Influence of pressure on fluidization properties

www.elsevier.com/locate/powtec

Powder Technology 141 (2004) 137–154

Influence of pressure on fluidization properties

Igor Sidorenko, Martin J. Rhodes*

Cooperative Research Centre for Clean Power from Lignite, Department of Chemical Engineering, Monash University, P.O. Box 36, Clayton,

Victoria 3800, Australia

Received 27 August 2003; received in revised form 12 December 2003; accepted 25 February 2004

Available online 13 April 2004

Abstract

An experimental study of the influence of pressure on fluidization phenomena was carried out in a bubbling fluidized bed over the range of

operating absolute pressures 101–2100 kPa with Geldart A and B bed materials in an apparatus 15 cm in diameter. Bed voidage data was

gathered using electrical capacitance tomography (ECT) and the resulting time series were analysed to give average bed voidage, average

absolute deviation, average cycle frequency and Kolmogorov entropy. Bed collapse tests were also performed to establish the influence of

operating pressure on dense phase voidage. For theGeldart B powderUmf was found to decrease slightly with pressure whilst voidage atUmf was

unaffected and no change in behavior from Geldart B to A was observed. Analysis of the ECT results for the Geldart B powder showed no

influence of pressure on cycle frequency and average absolute deviation, but an increase in mean bed voidage with pressure. For the Geldart A

powder, no influence of pressure on Umf, Umb and emf was observed, though emb was found to increase slightly with pressure. Analysis of the

ECT results for the Geldart A powder revealed a decrease in average bed voidage, cycle frequency and average absolute deviation with pressure.

The trend in Kolmogorov entropy within increasing pressure was not entirely revealed over the range of velocities used, but results suggest a

decrease.

D 2004 Elsevier B.V. All rights reserved.

Keywords: Fluidization; Pressure effect; Electrical capacitance tomography

1. Introduction tion velocity is not a quantity with a precise significance for

Fluidized bed technology originated as the Winkler pro-

cess for lignite gasification developed in the 1920s and the

fluidization technique as it is known now was born from the

successful development of the fluidized bed catalytic crack-

ing process in the early 1940s. It has been known for some

time that the use of higher operating pressures increases the

rate of processes in fluidized beds and improves fluidization

quality [1]. Although some data on the effect of pressure on

the behavior of fluidized beds have been obtained during the

past 50 years and several comprehensive review papers have

been published [2–5], areas remain where further experi-

mental work and analysis would be warranted.

In industrial practice fluidized bed reactors are mostly

operated at superficial gas velocities well above the minimum

fluidization velocities and, therefore, the minimum fluidiza-

0032-5910/$ - see front matter D 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.powtec.2004.02.019

* Corresponding author. Tel.: +61-3-9905-3445; fax: +61-3-9905-

5686.

E-mail address: [email protected] (M.J. Rhodes).

industrial applications, however discussion on accurate pre-

diction of the minimum fluidization velocity still seems to

remain of much scientific interest. The effect of pressure on

the minimum fluidization velocity has been studied by many

researchers [6–19]. The common finding is that the effect of

pressure on the minimum fluidization velocity depends on

particle size. The results of the experiments show a decrease

in the minimum fluidization velocity with increasing pressure

for particles larger than 100 Am (Geldart B and D materials);

however, for fine Geldart A particles the minimum fluidiza-

tion velocity is unaffected by pressure.

Under ideal experimental conditions, Abrahamsen and

Geldart [20] expected a very weak effect of pressure on the

minimum bubbling velocity. However, this effect is still

relatively unknown and findings in the literature are some-

what controversial. Sobreiro and Monteiro [18] found that

the influence of pressure on the minimum bubbling velocity

could be very different for different bed materials, but failed

to explain their results. According to Jacob and Weimer

[21], it has generally been found that with increasing

pressure for systems of Geldart A powders the minimum

Fig. 1. Pressurized fluidized bed apparatus used in this study.

I. Sidorenko, M.J. Rhodes / Powder Technology 141 (2004) 137–154138

bubbling velocity increases, however, the extent of the

pressure effect and the reason for it is uncertain.

There are also conflicting reports on the effect of increas-

ing pressure on the bed voidage at minimum fluidization (emf)

and minimum bubbling (emb) conditions. According to Bin

[22], for industrial design purposes the assumption that emf

does not change with pressure and that emf can be taken as that

experimentally determined at atmospheric conditions is jus-

tified for the majority of solids. Several researchers reported

that the voidage at minimum fluidization emf is essentially

independent of pressure [11,18,19]. Yang et al. [23]suggested

that the variation of the voidage at incipient fluidization

caused by increase in pressure is negligible for Geldart B

and D materials and expected an appreciable change with

pressure for Geldart A. Weimer and Quarderer [24] observed

only very small increase in bed voidage at minimum fluid-

ization with increase in pressure for Geldart A materials and

no change in voidage for Geldart B solids. Contrary to that,

Olowson and Almstedt [16] observed a slight increase of the

voidage at minimum fluidization with pressure for Geldart D

particles and did not notice any dependence of the voidage emf

on pressure for Geldart B particles.

Rowe [25] presented some of his calculations which

showed no measurable change in predicted minimum bub-

bling voidage, however, these predictions are not in agree-

ment with experimental findings of other researchers

[11,26,27] who generally observed an increase in minimum

bubbling voidage with pressure.

When fluidized beds operate in the bubbling regime, they

consist of the dense (emulsion) phase and bubbles. The

dense phase voidage is an important parameter in determin-

ing the fluidized bed performance because it is widely

believed that it has a direct influence on chemical reactions

in the fluidized bed. It affects the degree of gas–solid

contact and heat and mass transfer. It is also believed that

in Geldart A powders the equilibrium size of the bubbles

may be controlled by the dense phase voidage [28]. The

dense phase voidage is generally taken as being equal to its

value at the minimum fluidization point for Geldart B and D

materials, but can be higher for Geldart A powders [28,29].

A few methods have been used to measure the dense

phase voidage in bubbling fluidized beds. For Geldart A bed

materials the voidage can be measured by means of the bed

collapse technique. This method has a few variations but

generally it involves abruptly stopping the gas flow to a

vigorously bubbling fluidized bed and measuring the rate of

collapse of the bed surface (e.g. Refs. [30–32]).

Other techniques for measuring voidage in fluidized beds

involve the use of capacitance probes, optical fibres, X-ray

or g-ray attenuation and capacitance tomographic imaging

[29,33]. A non-invasive technique for fast measurement of

solids volume fraction in bubbling fluidized beds was

pioneered at Morgantown Energy Technology Centre

(METC) in the US and at the University of Manchester,

Institute of Science and Technology (UMIST) in the UK. By

the early 1990s both development groups had successfully

demonstrated the technique, which became known as the

electrical capacitance tomography (ECT).

In this work, the influence of pressure on fluidization

phenomena, such as minimum fluidization velocity, mini-

mum bubbling velocity, bed expansion and voidage was

studied; and a novel diagnostic technique suitable for online

monitoring of fluidized bed behavior was tested. Experi-

ments were conducted in a bubbling fluidized bed at

operating pressures up to 2100 kPa and ambient temperature

with Geldart A and B materials [34]. Pressure probe, visual

observation and the electrical capacitance tomography were

used for characterizing the fluidized bed behavior at elevat-

ed pressure. Although the ECT technique has been used in

fluidization research [35–37], it has not been previously

used to study the fluidized bed behavior in a pressurized

environment.

2. Experimental

2.1. Equipment

The process and instrumentation diagram of the pressur-

ized fluidized bed used in this study of pressure effects on

fluidization is given elsewhere [38]. The pressure vessel,

capable of operating at pressures up to 2600 kPa, was 2.38

m high and was equipped with five glass observation ports.

One of two 15-cm diameter plastic fluidized bed models

was inserted into the pressure vessel and used to study

physical behavior of gas-fluidized beds (Fig. 1). Both

I. Sidorenko, M.J. Rhodes / Powder Technology 141 (2004) 137–154 139

fluidized bed columns were identical and made of clear

plastic; however the second column had a permanently

mounted capacitance sensor described below.

The bottom flange of the column was bolted to a carbon

steel plenum chamber and distributor assembly, and the top

flange was bolted to an expanded top freeboard for solids

disengagement. The plenum chamber and distributor as-

sembly were designed and manufactured according to

recommendations given by Svarovsky [39]. Two types of

distributor plates were specially made for the study. The

first distributor was made from a sintered bronze plate with

nominal pore size of 12 Am and the pressure drop through

the plate of 4.9 kPa at a superficial gas velocity of 0.5 m/s.

The second distributor was designed for pressure drop

experiments for Geldart A materials so that the pressure

drop was approximately 6.5 kPa at a superficial gas

velocity of 0.01 m/s in accordance with recommendations

by Svarovsky [39].

