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

Подождите немного. Документ загружается.

18 Mechanical spouting

Tibor Szentmarjay, Elizabeth Pallai, and Judith T

oth

18.1 Principle of mechanical spouting

One of the main advantages of the conventional spouted bed is the characteristic recircu-

lation particle mixing provided by a high-velocity gas jet through a nozzle at the conical

bottom of the equipment. This ensures very effective “quasi-countercurrent” gas–solid

contact. Because of the circulation of the particles, local overheating within the bed can

be avoided and thus uniform product quality can be effected.

Disadvantages include ﬂuid motive-power consumption that occurs especially at the

start of operation, but also persists under stable conditions owing to the vertical particle

transport in the spout and the high pressure drop of the nozzle. In addition, conditions

for heat and mass transfer can be obtained only within a limited range of gas ﬂowrate

because the velocity of the gas, usually air, is determined mainly by the size and density

of the particles.

These disadvantages of the spouted bed can be eliminated, while keeping its advan-

tages, by t he mechanically spouted bed (MSB), in which the most important characteris-

tics of conventional spouted beds can be found.

According to the developed prototype,

the typical spouted circulating motion is ensured by an open conveyor screw, installed

along the vertical axis of the device, independent of the airﬂow rate. The diameter of the

screw is nearly equal to the diameter of the gas channel (spout), around which a similar

dense sliding layer (annulus) is formed. Because of mechanical particle circulation, ﬂuid

motive-power consumption is lower by 20–25 percent than for conventional spouting,

but this gain is reduced some 5–10 percent by the power consumption of the screw.

Moreover, airﬂow rate can be chosen in quite a wide range, independent of the particle

size or density. It depends mainly on the required operation (e.g., drying, coating,

granulation). With the high gas velocity (20–30 m/s) that is favorable from t he point of

view of heat and mass transfer, air is injected into the bed tangentially through specially

designed “whirling” rings at the bottom of the bed. The bed diameter-to-height ratio

can be selected quite freely, and particle circulation time can be controlled within wide

limits. The height of the screw is always greater than that of the bed, and the bed height

can be reduced until the desired particle circulation by the screw against the tangential

air jet is obtained.

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

Published by Cambridge University Press.



Cambridge University Press 2011.

298 Tibor Szentmarjay, Elizabeth Pallai, and Judith T

oth

Air

Figure 18.1. Principle of mechanical spouting, after Szentmarjay et al.

: (a) well-mixed drying

zone, (b) mechanical spout, (c) annulus.

In the MSB one can distinguish three zones, which differ signiﬁcantly from each other

both spatially and in their ﬂow pattern

5,6

(see Figure 18.1):

(1) a zone of intensive gas–solid contact with a characteristic turbulent particle ﬂow

pattern at the bottom of the bed in the vicinity of the gas inlet;

(2) a zone of particle transport vertically upward by the screw (mechanical spout)

co-current to the gas stream; and

(3) a dense phase annular region sliding downward countercurrent to the airﬂow.

18.2 Mechanically spouted bed drying

In the beginning the ﬁeld of applications focused only on drying of particulate solids, but

the mentioned advantages have created very effective possibilities in drying of pastelike

materials, pulps, suspensions, solutions, and heat-sensitive dispersions of high moisture

content, which cannot be directly spouted. For such applications, the suspension is f ed at

several points into the annular region of the inert particles, which provide a large surface

for contacting the gas. The suspension distributes itself on the surface of the inert particles

sliding downward, and forms an almost uniform thin layer, with a mean thickness of a

few microns. The wet coating in an optimum case is two to four times thicker than the

primary particle size of the suspended material to be dried. Besides effective drying, the

other goal of this process is usually reproduction in the dried product of the primary or

Mechanical spouting 299

Air out

Suspension

tank

Air

out

Product

Figure 18.2. Mechanically spouted bed dryer, after Szentmarjay and Pallai

: (1) cylindrical

column, (2) conical bottom, (3) inert particles, (4) houseless screw, (5) heat exchanger,

(6) tangential air slots, (7) feeding pump, (8) bag ﬁlter, (9, 10) temperature and RH meters.

near-primary particle size of the suspension. The coating thickness can be inﬂuenced

by diluting the suspension and by varying the bed height or the conveying rate of the

screw – that is, by proper adjustment of the circulation time of the inert particles. At the

bottom of the bed, the wet coating is contacted by the tangentially directed high-velocity

hot air. In the wet ﬁlmlike coating, diffusion resistance can be considered negligible, so

drying takes place in a very short time, even at relatively low temperature.

