), it was found that close to the minimum ﬂuidization velocity of

the annulus, the bed expanded. When V

slightly exceeded V

, hydraulic transport in

the draft tube began, and the annulus remained un-ﬂuidized. With a further increase in

, the solids ﬂow rate in the draft tube also increased, but without ﬂuidization of the

annulus solids. In addition to the spouting and annular ﬂow rates, the most important

factors affecting the ﬂuid ﬂow distribution between the annulus and draft tube sections

include the entrance region (including the inlet nozzle), the length of the inlet section,

cone angle, draft tube diameter, and particle size.

Had

zismajlovi

cetal.

and Povrenovi

investigated the hydrodynamic behavior of a

0.95-m-diameter spout-ﬂuid bed equipped with a 0.25-m-diameter draft tube. Polyethy-

lene particles 3.6 mm in diameter were spouted by water. Experiments were conducted

over a wide range of inlet nozzle sizes, annulus air ﬂowrates, and inlet section lengths.

Their results extended the Grabavci

cetal.

observations. A value of 66.7 was obtained

Spouted and spout-ﬂuid beds with draft tubes 131

600

(kg/m

500

400

300

200

100

0.0 0.5 1.0

Solid circulation

without nozzle flow

0.211

0.421

0.632

0.842

1.053

1.248

1.5 2.0

Figure 7.4. Effect of the spouting ﬂow rate on solids mass ﬂux in a draft tube system (d

= 1.94

mm, V

= 5.17 cm

/s, H = 585 mm). Adapted from Grbav

cetal.

for the critical D

ratio for stability, substantially higher than the value of 25 reported

by Chandnani and Epstein

for stable spouting of ﬁne particles in a gas-spouted bed

without a draft tube.

Had

zismajlovi

cetal.

proposed a movable nozzle (Figure 7.5). To initiate particle

ﬂow in the draft t ube, the nozzle tube was retracted to the desired inlet section length.

At the completion of a run, the inlet nozzle tube was reinserted into the draft tube to

prevent particle backﬂow from the annulus.

Cecen-Erbil

used pulse injection of a tracer to deﬁne ﬂuid streamlines and to deter-

mine the liquid ﬂow distribution between the draft tube and annulus in a semicircular

ﬂat-bottomed spouted bed, 80 mm in diameter, with an inlet tube of diameter 10 mm.

A semicircular draft tube of 20 mm internal diameter and 800 mm length was located

32 mm above the distributor. The particles were glass spheres of diameter 3 mm, and

the spouting ﬂuid was water. The superﬁcial liquid velocity through the draft tube was

found to vary linearly with the total ﬂow r ate through the bed. Under certain conditions,

it was found that liquid could leak from the annulus inlet ﬂow into the draft tube. An

empirical correlation was proposed to predict this leakage fraction.

Ijichi et al.

8–10

investigated the effects of gas ﬂowrate, inlet section length, conical base

angle, particle diameter, and bed weight on the solids circulation rate and annular gas

velocity. Both semicircular and circular draft tube columns were studied. The annular

gas velocity was found to increase with increasing inlet section length (L

), angle of

conical base, and mean particle diameter, but decreased with increasing bed mass. The

particle circulation rate increased with increasing tube cone clearance, mean particle

diameter, and bed mass.

132

Zeljko B. Grbav

c, Howard Littman, Morris H. Morgan III, and John D. Paccione

MOVABLE

NOZZLE

Figure 7.5. Draft tube spout-ﬂuid bed with movable nozzle.

A similar study by Zhao et al.

investigated the vertical particle velocity proﬁles

and solids circulation rate in a cylindrical-conical spouted bed, with and without a draft

tube. They devised a kinematic model to simulate the granular ﬂow in the annulus.

Experiments were conducted in a column of 0.196 m diameter and 1.2 m height, with a

◦

conical base, 0.0257 m inlet diameter, 0.42 m static bed depth of millet particles of

1.34 mm mean diameter and 1430 kg/m

density, at ambient temperature and pressure.