The assembled fluidized bed column was inserted into

the pressure vessel and positioned vertically. Provisions

were made in the design of the lid of the pressure vessel

for charging/discharging of solids and inserting bed differ-

ential pressure probe. Fluidizing gas for experiments was

supplied either from a building compressed air supply or

from dedicated nitrogen and compressed air cylinders and

metered by variable area flow meters. According to the

German Standard VDI/VDE 3513 Part 2, the rotameters of

the accuracy class 1.6 were properly compensated for

difference in gas properties at different operating pressures

and temperatures.

The operating pressure in the fluidized bed was set with a

backpressure control valve. The experimental facility was of

an open circuit type so the same gas was used for pressur-

izing the system and for fluidizing the solids. The apparatus

was pressurized to the desired operating pressure before the

fluidizing gas flow rate was adjusted to a desired value.

Bed pressure drop was measured using a simple differ-

ential pressure probe, having one leg at distributor level and

the other in the freeboard, and connected to a pressure

transducer and microprocessor-based data acquisition unit.

The electrical capacitance tomography (ECT) system

used in this work was a single plane system with driven

guard drive circuitry of PTL300 type manufactured by

Process Tomography.

The required electrode pattern for this work was designed

in accordance with PTL application note [40]. The maxi-

mum number of measurement electrodes (12) was used

together with driven guard electrodes. The sensor was

fabricated using standard printed circuit board design tech-

niques from flexible copper-coated plastic laminate that was

etched with the electrode pattern and wrapped tightly

around the fluidized bed plastic vessel. The sensor consisted

of 12 rectangular electrodes, 35 mm wide and 50 mm high.

The centers of the electrodes were located at the height of

250 mm above the distributor plate, or approximately at a

half of the static bed height used in the experiments.

In accordance with PTL recommendations, the measure-

ment and guard electrodes were connected to the data

acquisition module by 1.5-m-long screened coaxial cables,

terminated in miniature coaxial connectors. The capacitance

sensor had to be positioned inside the pressure vessel and,

since the data acquisition module was not designed for

operation at elevated pressure, it was located outside of the

pressure vessel. Therefore, it became necessary to direct 24

coaxial cables and an earth wire out of the pressure vessel

without gas leaks and pressure loss. The detail of how this

was achieved is given in Sidorenko [41].

Compressed air from a centralized building supply was

usually used for experiments when operating pressure did

not exceed 600 kPa. For other experiments that were usually

run at higher operating pressure, industrial grade nitrogen

supplied from a bank of 12 gas cylinders was used. Geldart

A FCC catalyst and Geldart B silica sand were used as bed

solids.

Geldart A FCC catalyst with the Sauter mean diameter of

77 Am contained 10.1% of fines less than 45 Am and had the

particle density of 1330 kg/m3. Particle size distributions

were measured using a Malvern Instruments Mastersizer

and particle density was determined by using a mercury

intrusion analysis. Seven kilograms of the powder were

used as the bed material for experiments to observe the

minimum fluidization and minimum bubbling conditions,

which gave a static bed height of 53.5 cm. For experiments

using the ECT system, 5 kg were used, and the bed height

was 44 cm. For bed collapse experiments, 6.56 kg of the

catalyst were used, which gave the bed height of 49.7 cm.

Geldart B silica sand of the narrow size distribution had

the Sauter mean diameter of 203 Am and the particle density

of 2650 kg/m3. Nine kilograms of the powder were used as

the bed material for experiments to observe the minimum

fluidization and minimum bubbling conditions, which gave

a static bed height of 40 cm. For experiments using the ECT

system, 10 kg were used, and the bed height was 42 cm.

The instruments downstream of the fluidized bed were

protected by a line filter capable of removing particles down

to 0.3 Am. Simple filter checks after each experiment

indicated that the fines carryover was negligible.

2.2. Minimum fluidization velocity

A single set of experiments was carried out in order to

determine the minimum fluidization velocity, minimum

bubbling velocity, bed expansion and voidage at minimum

fluidization and minimum bubbling conditions; and the

details of the experimental procedure are given here. Bed

pressure drop measurements were obtained for both increas-

ing and decreasing gas velocities. The minimum fluidization

velocity is taken as the velocity at the intersection point of

the line corresponding to the constant bed pressure drop in

the fluidized state and the extrapolated straight line of the

packed bed region. There is no agreed procedure for

determining the precise point of the minimum fluidization

I. Sidorenko, M.J. Rhodes / Powder T140

velocity and, usually, the two straight lines are obtained as

the gas flow rate is gradually reduced in increments from a

vigorously bubbling state. In this case, a slightly larger

value of the minimum fluidization velocity is obtained. That

makes sense for industrial applications as it provides a

maximum experimental value for the minimum fluidization

velocity and, therefore, the lowest limit for potential defluid-

ization of a process.

According to Svarovsky [39], better reproduction of

results can be obtained by allowing the bed to mix first

by bubbling freely before turning the gas flow rate down to

zero and then taking pressure drop measurements, while

increasing gradually the gas flow rate. However, various

researchers have described three different points represent-

ing the minimum fluidization velocity on the increasing gas

velocity curve for Geldart A materials, as reviewed by

Fletcher et al. [42].

A certain procedure was followed for preparing the bed

for the experiments. Prior to each experiment at ambient

conditions, the pressure vessel was open and the air supply

pressure was set to the gauge pressure of 100 kPa. The bed

was fluidized at superficial gas velocities well above the

onset of fluidization for at least 15 min, allowing bed solids

to fully mix. To achieve an initial packed bed of similar

structure in each experiment, the air supply was then slowly

turned off by closing the flow control valve.

In high-pressure experiments the same gas was used for

pressurizing the system and then for fluidizing the bed

material. Therefore, prior to each experiment, the back-

pressure controller was set to a pre-determined operating

pressure and the system was pressurized to that level first. At

the same time, the gas passed through the fluidized bed,

which was properly fluidized for much longer than at the

ambient conditions. Then the gas supply was slowly turned

off, allowing the bed to settle in the pressurized environment.

In actual experiments, measurements of the pressure drop

were taken for both increasing and decreasing gas velocities.

After each gas velocity change, the pressure in the system

was allowed to stabilize for at least 10 min and then pressure

drop measurements were recorded at a frequency of 1 Hz for

at least 3 min. For each experiment a straight line was drawn

from the origin through the series of bed pressure drop points

until it crossed the horizontal line corresponding to bed

weight per unit area (there was generally good agreement

between measured data and this horizontal line). The point of

intersection was taken as the value of minimum fluidization

velocity at a given experimental condition. Where measure-

ments showed a hysteresis effect between increasing and

decreasing gas velocities, both intersection points were

established as lower and upper values of the minimum

fluidization velocity at a given experimental condition. The

difference between the lower and upper values was generally

small, and for comparison between experiments at different

operating pressures the average of these two points was taken

as the minimum fluidization velocity. These experimental

curves are presented in Sidorenko [41].

2.3. Minimum bubbling velocity

The minimum bubbling velocity was measured by visual

observation of the point at which the first bubbles appeared

and by examination of the bed height variation with gas

velocity. Visual observation of the fluidized bed behavior

provides a common and simple but subjective way of

determining the minimum bubbling velocity. When the gas

flow is gradually increased, the gas velocity at which the first

distinct bubbles appear on the bed surface is recorded.

Alternatively, the gas velocity is noted at which bubbling

stops when the gas flow is decreased. According to Geldart

[28] and Svarovsky [39], the average of the two values

provides the minimum bubbling velocity, or more appropri-

ate values can be taken during decreasing gas flow. In this

work, measurements of the minimum bubbling velocity were

taken for both increasing and decreasing gas flow.

At each gas velocity the bed height was recorded and,

based on visual observation, general behavior of the bed

was noted. The first bubbles were not confused with the

small channels, which sometimes appeared and resembled

small volcanoes. Genuine bubbles usually appeared in

several places on the bed surface and were about 10 mm

in diameter; while the channels were smaller in diameter and

usually stayed in one place.

The minimum bubbling velocity was also determined

less subjectively as the velocity at which the maximum bed

height was observed [43].

2.4. Bubbling bed voidage

Experiments using the ECT equipment were carried out

at operating pressures ranging from 300 to 1900 kPa, using

Geldart A FCC catalyst as bed material, and ultra-pure

nitrogen and compressed air as fluidizing gases. Another

series of experiments was conducted at operating pressures

ranging from ambient to 1700 kPa, using Geldart B silica

sand as bed material, and industrial nitrogen and com-

pressed air as fluidizing gases.

Before commencing experiments, the ECT system was

calibrated by filling the sensor with the two reference

materials in turn and by measuring the inter-electrode

capacitances at the two extreme values of relative permit-

tivity (empty vessel and packed bed). The packed bed was

formed each time under relevant pressure and following the

same procedure (vigorous fluidization followed by gradual

defluidization).

Since some drift in the very sensitive electronic capac-

itance measuring circuitry was observed after a few hours of

fluidization, it became necessary to carry out such a cali-

bration twice a day and check its accuracy before and after

each experiment.