6–8

Because of

the intensive friction between the inert particles and the screw, the dried ﬁne coat wears

off the surface of the particles in the screw; the ﬁne product is then elutriated from the

bed by the drying air, and it can be collected in an appropriate cyclone or bag ﬁlter. A

sketch of an MSB dryer is shown in Figure 18.2.

Considering the overall drying process, the following sequential subprocesses have

basic signiﬁcance:

Formation of wet coating on the surface of the inert particles,

Drying of the coating,

Erosion of the dried coating from the surface of the particles, and

Elutriation of the product by the air stream.

6–8

300 Tibor Szentmarjay, Elizabeth Pallai, and Judith T

oth

Table 18.1. Most important dimensions of mechanically spouted beds.

Laboratory

scale Pilot plant Industrial scale

D, m 0.12−0.14 0.38−0.42 600−1000

H, m 0.25−0.50 0.40−0.50 0.70−0.80

H/D 1.5−4.0 1.2−1.5 1.0−1.2

(screw), m 0.040−0.05 0.12−0.15 0.20−0.35

s (pitch), m 0.028−0.032 0.090−0.095 0.145−0.250

D/D

2.8−3 2.8−33

s/D

0.65−0.75 0.72 0.72

18.3 Dimensions

The diameter D

of the screw and the size of the inert particles largely determine the

bed diameter. The bed is not practical if it is smaller than 120 mm in diameter, whereas,

according to scaleup considerations (see Chapter 17), t he upper limit of the diameter

is about 1.2 m. The most important dimensions of MSB dryers are summarized in

Table 18.1.

18.4 Inert particles

Spherical particles are generally preferred as the inert charge, but in some cases cylindri-

cal bodies of equivalent diameter and height can be applied successfully. Their makeup

depends on the ﬁeld of application and hence on the physical, chemical, or biological

properties of the wet material and on the applicable temperature range. Other important

particle properties are their solidity (1 – internal porosity) and their heat and wear resis-

tance. Adhesive properties are also important but, at the same time, the dried material

must be erodable from the surface.

Many kinds of inert solids have been tested; it has been found that particles made of

heat-resistant plastics, ceramic, or glass are mainly suitable. Effective erosion removal

of the dried coating can be achieved only if the particles of the inert charge have near-

uniform diameters. Because of the well-deﬁned shape and size of the particles, the total

surface of the inert bed can be calculated almost exactly. The most frequently used inert

particles and their properties are summarized in Table 18.2.

18.5 Cycle time distribution

The residence time of the particles in the zones of the bed can be determined by

measurement of cycle times. In one study, a single inert particle, tagged by Co-60

radioactive isotope, was traced in the bed for various bed depths, revolutions of the

screw, and gas velocities.

6,7

The detector screened the whole cross-section of the bed.

Mechanical spouting 301

Table 18.2. Properties of useful inert particles.

Particles Glass beads

Al-oxide

ceramic balls

Hostaform

(polyformaldehyde) Teﬂon

Shape Spherical Cylindrical

,mm 5−8 6.6; 7.4 7.6; 12 6 × 6; 12 × 12

, kg/m

2400−2600 3520−3640 1340 2070

Max. allowable

temp. (

◦

180 400 140 220

0.05

0.3

0.2

0.1

''Spouting'' velocity (m/s)

0.15

0.05

0.25

0.07 0.09

Peripheral speed of the screw (m/s)

Δ Q=0 m

Q=0.025 m

Q=0.017 m

0.11 0.13 0.15 0.17 0.19

Figure 18.3. Velocity of particles upward in the screw at three volumetric ﬂow rates, Q, of air:

D = 0.138 m, D

= 0.04 m, s = 0.028 m, H = 0.4 m, ρ

= 1340 kg/m

, d

= 7.6 mm.