Draft tubes of diameter 30 and 40 mm were tested. The particle velocity in the annulus

of such beds (DTSFB) increased slightly with increasing radial position, contrary to the

trend for conventional spouted beds (CSBs). Average particle velocities in the DTSFB

were much lower than in the CSB, but essentially equivalent to those of the draft-tube

spouted bed (DTSB).

Nagashima et al.

compared results for binary mixtures of glass beads in a draft

tube system with those for a bed of similar size containing coarser particles. Both the

minimum spouting velocity and gas velocity in the annulus decreased with increasing

mass fraction of ﬁner particles, and decreased with decreasing L

A three-phase investigation by Cecen-Erbil and Turan

focused on developing a

model for the energy dissipation inside the draft tube of spout-ﬂuid beds. Their goal was

to calculate the shear stress, velocity gradient, and turbulence ﬂuctuation parameters

inside the draft tube to assess the performance of a novel backwashing ﬁlter process.

Gas spouting inlet ﬂow and liquid annulus ﬂow were introduced through two indepen-

dent lines. Air was also introduced with the liquid into the annulus through the same

Spouted and spout-ﬂuid beds with draft tubes 133

distributor. The proposed model suggested that a desired shear stress/velocity gradient

could be achieved by judicious selection/combination of draft tube size, particle size,

and ﬂow rates.

7.2 Novel applications and experimental studies

7.2.1 Cylindrical and rectangular designs

Zhao et al.

studied the particle motion in a two-dimensional thin slot-rectangular

spouted bed (2DSB) with draft plates using par ticle image velocimetry (PIV). A dis-

crete element method (DEM) (see Chapter 4), in conjunction with computational ﬂuid

dynamics (DEM-CFD), was used to examine gas turbulence effects. The experiments

were conducted in a rectangular column, 152 mm wide and 15 mm thick, equipped with

a60

◦

-angle conical-base section. The two draft plates were 15 mm apart. The static bed

depth was 100 mm, and the inlet length ranged from 10 to 40 mm. The particles were

glass spheres of diameter 2 mm. The PIV results indicated that the inlet section length

had little inﬂuence on the vertical particle velocity, but a greater effect on the particle

circulation rate. DEM-CFD simulations did a good job of predicting the longitudinal

particle velocity proﬁle along the center line and showed that particles travel upward

individually, without clustering.

CFD simulations for grains of 0.22, 2.0, 3.7, and 1.0 mm diameter, by Szafran and

Kmiec,

conﬁrmed that ﬂuctuations are caused by particle clusters originating at the

bottom of the column. Solids were observed to cross into the jet, cover the column

inlet, and be transported periodically through the draft tube, contrary to the ﬁnding

of Zhao et al.

The ﬂuctuating solids inﬂow produces slugs and explains variations in

fountain height and porosity. Modiﬁed and extended scaling relationships were proposed

by Shirvanian and Calo

for conical-based rectangular spouted vessels with draft tubes.

A CFD model was devised to investigate the hydrodynamics and to predict the solid

volume fraction, solid and liquid velocities in the draft tube, and ﬂuid volume-average

fraction in the draft tube as a function of height. The CFD model predicted the solids

circulation rate to within 17 percent.

Another multizone spouting model was developed by Eng et al.

for nonisothermal

dynamic simulation of a DTSB for ultrapyrolysis of hydrocarbons. This model focused

on the nonisothermal dynamic response of the spout gas to disturbances and investigated

the effect of hydrodynamic behavior on unsteady heat transfer. The simulations predicted

that the spout gas responded in two different time scales: an initial pseudo-steady state

was established within 30 ms, followed by a long-time response after 15 min.

Takeuchi et al.

applied a DEM to a cylindrical system with a conical base. The

authors devised a new method for treating the boundary condition for the three-

dimensional gas ﬂow along the conical surface, based on a technique developed origi-

nally for particle-induced turbulent ﬂows. The simulated particle circulation and velocity

proﬁles in the spout and annulus agreed well with experimental observations. In the cylin-

drical region, the axial particle velocity proﬁles were self-similar, of Gaussian shape.