In actual experiments at pre-determined operating pres-

sure, the gas velocity was an experimental variable. After

each gas velocity change, the pressure in the system was

allowed to stabilize for at least 10 min. When both pressure

echnology 141 (2004) 137–154


and gas flow became stable, the tomographic data were

logged on an ECT dedicated computer. The ECT system

generated a cross-sectional image of the bed and showed the

solids volume fraction (the ratio of solids to gas) at sensor

level. Under each set of operating conditions, the behavior

of the bed was also characterized by visual inspection and

recording of bed height versus gas velocity.

2.5. Bed collapse method

Bed collapse experiments were carried out at room

temperature and the operating pressures ranging from atmo-

spheric to 1600 kPa, using 6.56 kg of Geldart A FCC

catalyst as bed material, and industrial nitrogen as fluidizing

gas. The experimental program was based on a ‘‘single-

vented’’ technique given by Geldart and Wong [44].

The bed collapse experiment was carried out as follows:

the gas flow rate was set at a predetermined level so that the

bed was bubbling, and the average bed height was recorded.

The gas supply was quickly isolated and the bed collapse

was recorded on a video camera at a rate of 25 frames per

second. The plenum chamber was not vented simultaneous-

ly. The combined internal volume of the chamber and the

pipeline downstream of the isolation valve was estimated to

be equal to 0.00115 m3. Since the bed height fluctuated due

to bubbling, five repeat tests were made and an average bed

height was taken. Similar tests were also carried out at

different predetermined gas flow rates.

Because of restrictions to our gas supply, we used initial

velocities in the range 0.054 m/s at atmospheric pressure to

0.008 m/s at 1600 kPa—somewhat below the 0.04 m/s

recommended by Geldart [45] and the higher values used by

Geldart (0.08 m/s in Geldart and Wong [44] and 0.1 m/s in

Geldart and Xie [30]).

3. Experimental results and discussion

3.1. Minimum fluidization velocity

Many researchers have previously studied the pressure

effect on the minimum fluidization velocity and found that

the minimum fluidization velocity is not affected by pressure

for fine Geldart A powders and decreases with pressure in-

crease for coarse materials. Numerous correlations for pre-

dicting the minimum fluidization velocity have been sugges-

ted with the most widely used correlation proposed by Wen

and Yu [46]. However, this popular correlation was based

only on data obtained at atmospheric pressure. Four similar

correlations based on experiments carried out at elevated pre-

ssures have also been proposed in the literature [6,9,15,17].

All the correlations are based on the Ergun [47] equation,

and are simplified by assuming certain fixed values for the

bed voidage at the minimum fluidization and constant

particle shape factor. Previously it was quite often found

that the absolute values of predictions based on the corre-

lations, available in the literature, were significantly in error

[12,14,16].

One of the better methods for prediction of the minimum

fluidization velocity, based on using the Ergun [47] equa-

tion, was suggested by Werther [48] and has been used in

the present work. Values of the minimum fluidization

velocity experimentally determined at ambient conditions

and the voidage at minimum fluidization, described in the

next section, were used to calculate a characteristic particle

diameter udp from the Ergun equation. Following this

method, which is described in Sidorenko and Rhodes [4],

the minimum fluidization velocity values at various operat-

ing pressures were calculated.

A comparison between the experimentally measured

values of the minimum fluidization velocity and the calcu-

lated values from the Ergun [47] equation and the available

correlations [6,9,15,17,46] is shown in Figs. 2 and 3.

For a Geldart A material, Fig. 2 shows that within the

studied pressure range the minimum fluidization velocity was

independent of the operating pressure, which is in line with

observations of other workers (e.g. Refs. [9,11,18,26,49,50]).

Many researchers (e.g. Refs. [6–19]) have observed the

decrease in the minimum fluidization velocity with increas-

ing pressure for Geldart B materials. Our results also show a

slight decrease in minimum fluidization velocity with in-

creasing pressure in the range up to 1600 kPa.

Depending on the physical properties of particles, the

correlations agree with the experimental results with vari-

able success. For the FCC catalyst, Fig. 2 shows that the

Ergun [47] equation and the correlations of Nakamura et al.

[15], and to some extent of Wen and Yu [46] predict the

minimum fluidization velocity values very well. However,

the correlations of Chitester et al. [9], and to larger extent of

Saxena and Vogel [17] and Borodulya et al. [6] overestimate

the minimum fluidization velocity values. Similar results

can be observed for the silica sand Geldart B material in Fig.

3, where only the Ergun [47] equation agrees well with the

experimental results and all the correlations more or less

overestimate the minimum fluidization velocity values.

3.2. Voidage at minimum fluidization conditions

Knowing the mass of the bed solids m and the bed cross-

sectional area A, the experimental values of the voidage at

minimum fluidization emf were determined, using measured

bed height Hmf values corresponding to the minimum

fluidization velocity at both increasing and decreasing gas

velocity. The average of the two values was used as a

parameter for comparison at various operating pressures.

It was experimentally found that the voidage at minimum

fluidization was essentially independent of pressure for:

� Geldart A FCC catalyst (emf = 0.42) in a pressure range

101–1100 kPa.� Geldart B silica sand (emf = 0.50) in a pressure range

101–1600 kPa.

Fig. 2. Variation of the minimum fluidization velocity with pressure for Geldart A FCC catalyst. Comparison of experimental values (clear squares) with

predictions from the Ergun equation [47,48] (line 4) and correlations of Borodulya et al. [6] (line 1); Chitester et al. [9] (line 3); Nakamura et al. [15] (line 5);

Saxena and Vogel [17] (line 2); and Wen and Yu [46] (line 6).


Other workers [11,18,19] also observed the indepen-

dence of the voidage at minimum fluidization on pressure.

Yang et al. [23] suggested, however, that the variation of the

voidage at minimum fluidization caused by a pressure

increase would be very small, if at all, for coarse materials

and substantial for Geldart A powders. That was not

observed in the present study. Weimer and Quarderer [24]

carried out their experiments at much higher pressures and

did not observe any voidage change for a Geldart B material

and noticed only very small increase in the voidage at

minimum fluidization for Geldart A materials.

Fig. 3. Variation of the minimum fluidization velocity with pressure for Gelda

predictions from the Ergun equation [47,48] (line 6) and correlations of Borodulya

Saxena and Vogel [17] (line 1); and Wen and Yu [46] (line 5).

3.3. Minimum bubbling velocity

The silica sand was found to exhibit Geldart B behavior,

with coincidence of minimum bubbling and minimum

fluidization velocities, for all pressures up to 1600 kPa.

This result is in agreement with the findings of King and

Harrison [11] at pressures up to 2500 kPa, but at odds with

the findings of Sobreiro and Monteiro [18] and Sciazko and

Bandrowski [51,52], who were able to establish separate

values for the minimum bubbling and minimum fluidization

velocities within the similar pressure range, as well as at

rt B silica sand. Comparison of experimental values (clear squares) with

et al. [6] (line 2); Chitester et al. [9] (line 3); Nakamura et al. [15] (line 4);


higher pressures up to 3500 kPa. These results are also

contrary to those of Varadi and Grace [53], obtained within

the pressure range from atmospheric to 2170 kPa. However,

it should be noted that the often-cited work of Varadi and

Grace [53] is only a preliminary study carried out in a two-

dimensional bed.

FCC catalyst showed a typical Geldart A behavior,

expanding without bubbling within a certain gas velocity

range at ambient conditions. At all the experimental con-

ditions the height of the homogeneously expanded bed of

the FCC catalyst was at least 0.6 m. In all cases the

minimum bubbling velocity values, determined visually,

coincided for both increasing and decreasing gas flow and

were equal to values, obtained from the maximum bed

height.

Only one correlation for determining the minimum

bubbling velocity has been proposed [20]. This equation

is not commonly used on its own as Abrahamsen and

Geldart [20] combined it with another correlation for the

minimum fluidization velocity [54] and proposed a more

popular relation:

Umb

Umf

¼2300q0:126

g l0:523expð0:716FÞd0:8p g0:934ðqp � qgÞ

0:934ð1Þ

Thus, Geldart A behavior corresponds to the velocity

ratio being more than unity. The minimum bubbling to

minimum fluidization velocity ratios calculated from the

experimental values of the minimum bubbling and mini-

mum fluidization velocities and compared to the predictions

of Eq. (1) are presented in Fig. 4.

As can be seen in Fig. 4, Eq. (1) predicts an increase

in the velocity ratio with increasing pressure but the

accuracy of prediction is quite low. It overestimates the

Fig. 4. Variation of the minimum bubbling to minimum fluidization velocity ratio w

(clear squares) with predictions from the Abrahamsen and Geldart [20] correlatio

experimental values for the FCC catalyst which show no

pressure effect on the velocity ratio over the range

considered.

Within the higher pressure range (2140 to 6900 kPa) and

in a smaller column (2.7 cm in diameter), Crowther and

Whitehead [55] fluidized 50 g of fine synclyst particles with

a high density gas (CF4) and found that the minimum

bubbling velocity was equal to 0.008 m/s at 2140 and

2820 kPa, and increased to 0.009 m/s at 4180 kPa and to

0.011 at 5540 kPa.