Screening started at the top; when the detector was placed at a lower height, the path of

the traced particle, both upward and downward, could be followed. With the detector at

the top of the bed, only one peak appeared in the intensity-versus-time diagram during a

cycle. At lower heights, however, two peaks were observed, and the time elapsed between

the peaks was proportional to the distance the particle traveled from the detector to the

top and back again. Thus, the particle passed the detector twice during one cycle, moving

upward in the screw and downward in the annulus. Knowing the height of the bed and

the geometry, the residence time of the tracer particle in selected sections of the bed

could be obtained. It was found that a particle spent 70 percent to 80 percent of the cycle

time in the cylindrical annulus, 5 percent to 10 percent in the screw, and 15 percent to

20 percent in the conical bottom. From the variances of the cycle time distributions and

of the residence time distributions in the three zones, it was concluded that intensive

particle mixing occur red in the vicinity of the air injection, whereas the ﬂow pattern of

the particles in the annulus could be characterized as plug ﬂow. The particles traveling

upward in the screw were surrounded by the annulus as a closed entity, and cross-

mixing between the two sections was negligible. Furthermore, the particle velocity was

independent of the gas velocity, both in the screw and in the annulus. In Figure 18.3, the

302 Tibor Szentmarjay, Elizabeth Pallai, and Judith T

oth

“spouting” velocity of the tagged particle is plotted against the peripheral speed of the

screw for different gas ﬂowrates.

It follows from the foregoing that the velocity of the particles in both the screw and the

annulus is determined essentially by the conveying rate of the screw. This was conﬁrmed

by means of a specially designed cylindrical screen basket with an oriﬁce at the bottom,

the diameter of which was identical to that of the screw. The basket was placed over

the top of the bed; after a short operating time, the screw was stopped and the particles

were weighed. It was proved that the mass ﬂow rate of the particles is a function of the

dimensions and revolution of the screw, as well as of the particle density and the voidage

ε in the screw, but is independent of the gas velocity. Thus, for calculation of the mass

ﬂow rate

m of the particles, the following expression can be used

m =

− d

πsnρ

(

1 − ε

)

(18.1)

where s is the pitch, n is the speed of revolution of the screw, d

is the screw’s axle

diameter, and ε is the loose-packed voidage.

From the mass ﬂow rate and the bed volume, as well as from the volume of the

characteristic zones, the residence time of the particles in each zone can be estimated

and controlled; thus the drying process can also be well controlled. This information

is necessary for scaleup calculations. Because of the mechanical spouting, the airﬂow

rate can be chosen very freely, ﬁrst considering the requirements for drying, given the

moisture content and other properties of the solutions or slurries to be dried. To maintain

the characteristic ﬂow pattern of the MSB, the gas velocity should be between the

elutriation velocity U

of the ﬁne product and the minimum ﬂuidization velocity U

the inert particles, that is:

< U < U

. (18.2)

18.6 Pressure drop

Gas velocity plays the primary role in determining the pressure drop across the bed,

but some effect of the screw must also be taken into account. In the annulus, the gas

velocity has been found to be mostly radially uniform, although a small increase has

been observed from the column wall to the screw.

The pressure drop p across the

bed in the MSB can be well approximated by the following modiﬁcation of p

,the

pressure drop by Ergun’s equation

for a packed bed of height H:

p = p

(0.85 − 0.09U

), (18.3)

which is valid in the range of peripheral speed of the screw 0.4 < U

< 1.6. The speed

of rotation U

(m/s) depends on the inert particles and the dimensions of the screw, as

well as on the drying task.

Mechanical spouting 303

18.7 Drying mechanism

With mechanical spouting, not only can the efﬁciency of drying be improved vis-

a-vis

conventional spouting, but special drying tasks, such as gentle drying of pastelike, sticky,

heat sensitive, and thermoplastic materials, are also made possible.

6,8,9

The drying process takes place most intensively at the conical bottom of the bed,

at a constant rate in a ver y short time. Steady-state conditions can be achieved only

when the total operating time required by the successive subprocesses does not exceed

the cycle time – in other words, the whole drying process must take place during one

circulation. According to circulation measurements and drying experiments, the inert

particles spend 10 percent to 15 percent of the cycle in the intensive drying zone,

which means in practice that a particle can reside there for an average of 5 to 10 s.