A small reverse gas ﬂow was observed near the bottom of the conical base, decreasing

134

Zeljko B. Grbav

c, Howard Littman, Morris H. Morgan III, and John D. Paccione

(a) (b) (c) (d) (e) (f) (g)

Figure 7.6. Draft tube spouted/spout-ﬂuid bed designs: (a) porous draft tube spouted bed, adapted

fromClaﬂinandFane

and Ishikura et al.

; (b) ﬂat s creen gas inlet; (c) conical screen gas inlet;

(d) nozzle gas inlet; (e) oriﬁce gas inlet. Subﬁgures b, c, d, and e are adapted from Hattori

et al.

27– 29

; f and g are adapted from Hattori et al.

30, 31

with increasing cone angle. Kmiec and Ludwig

19,20

proposed a simple one-dimensional

model for the axial proﬁles of particle velocity, gas velocity, and voidage along the draft

tube of such systems.

7.2.2 Conical and other design modiﬁcations

Altzibar et al.

pointed out that conical spouted beds have low particle segregation,

making them excellent for treating particles of wide size distribution. They investigated

experimentally the hydrodynamics of draft-tube conical-spouted beds. Correlations were

proposed for the minimum spouting velocity, pressure drop, and peak pressure drop as

functions of bed and column characteristics. Altzibar et al. also found that the minimum

spouting velocity decreased with a reduction in ﬂuid inlet diameter and that the spouting

pressure drop increased with increasing inlet section length. A very pronounced hystere-

sis in the pressure drop versus air velocity occurs in conical DTSB units, much larger

than in conventional conical spouted beds.

San Jos

eetal.

studied the effect of draft tube diameter and inlet length on stability

of a conical contactor with the top of the central draft tube at the same level as the upper

bed surface. Bed stability was enhanced when D

≥ 1, providing that 5 ≤ D

≤

50 and L

> 10. They provided a correlation for the minimum spouting velocity for

conical spouted beds with and without draft tubes.

Several variants of porous draft tube systems (see Figure 7.6a) were studied by Claﬂin

and Fane

and Ishikura et al.,

with an emphasis on the effect of design changes on

annular ﬂuid ﬂow. In the Ishikura et al.

24–26

investigation, the porous draft tubes were

0.30 m in length and varied from 0.012 to 0.018 m in diameter. The inlet section length,

, was restricted to between 0.02 and 0.04 m. Their porous-wall draft tube was made

of sintered stainless steel containing pores 120 μm in diameter, with a void fraction of

∼40 percent. Binary mixtures of glass beads (d

= 1.35 mm and 0.477 mm) were the

spouting solids. In most experiments a small fraction of ﬁne particles was added to

Spouted and spout-ﬂuid beds with draft tubes 135

0 0.2 0.4 0.6 0.8 1 1.2

= 1.351 / 0.477 mm

Mass fraction of fines = 0.05

H = 0.3 m

S-DTSB

P-DTSB

U (m/s)

(kPa)

Figure 7.7. Typical pressure drop–superﬁcial gas velocity curves for a binary mixture spouted in

systems without and with a draft tube (SB; conventional spouted bed, S-DTSB; spouted bed with

nonporous draft tube, P-DTSB; spouted bed with porous draft tube). Adapted from Ishikura

et al.

the coarser particles. As anticipated, the porous draft tube led to a higher annular gas

through-ﬂow than a nonporous one. Figure 7.7 shows the dependence of the spouting

pressure drop on the gas superﬁcial velocity for a conventional spouted bed (SB), a

spouted bed with a nonporous draft tube (S-DTSB), and a spouted bed with a porous

draft tube (P-DTSB).