In another study within the pressure range similar to

present work, Richardson [56] fluidized Perspex particles at

pressures between atmospheric and 1400 kPa and reported a

‘‘slight increase’’ in the velocity ratio from 2.2 to 2.8.

However, when Richardson [56] fluidized similarly sized

phenolic resin particles, the velocity ratio trend was not

obvious.

It should be noted that, since both the minimum fluid-

ization and minimum bubbling velocities are very low, a

small variation of just 0.5 mm/s in either of them could

position the velocity ratio anywhere in the range between

two and three, as can be illustrated by substantial error bars

in Fig. 4.

3.4. Voidage at minimum bubbling conditions

Abrahamsen and Geldart [20] correlated the maximum

non-bubbling bed expansion ratio in the following way:

Hmb

Hmf

¼5:50q0:028

g l0:115expð0:158FÞd0:176p g0:205ðqp � qgÞ0:205

ð2Þ

According to this correlation, the maximum bed expan-

sion ratio should slightly increase with pressure. A compar-

ith pressure for Geldart A FCC catalyst. Comparison of experimental values

n.


ison between the maximum bed expansion ratio determined

experimentally and that predicted from Eq. (2) is shown in

Fig. 5.

As can be seen in Fig. 5, the effect of pressure on the

maximum bed expansion within the pressure range 101–

1100 kPa was found to be negligible for the Geldart A

material used in this work. Eq. (1) overestimates the

experimental values for the FCC catalyst. The experimental

results and predictions from Eq. (2), however, contradict the

experimental results presented by Piepers et al. [26] who

reported a substantial increase in the total bed expansion

with pressure increase up to 1500 kPa.

Values of the voidage at minimum bubbling emb were

determined from bed height measurements. In this study, the

bed voidage at minimum bubbling for FCC catalyst was

found to increase from 0.47 at ambient conditions to 0.49 at

1000 kPa. Similar small increases caused by pressure was

noticed for some Geldart A materials by other workers

[11,18,49,57–59]. Piepers et al. [26] observed slightly

greater increase in the minimum bubbling voidage for 59-

Am sized catalyst within the similar pressure range (from

0.58 at atmospheric pressure to 0.63 at 1500 kPa).

3.5. Bubbling bed voidage at elevated pressure

For FCC catalyst and silica sand the ECT system

generated two types of data:

� time series image frames of the local solids volume

fraction distribution at a given horizontal level in the

form of a 32� 32 pixels matrix;� time series data points of the average solids volume

fraction representing the whole bed at a given horizontal

level.

Fig. 5. Variation of the maximum bed expansion ratio with pressure for Geldart

predictions from the Abrahamsen and Geldart [20] correlation.

It is accepted that the tomographic images obtained from

the capacitance measurements are usually of relatively low

resolutzion (e.g. Refs. [60–62]). A linear back-projection

algorithm, supplied as standard with the PTL300 system, is

mathematically simple and fast because the image recon-

struction process is reduced to matrix–vector multiplication.

However, the images are only approximate and suffer from

blurring. Although, image resolution and accuracy can be

improved by employing an iterative linear back-projection

algorithm, this and other reconstruction algorithms are still

under development [61,63–65].

According to the Process Tomography Application Note

[66], the linear back-projection algorithm will always un-

derestimate areas of high permittivity and overestimate areas

of low permittivity. The images produced by this method

will always be approximate, since the method spreads the

true image over the whole sensor area. Because the image

has been spread out over the whole area, the magnitude of

the image pixels will be less than the true values. However,

the sum of all of the pixels will approximate to the true

value, and the linear back-projection algorithm is therefore

useful for calculating the average voidages.

At each operating pressure, measurements of relative

solids volume fraction were taken at a number of superficial

gas velocities above the minimum fluidization velocity. At

each given gas velocity, 16,000 frames of data were logged

over more than 3 min of dynamic operation and recorded by

the ECT system at the rate of approximately 81 frames per

second. The output, which was in the form of 32� 32 pixels

matrix was averaged and provided for each frame the

average volume fraction as a single value between 0 (gas)

and 100 (packed bed). From 16000 values of the relative

voidage and the sample frequency the relevant time series

statistics were calculated.

A FCC catalyst. Comparison of experimental values (clear squares) with


Three parameters were selected for further analysis—

average, average absolute deviation and average cycle

frequency. Although some analysis techniques were origi-

nally proposed for pressure fluctuations time series, in this

work they have been applied to the analysis of the voidage

fluctuations resulting from the ECT measurements. Accord-

ing to Schouten et al. [67], the average absolute deviation of

the data points from the solids volume fraction average

value is a robust estimator of the width of time series around

the mean. According to Zijerveld et al. [68], the average

absolute deviation is comparable with the standard deviation

and shows similar dependence on operating conditions. The

average cycle frequency is the reciprocal of the average

cycle time, which is defined as the average time to complete

a full cycle after the first passage through the time series

average. These three parameters make it possible to com-

pare results between measurements at various gas velocities

and different operating pressures, since their values are

unambiguous and can be readily calculated. The time series

data were also used in the reconstruction of an attractor from

which the Kolmogorov entropy was estimated for use in

chaos analysis.

3.5.1. Average total bed voidage

The experimental values of the average relative solids

volume fraction were expressed in percentage points, and

under the existing operating conditions fluctuated some-

where between 100 in a non-fluidized state and 55 in a

bubbling regime. The value of 100 was established during

the calibration procedure and corresponded to a packed bed.

Based on a known packed bed height, the packed bed

voidage was calculated to be equal to 0.476 for the FCC

catalyst and 0.463 for the silica sand. For both materials,

calibration procedures were completed at various operating

Fig. 6. Variation in average bed voidage with gas velocity for FCC catalyst over a

the eye).

pressures and the packed bed height, and therefore voidage,

did not vary with pressure.

Fig. 6 shows the measured influence of superficial gas

velocity and operating pressure on the average bed voidage

for FCC catalyst. Average bed voidage reduces with in-

creasing gas velocity above the minimum bubbling velocity

of 0.006 m/s and up to about 0.03 m/s, and generally

increases thereafter. This result is consistent with the find-

ings of other researchers at ambient pressure [44,45,69,70]

who found that at superficial gas velocities up to 0.03 m/s

the bed voidage decreases with increase in gas velocity.

It might be expected that the total bed expansion would

increase if more gas is put into the fluidized bed as bubbles.

However, according to Geldart et al. [69], the reduction in

the bed expansion just above the minimum bubbling veloc-

ity in Geldart A powders occurs because the dense phase

volume in the bubbling bed is reduced more rapidly than the

bubble hold-up increases. According to Geldart and Radtke

[70], this reduction in dense phase voidage is due to small

interparticle forces which make the powder slightly cohe-

sive and are continually disrupted by the bubbles passage

and increase in powder circulation.

The effect of operating pressure on the cross-section

averaged total bed voidage for the Geldart A FCC catalyst

can be clearly seen in Fig. 6. This diagram shows that any

pressure increase is accompanied by the reduction of the

average bed voidage.

The ECT system used in this work can only provide

information about the total bed voidage, consisting of the

dense phase voidage and bubbles volume held up within the

bed and averaged across one cross-sectional plane. There-

fore, based on these results alone, we cannot determine

whether the decrease in the bed voidage with the increase in

operating pressure, seen in Fig. 6, is caused by a reduction

range of operating pressures from 300 to 1900 kPa (lines are used to guide


in dense phase voidage or volume of bubbles in the bed, or

both.

However, the bed collapse experiments (see below)

demonstrate that for FCC catalyst the dense phase voidage

increased with the increase in pressure. The decrease in the

bubbling bed voidage or expansion with increase in pres-

sure, found in this work from the ECT dynamic measure-

ments, can therefore be explained by the reduction in

volume of bubbles held up within the bed at higher

operating pressures.

This might be in line with the observations of other

researchers [24,59,71–75], who reported that fluidization

became smoother at higher pressures and attributed that to

smaller bubbles in Geldart A powders at increased pressure.

However, in the present work, because of the ECT limi-

tations in image generating, no conclusion on the volume or

size of individual bubbles could be reached.

The results of the ECT experiments with the Geldart B

silica sand are different from those for the FCC catalyst. As

can be seen in Fig. 7, a considerable increase in the bed

voidage with increasing gas velocity takes place at each

operating pressure. This is expected as more gas is put into

the fluidized bed as bubbles. In comparison to Geldart A

powders, the interparticle forces in Geldart B materials are

negligible compared with the hydrodynamic forces acting in

the fluidized bed.

Fig. 7 also shows that increase in operating pressure

produces an increase in average bed voidage, consistent

with the trends reported for Geldart B powders by other

workers [8,9].

However, some unexplained changes in bed expansion

at higher pressures were observed by Olowson and Alm-

stedt [76] and Llop et al. [13,77], who fluidized silica

sand and found that the bed expansion strongly increased

with pressure up to a maximum at approximately 800 kPa

and then stayed constant or even slightly decreased

Fig. 7. Variation in average bed voidage with gas velocity for silica sand

thereafter at pressures up to 1600 kPa. Although in the

present work, no obvious decrease in total bed voidage

was observed at higher operating pressures, it was found

that the pressure influence was stronger at lower pressures

(see Fig. 7).