Naturally the subprocesses cannot be sharply separated from one another. Because

of the relatively slow descent velocity of the particles in the annulus, sufﬁcient time

is available for formation of the wet coating. Moreover, the drying takes place in a

bed height of a few centimeters. Therefore, the time required by the erosion process

determines the total bed height and thereby the length, as well as the rotation speed, of the

screw.

18.8 Scaleup

The optimum parameters for scaleup – t he times required by the sequential subpro-

cesses, the speciﬁc rate of water evaporation, the gas velocity, and the type and size

of inert particles – can be obtained from l aboratory-scale experiments. The height of

the drying zone can be computed from t he heat balance recommended by Mathur and

Epstein,

H =

ρc

(

1 − ε

)

− T

ou t

− T

, (18.4)

in which, for calculation of the ﬂuid-particle heat transfer coefﬁcient h

at Re

> 1000,

the equation of Rowe and Claxton

can be used.

Knowing the height of the drying zone and the speciﬁc drying performance, as well

as the amount of water to be evaporated, the total particle surface area and the volume

of the drying zone can be obtained. From the bed volume and the time required by

the drying process, the mass ﬂow rate of the inert particles can be calculated based on

Eq. (18.1). To ensure the required circulation time for the subprocesses, the cycle time

divided by the bulk-volumetric ﬂow rate of the particles yields the total volume of the

bed. Finally, dimensions of the screw should be given by calculation from the peripheral

speed of rotation in the laboratory tests.

9,13

Several large scale MSBs currently operate for production of curatives and of trace

elements for animal foods, as well as for production of basic materials for high-purity

industrial ceramics.

304 Tibor Szentmarjay, Elizabeth Pallai, and Judith T

oth

18.9 Developments

The bed height, and simultaneously the pressure drop across the bed, can be reduced

without reducing the time and length required by the erosion subprocess by the following

method: The upper section of the screw is surrounded by a tube, the l ength of which can

be extended down to the feeding point of the suspension. Thereby, the height of the bed

retaining the original annulus cross sectional area can be reduced by the length of the

tube, whereas the screw can transport the inert particles upward so the time and length

required by the erosion subprocess remain the same as before.

Another direction of recent developments i s combination of the MSB dryer with

microwave heating for drying and for heat treatment of grains. Laboratory-scale inves-

tigation has proved that very good drying efﬁciency can be achieved in this manner.

Acknowledgment

The authors express their gratitude to Professor Norman Epstein for the detailed checking

of this chapter.

References

1. K. B. Mathur and N. Epstein. Spouted Beds (New York: Academic Press, 1974).

2. A. S. Mujumdar. Spouted bed technology – a brief review. In Drying ’84, ed. A. S. Mujumdar

(New York: Hemisphere-McGraw-Hill, 1984), pp. 151–157.

3. Hungarian Pat. 176,030, U.S. Patent 4,203,226 (1977), French Patent 30,230 (1977).

4. T. Blickle, E. Aradi, and E. Pallai. Use of up-to-date processes for the drying of agricultural

products. In Drying ’80, ed. A. S. Mujumdar (New York: Hemisphere, 1980), pp. 265–271.

5. T. Szentmarjay and E. Pallai. Drying of suspensions in a modiﬁed spouted bed drier with inert

packing. Drying Technol., 7 (1989), 523–536.

6. E. Pallai, T. Szentmarjay, and A. S. Mujumdar. Spouted bed drying. In Handbook of Indus-

trial Drying, ed. A. S. Mujumdar ( London: CRC Press/Taylor and Francis Group, 2006),

pp. 363–384.

7. T. Szentmarjay, E. Pallai, and A. Szalay. Drying process on inert particles in mechanically

spouted bed dryer. Drying Technol., 13 (1995), 1203–1219.

8. T. Szentmarjay, E. Pallai, and Zs. Reg

enyi. Short-time drying of heat-sensitive biologically

active pulps and pastes. Drying Technol., 14 (1996), 2091–2115.

9. T. Szentmarjay and E. Pallai. Drying experiments with AlO(OH) suspension of high purity

and ﬁne particulate size to design an industrial scale dryer. Drying Technol., 18 (2000),

759–776.