For both the porous and the nonporous draft tube systems, P

and solids circulation

increased with increasing gas superﬁcial velocity. In addition, P

and U

for the

spouted bed with nonporous draft tube was always lower than P

and U

for the

spouted bed with a porous draft tube of the same size. The addition of even a small

fraction of ﬁne particles was found to reduce the minimum spouting gas velocity.

Hattori et al.

27–29

investigated four modiﬁcations of the conventional draft tube

spouted bed design, as shown in Figure 7.6: (b) a ﬂat screen gas inlet, (c) a conical

screen gas inlet, (d) a nozzle gas inlet, and (e) an oriﬁce gas inlet. Conﬁgurations (b) and

(0.55 and 0.61 m/s, respectively) than conﬁgurations (d) and (e) (0.35 and 0.36 m/s,

respectively), for similar particles and bed dimensions. The solids circulation rate of the

screen-bottomed spouted beds was considerably larger than the rate for the nozzle or

oriﬁce designs (Figure 7.8a). In addition, the gas velocity in the annulus of screen bottom

units was almost three times higher than that for the nozzle and oriﬁce conﬁgurations

(Figure 7.8b).

Hattori et al.

30–32

also developed a side-outlet draft tube spouted bed (Figure 7.6f).

Particles as small as 0.27 mm could be effectively spouted in air without instabilities

136

Zeljko B. Grbav

c, Howard Littman, Morris H. Morgan III, and John D. Paccione

0.5

1 1.5

Nozzle or orifice

Screen bottom

flat screen

conical screen

Screen bottom

= 1.06 m/s

flat screen

conical screen

nozzle

orifice

Nozzle or orifice

nozzle

orifice

2 2.5

U/U

1 1.5

(b)

(a)

2 2.5

V [m

/s]

/ U

[–]

[x10

–5

]

Figure 7.8. Solids circulation rate (a) and gas velocity in the annulus (b) versus superﬁcial gas

velocity in a modiﬁed draft tube spouted bed. Adapted from Hattori et al.

in the draft tube. It was later shown

that, for any given geometrical conﬁguration

of column and draft tube, there is an optimum vertical location of the side outlet that

produces a maximum conversion for a catalytic reaction. To avoid screen plugging by

ﬁne particles, Hattori et al. also proposed the conﬁguration shown in Figure 7.6(g).

Kim

reported that proper selection of the ratio of upper column diameter to lower

column diameter was crucial for stability. Ijichi et al.

demonstrated that very ﬁne

particles (Group C in the Geldart

classiﬁcation), could be processed effectively in a

draft tube spouted bed if it was equipped with a converging nozzle inlet.

Ar and Uysal

studied the potential application of liquid spouting with multiple

draft tubes for nonfouling heat exchangers. Liquid was introduced through the center

multitubular zone. A conical barrier just below the return pipe diverted the liquid ﬂow

toward the draft tube entrances around a return pipe. Both entrained solid particles

coming from the return pipe and liquid rose in the draft tubes. Predictive models for the

solids circulation rates proposed by Grbav

cetal.

and Nitta and Morgan

met with

limited success.

Follansbee et al.

proposed a continuous photocatalytic reactor design for removal

of organic compounds from wastewater. Their reactor (Figure 7.9) contains an annular

moving packed bed of TiO

-activated carbon catalyst supported on large silica particles

=1.94 mm). The contaminated water stream is in continuous countercurrent contact

with a photocatalytic adsorbent catalyst. Treated water leaves the column at the top of

the annulus, whereas spent adsorbent is transpor ted through the draft tube into the UV

exposure chamber. Regenerated particles fall onto the annulus top for another adsorp-

tion cycle. The UV chamber serves two functions: (1) activation of the photocatalytic

component to promote surface oxidation and (2) regeneration of the sorbent. The authors

achieved 90 percent removal of the organic contaminant from the wastewater.

Arsenijevic et al.

developed a combined system for ethylene oxide (EtO) removal

based on a draft tube spouted bed (DTSB) loaded with Pt/Al

catalyst. The annular

region was partitioned into a “hot” zone (desorber and catalytic incinerator) containing