Olowson and Almstedt [78] found in their fluidization

work with Geldart B sand that the bubble size is deter-

mined by a complex balance between splitting and coa-

lescence, and an increase in pressure may cause either

increase or decrease in bubble size, depending on the

location in the bed, gas velocity and the pressure level.

In a comprehensive summary and analysis of previous

research on bubble size under pressurized conditions,

applicable for the pressurized fluidized bed combustion,

Cai et al. [79] reached the conclusion that in fluidized beds

of coarse materials the bubble size decreases with increas-

ing pressure except when gas velocity is low. At low gas

velocities, there is a dual effect of pressure on bubble size,

i.e. there is an initial increase in bubble size in the pressure

range less than 1000 kPa and then a size decrease with

further increase in pressure. The bed voidage results,

obtained from the dynamic ECT data in this study, are

therefore in accordance with the views of Olowson and

Almstedt [78] and Cai et al. [79].

3.5.2. Average absolute deviation

In bubbling fluidized beds the main source of signal

(pressure or voidage) fluctuations originates from the for-

mation of bubbles. The intensity of the signal can be

measured and compared in a statistical manner by using

the average absolute deviation. Therefore it was expected

that the average absolute deviation in the bed voidage would

steadily increase with increasing gas velocity when more

and more gas is put into the bed as bubbles. This was indeed

the case for both Geldart A and Geldart B materials, as can

be seen in Figs. 8 and 9.

over a range of operating pressures from atmospheric to 1700 kPa.

Fig. 8. Variation in average absolute deviation in bed voidage with gas velocity for FCC catalyst over a range of operating pressures from 300 to 1900 kPa

(lines are used to guide the eye).


For the Geldart A FCC catalyst (Fig. 8) average absolute

deviations of bed voidage was found to increase steadily

with gas velocity, as expected, and decrease with increasing

operating pressure—this effect being stronger at lower

pressure, as was found for average bed voidage. The

decrease in average absolute deviations of bed voidage with

increasing operating pressure is consistent with smaller

bubble size at higher pressures.

For the Geldart B silica sand, as can be seen in Fig. 9,

increasing operating pressure from ambient to 2100 kPa

has practically no effect on the average absolute deviation

in bed voidage and bubbling behavior of the bed. Once

again, the fluidized bed remains in the bubbling state

within the range of gas velocities studied, as can be seen

Fig. 9. Variation in average absolute deviation in bed voidage with gas velocity

2100 kPa.

from the increase in the average absolute deviation with

increasing gas velocity.

3.5.3. Average cycle frequency

The dependence of the measured average cycle frequen-

cy on gas velocity and operating pressure for the Geldart A

FCC catalyst and the Geldart B silica sand is shown in Figs.

10 and 11, respectively.

For the Geldart A FCC catalyst the average cycle

frequency is found to be variable at low gas velocities,

but tend towards a steady value once the bed was freely

bubbling (Umf is approximately 0.003 m/s and Umb is

approximately 0.006 m/s). Similar large differences between

different pressure fluctuations measurements at velocities

for silica sand over a range of operating pressures from atmospheric to

Fig. 10. Variation in average cycle frequency with gas velocity for FCC catalyst over a range of operating pressures from 300 to 1900k Pa (lines are used to

guide the eye).


near to Umf and Umb were observed at ambient conditions

by van der Stappen et al. [80] and Daw and Halow [81].

It can be seen from Fig. 10 that average cycle frequency

decreases from 4.5 Hz at 300 kPa to 2.3 Hz at 1900 kPa. We

note that the influence of pressure on the average cycle

frequency is stronger at low pressures, as was the case for

average bed voidage and average absolute deviation in

voidage. According to Bai et al. [82], experimental results

presented by van der Stappen et al. [83] suggest that higher

cycle frequencies correspond to more complex dynamic

systems. Therefore, the ECT experimental results, presented

in Fig. 10, show that increasing the operating pressure leads

Fig. 11. Variation in average cycle frequency with gas velocity for silica sand over

to guide the eye).

to a less dynamic fluidized bed system in line with other

researchers’ observations of smoother fluidization for Gel-

dart A materials at high pressure, which is generally

attributed to smaller bubbles.

For the Geldart B silica sand, in the intermediate regime

near the minimum fluidization and at gas velocities below

approximately 0.04 m/s, the average cycle frequency is

very sensitive and the differences between different meas-

urements are quite large, as shown in Fig. 11. When the

bubbling bed is developed, the average cycle frequency of

voidage fluctuations is about constant at a value of about

3.2 Hz at all the operating pressures tested in the range

a range of operating pressures from atmospheric to 2100 kPa (lines are used


from atmospheric to 2100 kPa. Similarly, the constant

value of the average cycle frequency of pressure fluctua-

tions of about 3.3 Hz in a bubbling bed was reported by

van der Stappen et al. [80] who fluidized Geldart B

polystyrene particles at ambient conditions. From Fig.

11, it can be observed that in the fully developed bubbling

regime the average cycle frequency is practically indepen-

dent of pressure.

3.5.4. Deterministic chaos analysis

Recently it has been suggested that the irregular behavior

of the fluidized bed dynamics is due to the fact that the

fluidized bed is a chaotic non-linear system. According to

some researchers in this area [68,84], chaotic systems are

governed by non-linear interactions between the system

variables and all methods of non-linear chaos analysis are

based on the construction of an attractor of the system in

state-space. The attractor forms a fingerprint of the system

and reflects its hydrodynamic state. The most common

method to characterize the attractor is the evaluation of

the correlation dimension and the Kolmogorov entropy,

where the correlation dimension expresses the number of

degrees of freedom of the system and the Kolmogorov

entropy is a measure of the predictability of the system

and the sensitivity to the initial state.

The Kolmogorov entropy, K, is a number expressed in

bits/s and can be calculated from a time series of only one

characteristic variable of the system, solids volume fraction

in this case. Generally, the Kolmogorov entropy is small in

cases of more or less regular behavior, and large for very

irregular dynamic behavior such as fluctuations in turbulent

flow.

The average absolute deviation is a measure of the

characteristic length and the average cycle frequency is a

Fig. 12. Kolmogorov entropy as a function of gas velocity for FCC catalyst over a

the eye).

measure of the time scale in the time series, and both

invariants were used in the reconstruction of the attractor,

from which the Kolmogorov entropy was calculated. Since

for the Geldart B material studied, both invariants were

found to be practically independent of pressure, it is

expected that the Kolmogorov entropy for this material

would be also more or less unaffected by pressure.

In order to observe the dynamic processes in fluidized

beds and be able to apply fully a chaotic time series

analysis, Daw and Halow [81] recommend acquiring time

series fluidized bed pressure drop or voidage measurements

at an optimum sampling rate of 200 Hz. Schouten and van

den Bleek [85] have found that in order to sufficiently

accommodate the chaos analysis attractor in its state space,

the required sampling frequency should be at approximately

100 times the average cycle frequency. Based on the average

cycle frequency of 3.2 Hz for the bubbling bed of the

Geldart B sand, the required sampling frequency should be

more than 300 Hz.

However, the ECT system used in this work was not

capable of sampling frequency above 90 Hz. Taking into

consideration this limitation of the equipment, it was ac-

cepted that the chaos analysis of the experimental data

might not be complete. According to Kuehn et al. [86], a

number of points per cycle in the voidage time series from

the ECT measurements is rather low which may lead to an

over-estimation of the Kolmogorov entropy. The Kolmo-

gorov entropy calculated from the results of the ECT

experiments with the Geldart A material is presented in

Fig. 12.

As can be seen in Fig. 12, at minimum bubbling

(Umbg0.006 m/s), the Kolmogorov entropy seems to have

a minimum, generally less than 5 bits per second, which is

lower than in the freely bubbling state. This is in good

range of operating pressures from 300 to 1900 kPa (lines are used to guide


agreement with results of van der Stappen et al. [80], who

qualitatively explained this in the following way—at gas

velocities where the first bubbles appear, the bubbles are

independent of each other and rise through the bed with no

or minor interaction. This behavior is relatively simpler and

more periodic compared to the developed bubbling state,

where bubbles are influencing each other in a more complex

way.

In the freely bubbling state, the Kolmogorov entropy is

expected to settle at a certain value, which could be as high

as 32 at the operating pressure of 300 kPa or as low as 12 at

1900 kPa. However, because of the equipment limitations

previously explained, Fig. 12 shows possibly not the abso-

lute values but merely trends. Examination of Fig. 12 alone

does not yield any trends. However, we could reasonably

make the comparison with the results for average cycle

frequency (Fig. 10), where the trend established itself at gas

velocities beyond 0.03 m/s. The variation in Kolmogorov

entropy with operating pressure may therefore be revealed

by the final data point in each series—i.e. Kolmogorov

entropy decreases with increasing pressure.

When the Kolmogorov entropy decreases, the fluidized

bed hydrodynamics become more predictable. Zijerveld et

al. [68] explain this in terms of an increase in bubble size at

the bed surface. This is in agreement with Schouten et al.