10. T. Szentmarjay, A. Szalay, and E. Pallai. Hydrodynamic measurements in mechanically

spouted bed. Hung. Ind. Chem., 20 (1992), 219–224.

11. S. Ergun. Fluid ﬂow through packed columns. Chem. Eng. Progr., 48 (1952), 89–94.

12. P. N. Rowe and K. T. Claxton. Heat and mass transfer from a single sphere to ﬂuid ﬂowing

through an array. Trans. Instn Chem. Engrs., 43 (1965), T321–T331.

13. T. Szentmarjay, A. Szalay, and E. Pallai. Scale-up aspects of the mechanically spouted bed

dryer with inert particles, Drying Technol., 12 (1994), 341–350.

19 Catalytic reactors and their modeling

Giorgio Rovero and Norberto Piccinini

In an earlier review on modeling of catalytic gas–solid ﬂuidized bed reactors,

spouted

beds were treated as a special case of ﬂuidized beds, with the jet region extending

to the upper free surface of the solids emulsion. The spout and the annulus are two

well-distinguished zones, characterized by such different gas–solid contacting that they

have been treated distinctly since the ﬁrst approach

for modeling spouted bed chemical

reactors. The spout zone acts as a dilute vertical transport system, whereas the peripheral

annulus has many similarities with a countercurrent downﬂow packed bed, the major

peculiarity being gas percolation into the dense emulsion from the spout. Therefore,

solids motion in a spouted bed is a major factor affecting mass and heat transfer,

mixing, contacting, and hence the choice of reactor model. The solids can react (e.g., in

combustion, where they are depleted over time), constitute inerts (e.g., sand diluting a

solid fuel and regulating the system hydrodynamics, as well as providing a heat buffer),

or act as a catalyst.

Successful design and operation of a reactor depend on the ability to predict the

system behavior, especially the hydrodynamics, mixing of individual phases, heat and

mass transfer rates, and kinetics of the reactions involved. A model for spouted bed

reactors must account for design equations (e.g., H

, U

), basic hydrodynamics (e.g.,

, U

(z), U

, U

, distributed P proﬁles), geometric and physical parameters (e.g.,

D, D

, θ , d

, ρ

), solids circulation pattern and mixing, mass or mole balances, transfer

between phases (spout and annulus, gas and catalyst particles), hydraulic assumptions

and phenomenological observations (e.g., perfect mixing, plug ﬂow, and radial or axial

dispersion), transport mechanism hypotheses, heat and mass balances, and kinetic rate

expressions.

19.1 Fundamental modeling

The earliest fundamental model to predict the chemical conversion in a spouted bed

reactor, by Mathur and Lim,

and recapitulated by Mathur and Epstein,

established the

basis both f or subsequent theoretical studies and for experimental testing, and presented

a general approach for cases in which the solid phase acts as a catalyst or heat carrier.

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

Published by Cambridge University Press.



Cambridge University Press 2011.

306 Giorgio Rovero and Norberto Piccinini

Figure 19.1. Two-region model of a spouted bed catalytic reactor.

The model assumes t he existence of two regions, the spout and annulus; in both, the

gas moves upward in “one-dimensional ﬂow” according to the scheme presented in

Figure 19.1. The model was developed for an isothermal constant-volume reaction.

Mass balances in the two regions, together with continuity equations, provide a full ﬂow

description. A ﬁrst-order reaction was promoted by a highly porous particulate catalyst,

in which mass transfer and diffusional resistances can be neglected. The convective gas

crossﬂow from the spout to the annulus satisﬁed the Mamuro and Hattori equation.

was either derived from previous experiments or calculated for D in the range 0.15 to

0.61 m, with U

and D

based on previous correlations. The spout voidage was set at

a constant value of 0.95 for the largest column or matched to observed proﬁles for the

smallest unit; in all cases the contribution to the overall conversion from the gas leaving

the spout was predicted to be minimal. No contribution to the reaction was included for

the catalyst in the fountain, even though the spout was extended to reach the highest

elevation of the fountain itself. The conversion at the exit is given by the weighted

average of the concentrations of the two regions, obtained by integrating the differential

gas mass balance in the spout and the annulus:

+ k

(

1 − ε

)

= 0 (19.1)