[87], who related the Kolmogorov entropy K to the bubble

size db as follows:

K ¼ 1:5ðU � Umf ÞDd2b

ð3Þ

According to this equation, the Kolmogorov entropy

decreases with an increase in bubble size. Based on this

and the trends presented for the Geldart A material in Fig.

12, it appears that the bubble size increases with increasing

pressure. However, this is contradictory to the previous

research and the results of this study presented in Fig. 6.

As previously mentioned, generally it has been reported

that fluidization becomes smoother with high pressure, which

has been attributed to smaller bubbles at increased pressure.

Many workers found that increasing the operating pressure

causes bubble size to decrease in Geldart A materials

[24,58,71–75]. A possible explanation might be that Eq.

(3) was developed based on the pressure fluctuations

obtained during the experiments with coarse Geldart B

polyethylene particles (0.56 mm) and silica sand (0.4 mm)

carried out at ambient conditions. As previouslymentioned, it

was found that bubbling behavior in fluidized beds of coarse

Geldart B materials differs noticeably from that in fluidized

beds of Geldart A powders even at ambient conditions.

Even with the Geldart B materials, there are some

apparently conflicting reports coming from the same school.

For example, Zijerveld et al. [68] found that in the bubbling

bed the Kolmogorov entropy decreases with superficial gas

velocity and decreases with an increase in bubble size.

Earlier van der Stappen et al. [80] found that in a freely

bubbling bed the Kolmogorov entropy settles at a value of

about 17 bits per second. In another related study [86],

higher gas velocity and more turbulent hydrodynamics was

found to cause an increase of the Kolmogorov entropy,

though bubble size in Geldart B powders is known to

increase with gas velocity.

Obviously, deterministic chaos theory can provide nu-

merical and graphical tools that can quantify and visualize

the fluidized bed hydrodynamics, however, it is still in its

early days and much more work has to be done. It is

necessary to verify further the correlations and incorporate

other particle systems. So far, it seems that the chaos

analysis method has been mainly applied to scale-up of

the dynamic behavior of the fluidized bubbling reactors and

fluidization regime transitions in the circulating fluidized

beds [68,87].

3.6. Bed collapse tests

Bed collapse experimental results had been obtained in a

form of video recording, from which the plots of bed height

decrease with time were constructed. The results of the

second intermediate stage when the bed surface falls at a

constant rate of collapse velocity Uc, until the bed height

approaches a certain critical height Hc, are plotted in Fig. 13.

It should be noted that the error bars for the curve at

atmospheric pressure are based on the total of 20 experi-

ments, five for each of four initial fluidization gas velocities.

Similar collapse curves for a FCC catalyst at various

operating pressures were observed by Foscolo et al. [49].

The straight line portions of the bed collapse curves were

extrapolated back to time zero and it was assumed that the

intercept with the bed height axis gives the height which the

dense phase would occupy in the bubbling bed.

Abrahamsen and Geldart [45] correlated the dense phase

expansion (Eq. (4)) which includes the gas density effect.

Hd

Hmf

¼2:54q0:016

g l0:066expð0:09ÞFd0:1p g0:118ðqp � qgÞ0:118H0:043

mf

ð4Þ

According to this correlation, the dense phase expansion

should slightly increase with increasing the operating pres-

sure. A comparison between the dense phase expansion

determined experimentally from the bed collapse tests at

various operating pressures and that predicted from the

correlation by Abrahamsen and Geldart [45] is presented

in Fig. 14. For comparison with the maximum non-bubbling

bed expansion for this material shown in Fig. 5, the scale of

both graphs is the same.

As can be seen in Fig. 14, the dense phase expansion

linearly increases with increasing pressure within the stud-

ied range 101–1600 kPa. Compared to the predictions of

the correlation by Abrahamsen and Geldart [45], the exper-

imental results show a slightly higher rate of increase, and

therefore, a stronger influence of operating pressure than the

factor qg0.016 included in the correlation.

Fig. 13. Bed surface collapse curves for FCC catalyst at various operating absolute pressures.


Similar results were observed by other workers [24,26].

As noted by Barreto et al. [88], the extensive experiments of

Abrahamsen and Geldart were carried out with air at

atmospheric pressure and the effect of gas density arose as

a consequence of the correlation method. According to the

correlation, it was expected that a pressure increase from

atmospheric to 1500 kPa should cause only 4.5% increase in

the dense phase height for a FCC catalyst fluidized with

nitrogen [26]. However, their experimental results demon-

strated a much higher 21% increase. Weimer and Quarderer

[24] also found that the correlation proposed by Abraham-

sen and Geldart [45] largely underestimated the effect of

pressure on the dense phase voidage for a Geldart A powder.

Fig. 14. Variation of the dense phase expansion with pressure for FCC catalyst. C

Abrahamsen and Geldart [45] correlation.

Bed collapse experiments are primarily intended to

obtain a direct value of the dense phase voidage. Once the

initial heights of the dense phase of the bed had been

determined from the collapse curves obtained at various

operating pressures (Fig. 13), the experimental values of the

dense phase voidage ed were calculated. The experimental

values of all the voidages—dense phase voidage, settled bed

voidage, minimum bubbling voidage and minimum fluid-

ization voidage—can now be compared (Fig. 15).

As can be seen in Fig. 15, within the studied pressure

range, the dense phase voidage was found to be higher

than the voidage at minimum fluidization and to increase

steadily with increasing pressure for a Geldart A powder.

omparison of experimental values (clear squares) with predictions from the

Fig. 15. Variation with pressure of dense phase voidage (ed), settled bed voidage (es), minimum fluidization voidage (emf) and minimum bubbling voidage (emb)

for FCC catalyst.


This is in line with the observations of other researchers

[24,26,49,58,71].

4. Conclusions

A comprehensive experimental study of the influence of

pressure on fluidization phenomena, such as minimum

fluidization and minimum bubbling velocities, bed expan-

sion and voidage was carried out in a bubbling fluidized bed

over the range of operating absolute pressures 101–2100

kPa with Geldart A and B bed materials.

As expected, it was found that the minimum fluidization

velocity decreased slightly with increasing pressure for a

Geldart B material and was independent of pressure for a

Geldart A material. In the pressure range studied in this

work, bed voidage at minimum fluidization was found to be

practically independent of pressure for both Geldart A and

Geldart B materials.

Although it has been implied that with pressure increase

Geldart B solids could demonstrate Geldart A type behavior,

in this work, we observed no shift in behavior over the

pressure range studied.

For the Geldart A powder, the minimum bubbling

velocity and the maximum bed expansion ratio were found

to be practically unaffected by increasing pressure. Predic-

tions using existing correlations for the minimum bubbling

velocity and the maximum bed expansion ratio were not in

agreement with the experimental values. In line with obser-

vations of other researchers, bed voidage at minimum

bubbling was found to increase slightly within the studied

pressure range.

Electrical capacitance tomography was used to provide a

time series analysis of fluctuations of the average solids

volume fraction, from which a quantitative description of

the bubbling bed dynamics as a function of operating

pressure was obtained. For Geldart A catalyst, it was found

that average bed voidage decreased with increasing operat-

ing pressure. For Geldart B sand, however, the opposite

trend was observed.

The intensity of the voidage fluctuations caused by the

formation of bubbles was measured and compared by using

the average absolute deviation. For Geldart A material,

increase in operating pressure gave rise to a decrease in

the average absolute deviation in bed voidage, suggesting

more uniform bubbling at elevated pressure. For Geldart B

material, increasing pressure up to 2100 kPa had no effect

on the average absolute deviation in voidage fluctuations.

For Geldart A material, the average cycle frequency

decreased with increasing operating pressure. This sug-

gested that increasing the operating pressure lead to a less

dynamic fluidized bed system in line with other researchers’

observations of smoother fluidization for Geldart A materi-

als at high pressure. For Geldart B solids, the average cycle

frequency was found to be independent of pressure.

Bed collapse tests were carried out at various pressures

up to 1600 kPa, using Geldart A solids. The dense phase

voidage values derived from these experiments were found

to be higher than the values of voidage at minimum

fluidization and were found to increase steadily with in-

creasing pressure.

References

[1] M.I. Fridland, The effect of pressure on productivity and on entrain-

ment of particles in fluidized bed equipment, Int. Chem. Eng. 3 (4)

(1963) 519–522.

[2] J.S.M. Botterill, Fluidized bed behaviour at high temperatures and


pressures, in: L.K. Doraiswamy, A.S. Mujumdar (Eds.), Transport in

Fluidized Particle Systems, Elsevier, Amsterdam, 1989, pp. 33–70.

[3] T.M. Knowlton, Pressure and temperature effects in fluid–particle

systems, in: W.C. Yang (Ed.), Fluidization Solids Handling and Pro-

cessing: Industrial Applications, Noyes Publications, Westwood,

1999, pp. 111–152.

[4] I. Sidorenko, M.J. Rhodes, Pressure effects on gas–solid fluidized

bed behavior, Int. J. Chem. React. Eng. 1 (2003) R5.

[5] J.G. Yates, Review article number 49. Effects of temperature and

pressure on gas–solid fluidization, Chem. Eng. Sci. 51 (2) (1996)

167–205.

[6] V.A. Borodulya, V.L. Ganzha, V.I. Kovensky, Gidrodynamika i

teploobmen v psevdoozhizhenom sloye pod davleniem (in Russian),

Nauka i Technika, Minsk, 1982.

[7] R. Bouratoua, Y. Molodtsof, A. Koniuta, ‘‘Hydrodynamic character-

istics of a pressurized fluidized bed’’. Paper presented at 12th Inter-

national Conference on Fluidized Bed Combustion, La Jolla, 1993.

[8] S. Chiba, J. Kawabata, T. Chiba, Characteristics of pressurized gas-

fluidized beds, in: N.P. Cheremisinoff (Ed.), Encyclopedia of Fluid

Mechanics, Gulf Publishing, Houston, 1986, pp. 929–945.

[9] D.C. Chitester, R.M. Kornosky, L.S. Fan, J.P. Danko, Characteristics of

fluidization at high pressure, Chem. Eng. Sci. 39 (2) (1984) 253–261.

[10] M.A. Gilbertson, D.J. Cheesman, J.G. Yates, Observations and meas-

urements of isolated bubbles in a pressurized gas-fluidized bed, in:

L.S. Fan, T.M. Knowlton (Eds.), Fluidization IX, Engineering Foun-

dation, New York, 1998, pp. 61–68.

[11] D.F. King, D. Harrison, The dense phase of a fluidized bed at elevated

pressures, Trans. Inst. Chem. Eng., (1) (1982) 26–30.

[12] T.M. Knowlton, High-pressure fluidization characteristics of several

particulate solids, primarily coal and coal-derived materials, AIChE

Symp. Ser. 73 (161) (1977) 22–28.

[13] M.F. Llop, J. Casal, J. Arnaldos, Incipient fluidization and expansion

in fluidized beds operated at high pressure and temperature, in: J.-F.

Large, C. Laguerie (Eds.), Fluidization VIII, Engineering Foundation,

New York, 1995, pp. 131–138.

[14] A. Marzocchella, P. Salatino, Fluidization of solids with CO2 at pres-

sures from ambient to supercritical, AIChE J. 46 (5) (2000) 901–910.

[15] M. Nakamura, Y. Hamada, S. Toyama, A.E. Fouda, C.E. Capes, An

experimental investigation of minimum fluidization velocity at ele-

vated temperatures and pressures, Can. J. Chem. Eng. 63 (1) (1985)

8–13.

[16] P.A. Olowson, A.E. Almstedt, Influence of pressure on the minimum

fluidization velocity, Chem. Eng. Sci. 46 (2) (1991) 637–640.

[17] S.C. Saxena, G.J. Vogel, The measurement of incipient fluidization

velocities in a bed of coarse dolomite at temperature and pressure,

Trans. Inst. Chem. Eng. 55 (3) (1977) 184–189.

[18] L.E.L. Sobreiro, J.L.F. Monteiro, The effect of pressure on fluidized

bed behavior, Powder Technol. 33 (1982) 95–100.

[19] C. Vogt, R. Schreiber, J. Werther, G. Brunner, Fluidization at super-

critical fluid conditions, in: M. Kwauk, J. Li, W.C. Yang (Eds.),

Fluidization X, United Engineering Foundation, New York, 2001,

pp. 117–124.

[20] A.R. Abrahamsen, D. Geldart, Behavior of gas-fluidized beds of fine

powders: part I. Homogeneous expansion, Powder Technol. 26 (1980)

35–46.

[21] K.V. Jacob, A.W. Weimer, High-pressure particulate expansion and

minimum bubbling of fine carbon powders, AIChE J. 33 (10) (1987)

1698–1706.

[22] A.K. Bin, Minimum fluidization velocity at elevated temperatures and

pressures, Can. J. Chem. Eng. 64 (5) (1986) 854–857.

[23] W.C. Yang, D.C. Chitester, R.M. Kornosky, D.L. Keairns, A general-

ized methodology for estimating minimum fluidization velocity at el-

evated pressure and temperature, AIChE J. 31 (7) (1985) 1086–1092.

[24] A.W. Weimer, G.J. Quarderer, On dense phase voidage and bubble

size in high pressure fluidized beds of fine powders, AIChE J. 31 (6)

(1985) 1019–1028.

[25] P.N. Rowe, The dense phase voidage of fine powders fluidised by gas

and its variation with temperature, pressure and particle size, in: J.R.

Grace, L.W. Shemilt, M.A. Bergougnou (Eds.), Fluidization VI, En-

gineering Foundation, New York, 1989, pp. 195–202.

[26] H.W. Piepers, E.J.E. Cottaar, A.H.M. Verkooijen, K. Rietema, Effects

of pressure and type of gas on particle–particle interaction and the

consequences for gas–solid fluidization behavior, Powder Technol.

37 (1984) 55–70.

[27] C. Vogt, R. Schreiber, J. Werther, G. Brunner, ‘‘The expansion be-

havior of fluidized beds at supercritical fluid conditions’’. Paper pre-

sented at World Congress on Particle Technology 4, Sydney, 2002.

[28] D. Geldart, Single particles, fixed and quiescent beds, in: D. Gel-

dart (Ed.), Gas Fluidization Technology, Wiley, Chichester, 1986,

pp. 11–32.

[29] J.G. Yates, Experimental observations of voidage in gas fluidized

beds, in: J. Chaouki, F. Larachi, M.P. Duducovic (Eds.), Non-Inva-

sive Monitoring of Multiphase Flows, Elsevier, Amsterdam, 1997,

pp. 141–160.

[30] D. Geldart, H.Y. Xie, The collapse test as a means for charac-

terizing Group A powders (preprints), in: J.-F. Large, C. Laguerie

(Eds.), Fluidization VIII, Engineering Foundation, Tours, 1995,

pp. 263–270.

[31] J.R. Grace, Agricola aground: characterization and interpretation of

fluidization phenomena, AIChE Symp. Ser. 88 (289) (1992) 1–16.

[32] K. Rietema, The Dynamics of Fine Powders, Elsevier, London, 1991.

[33] M. Louge, Experimental techniques, in: J.R. Grace, A.A. Avidan,

T.M. Knowlton (Eds.), Circulating Fluidized Beds, Blackie, London,

1997, pp. 312–368.

[34] D. Geldart, Types of gas fluidization, Powder Technol. 7 (1973)

285–292.

[35] J.S. Halow, P. Nicoletti, Observations of fluidized bed coalescence

using capacitance imaging, Powder Technol. 69 (1992) 255–277.

[36] Y.T. Makkawi, P.C. Wright, ‘‘Application of process tomography as a

tool for better understanding of fluidization quality in a conventional

fluidized bed’’. Paper presented at 2nd World Congress on Industrial

Process Tomography, Hannover, 2001.

[37] S.J. Wang, T. Dyakowski, C.G. Xie, R.A. Williams, M.S. Beck, Real

time capacitance imaging of bubble formation at the distributor of a

fluidized bed, Chem. Eng. J. 56 (3) (1995) 95–100.

[38] I. Sidorenko, M.J. Rhodes, ‘‘The use of Electrical Capacitance Tomog-

raphy to study the influence of pressure on fluidized bed behaviour’’.

Paper presented at 6th World Congress of Chemical Engineering,

Melbourne, 2001.

[39] L. Svarovsky, Powder Testing Guide: Methods of Measuring the

Physical Properties of Bulk Powders, Elsevier, London, 1987.

[40] Engineering design rules for ECT sensors (Wilmslow, Process To-

mography), 2001a.

[41] I. Sidorenko, Pressure Effects on Fluidized Bed Behaviour, PhD the-

sis, Monash University, Clayton, 2003.

[42] J.V. Fletcher, M.D. Deo, F.V. Hanson, Fluidization of a multi-sized

Group B sand at reduced pressure, Powder Technol. 76 (1993)

141–147.

[43] P. Harriott, S. Simone, Fluidizing fine powders, in: N.P. Cheremisinoff,

R. Gupta (Eds.), Handbook of Fluids in Motion, Science, Ann Arbor,

1983, pp. 653–663.

[44] D. Geldart, A.C.-Y. Wong, Fluidization of powders showing degrees

of cohesiveness: II. Experiments on rates of de-aeration, Chem. Eng.

Sci. 40 (4) (1985) 653–661.

[45] A.R. Abrahamsen, D. Geldart, Behavior of gas-fluidized beds of fine

powders: part II. Voidage of the dense phase in bubbling beds, Pow-

der Technol. 26 (1980) 47–55.

[46] C.Y. Wen, Y.H. Yu, A generalized method for predicting the minimum

fluidization velocity, AIChE J. 12 (3) (1966) 610–612.

[47] S. Ergun, Fluid flow through packed columns, Chem. Eng. Prog. 48

(1952) 89–94.

[48] J. Werther, Stromungsmechanische grundlagen der wirbelschichttech-

nik (in German), Chem. Ing. Tech. 49 (3) (1977) 193–202.

[49] P.U. Foscolo, A. Germana, R. Di Felice, L. De Luca, L.G. Gibilaro,


An experimental study of the expansion characteristics of fluidized

beds of fine catalysts under pressure, in: J.R. Grace, L.W. Shemilt,

M.A. Bergougnou (Eds.), Fluidization VI, Engineering Foundation,

New York, 1989, pp. 187–194.

[50] P.N. Rowe, P.U. Foscolo, A.C. Hoffmann, J.G. Yates, Fine powders

fluidized at low velocity at pressures up to 20 bar with gases of

different viscosity, Chem. Eng. Sci. 37 (7) (1982) 1115–1117.

[51] M. Sciazko, J. Bandrowski, Effect of pressure on the minimum bub-

bling velocity of polydisperse materials, Chem. Eng. Sci. 40 (10)

(1985) 1861–1869.

[52] M. Sciazko, J. Bandrowski, Effect of pressure on the minimum bub-

bling velocity of polydisperse materials. Reply to comments, Chem.

Eng. Sci. 42 (2) (1987) 388.

[53] T. Varadi, J.R. Grace, High pressure fluidization in a two-dimensional

bed, in: J.F. Davidson, D.L. Keairns (Eds.), Fluidization, Cambridge

Univ. Press, Cambridge, 1978, pp. 55–58.

[54] J. Baeyens, D. Geldart, Predictive calculations of flow parameters in

gas fluidized beds and fluidization behavior of various powders, in:

H. Angelino, J.P. Couderc, H. Gibert, C. Laguerie (Eds.), Fluidization

and its Application, Societe de Chimie Industrielle, Toulouse, 1974,

pp. 263–273.

[55] M.E. Crowther, J.C. Whitehead, Fluidization of fine particles at ele-

vated pressure, in: J.F. Davidson, D.L. Keairns (Eds.), Fluidization,

Cambridge Univ. Press, Cambridge, 1978, pp. 65–70.

[56] J.F. Richardson, Incipient fluidization and particulate systems, in:

J.F. Davidson, D. Harrison (Eds.), Fluidisation, Academic Press, Lon-

don, 1971, pp. 25–64.

[57] K. Godard, J.F. Richardson, The behaviour of bubble-free fluidised

beds, IChemE Symp. Ser. 30 (1968) 126–135.

[58] J.R.F. Guedes de Carvalho, Dense phase expansion in fluidized beds

of fine particles. The effect of pressure on bubble stability, Chem.

Eng. Sci. 36 (2) (1981) 413–416.

[59] J.R.F. Guedes de Carvalho, D.F. King, D. Harrison, Fluidization of

fine particles under pressure, in: J.F. Davidson, D.L. Keairns (Eds.),

Fluidization, Cambridge Univ. Press, Cambridge, 1978, pp. 59–64.

[60] M. Byars, ‘‘Developments in Electrical Capacitance Tomography’’.

Paper presented at 2nd World Congress on Industrial Process Tomog-

raphy, Hannover, 2001.

[61] T. Dyakowski, L.F.C. Jeanmeure, A.J. Jaworski, Applications of elec-

trical tomography for gas–solids and liquid–solids flows—a review,

Powder Technol. 112 (2000) 174–192.

[62] Y.T. Makkawi, P.C. Wright, Optimization of experiment span and data

acquisition rate for reliable electrical capacitance tomography mea-

surement in fluidization studies—a case study, Meas. Sci. Technol. 13

(2002) 1831–1841.

[63] T. Dyakowski, personal communication, Clayton, 2002.

[64] O. Isaksen, A review of reconstruction techniques for capacitance

tomography, Meas. Sci. Technol. 7 (3) (1996) 325–337.

[65] O. Isaksen, J.E. Nordtvedt, A new reconstruction algorithm for pro-

cess tomography, Meas. Sci. Technol. 4 (3) (1993) 1464–1475.

[66] Generation of ECT images from capacitance measurements (Wilm-

slow, Process Tomography) 2001b.

[67] J.C. Schouten, F. Takens, C.M. van den Bleek, Estimation of the di-

mension of a noisy attractor, Phys. Rev., E 50 (3) (1994) 1851–1861.

[68] R.C. Zijerveld, F. Johnsson, A. Marzocchella, J.C. Schouten, C.M.

van den Bleek, Fluidization regimes and transitions from fixed bed to

dilute transport flow, Powder Technol. 95 (1998) 185–204.

[69] D. Geldart, N. Harnby, A.C.-Y. Wong, Fluidization of cohesive pow-

ders, Powder Technol. 37 (1984) 25–37.

[70] D. Geldart, A.L. Radtke, The effect of particle properties on the

behavior of equilibrium cracking catalysts in standpipe flow, Powder

Technol. 47 (1986) 157–165.

[71] G.F. Barreto, J.G. Yates, P.N. Rowe, The effect of pressure on the

flow of gas in fluidized beds of fine particles, Chem. Eng. Sci. 38 (12)

(1983) 1935–1945.

[72] I.H. Chan, C. Sishtla, T.M. Knowlton, The effect of pressure on

bubble parameters in gas-fluidized beds, Powder Technol. 53 (1987)

217–235.

[73] D.F. King, D. Harrison, The bubble phase in high-pressure fluidized

beds, in: J.R. Grace, J.M. Matsen (Eds.), Fluidization, Plenum, New

York, 1980, pp. 101–108.

[74] P.N. Rowe, H.J. MacGillivray, A preliminary X-ray study of the effect

of pressure on a bubbling gas fluidised bed, in: Fluidised Combustion:

Systems and Applications, vol. IV-1, Institute of Energy, London,

1980, pp. 1–9.

[75] M.P. Subzwari, R. Clift, D.L. Pyle, Bubbling behavior of fluidized

beds at elevated pressures, in: J.F. Davidson, D.L. Keairns (Eds.),

Fluidization, Cambridge Univ. Press, Cambridge, 1978, pp. 50–54.

[76] P.A. Olowson, A.E. Almstedt, Influence of pressure and fluidization

velocity on the bubble behavior and gas flow distribution in a fluid-

ized bed, Chem. Eng. Sci. 45 (7) (1990) 1733–1741.

[77] M.F. Llop, J. Casal, J. Arnaldos, Expansion of gas–solid fluidized

beds at pressure and high temperature, Powder Technol. 107 (2000)

212–225.

[78] P.A. Olowson, A.E. Almstedt, Hydrodynamics of a bubbling fluidized

bed: influence of pressure and fluidization velocity in terms of drag

force, Chem. Eng. Sci. 47 (2) (1992) 357–366.

[79] P. Cai, M. Schiavetti, G. De Michele, G.C. Grazzini, M. Miccio,

Quantitative estimation of bubble size in PFBC, Powder Technol.

80 (1994) 99–109.

[80] M.L.M. van der Stappen, J.C. Schouten, C.M. van den Bleek, Appli-

cation of deterministic chaos theory in understanding the fluid dy-

namic behavior of gas– solids fluidization, AIChE Symp. Ser. 89

(296) (1993) 91–102.

[81] C.S. Daw, J.S. Halow, Evaluation and control of fluidization quality

through chaotic time series analysis of pressure-drop measurements,

AIChE Symp. Ser. 89 (296) (1993) 103–122.

[82] D. Bai, A.S. Issangya, J.R. Grace, Characteristics of gas-fluidized

beds in different flow regimes, Ind. Eng. Chem. Res. 38 (3) (1999)

803–811.

[83] M.L.M. van der Stappen, J.C. Schouten, C.M. van den Bleek, ‘‘De-

terministic chaos analysis of the dynamical behaviour of slugging and

bubbling fluidized beds’’. Paper presented at 12th International Con-

ference on Fluidized Bed Combustion, La Jolla, 1993.

[84] F. Johnsson, R.C. Zijerveld, J.C. Schouten, C.M. van den Bleek,

B. Leckner, Characterization of fluidization regimes by time-series

analysis of pressure fluctuations, Int. J. Multiph. Flow 26 (4)

(2000) 663–715.

[85] J.C. Schouten, C.M. van den Bleek, Monitoring the quality of fluid-

ization using the short-term predictability of pressure fluctuations,

AIChE J. 44 (1) (1998) 48–60.

[86] F.T. Kuehn, J.C. Schouten, R.F. Mudde, C.M. van den Bleek, B.

Scarlett, Analysis of chaos in fluidization using electrical capaci-

tance tomography, Meas. Sci. Technol. 7 (3) (1996) 361–368.

[87] J.C. Schouten, M.L.M. van der Stappen, C.M. van den Bleek, Scale-

up of chaotic fluidized bed hydrodynamics, Chem. Eng. Sci. 51 (10)

(1996) 1991–2000.

[88] G.F. Barreto, G.D. Mazza, J.G. Yates, The significance of bed col-

lapse experiments in the characterization of fluidized beds of fine

powders, Chem. Eng. Sci. 43 (11) (1988) 3037–3047.