Application of Ultrasonic Technique in Multiphase

Maytinee Vatanakul,Ying Zheng,*and Michel Couturier

Department of Chemical Engineering,University of New Brunswick,15Dineen Drive,P.O.Box4400,

Fredericton,New Brunswick E3B5A3,Canada

Ultrasonic technique was used as a tool for phase holdup measurement in multiphase systems.

A new statistical approach using fluctuations of the ultrasound signals was applied to

simultaneously detect the dispersed phase holdups in gas-liquid-solid flows.These phase

holdups,measured by two methods s the ultrasonic technique and the applications of pressure

transducers and conductivity probes s agreed well.In the system of a high liquid velocity(above

14cm/s),the ultrasonic technique still provided reliable results.The local phase holdups and

their radial distribution were measured in gas-liquid-solid three-phase circulating fluidized

beds at two different elevations.A nonuniform radial distribution and a uniform axial distribution

of gas and solid phases were observed.The nonuniform radial distribution of solid holdup did

not depend on the solid circulating rate.The radial distributions of both phase holdups became

uniform with increasing liquid velocity.

Introduction

Gas-liquid-solid fluidization is a complex system that consists of dispersion of gas bubbles and solid particles in a continuous liquid phase.These systems are widely used for chemical,physical,petrochemical, and biochemical processes.1Gas-liquid-solid three-phase circulating fluidized beds(GLSCFBs)are a special type of three-phase fluidized bed with a continu-ous outer particle circulation.For multiphase industrial applications,GLSCFBs provide many efficient proper-ties,such as good phase contact,minimum dead zone, excellent heat and mass transfer characteristics,and high operational flexibility.GLSCFBs are suitable for the reactions involving light and/or small solid particles. Compared to the expanded three-phase fluidized bed, these reactors are commonly operated at high gas and liquid velocities,resulting in higher capacities. Thus,it is important to develop and achieve an efficient design of the fluidization systems.The hydro-dynamic behaviors and phase holdup distributions need to be studied and well-understood.Several measure-ment techniques have been applied for phase holdups in multiphase systems,such as static pressure,direct sampling,shutter,optical-probe,and electrical-probe techniques.However,these methods involve some dif-ficulties in obtaining reliable data.For example,the static pressure method is ineffective in the system where solids and liquids have similar densities.It is difficult to obtain reliable results with the direct optical-probe and the electrical-probe techniques.The direct sampling method is not accurate in measuring solid particles in the slurry.2

Ultrasonic measurement offers many advantages, such as high accuracy and rapidity.It is also suitable for optically opaque systems and potentially applicable to high temperature and pressure conditions.Further-more,the ultrasonic technique is safer and simpler to use in comparison with existing techniques such as γ-ray and X-ray.There are several researchers develop-ing this method to study the hydrodynamic character-istics of multiphase fluidization systems.However,the disadvantage of this technique is temperature sensi-tivity.3-5

For this ultrasonic system,the sound beams are transmitted through the studied volume.The acoustic properties of the transmitted ultrasound,such as veloc-ity and amplitude,vary as the beam comes across different media along its path.In multiphase flows,a fluid contains dispersed inhomogeneities,which are the solid particles and/or gas bubbles.When an acoustic wave strikes the boundary between two different media and the acoustic impedances of the two media differ, some acoustic energy is reflected and some is transmit-ted.The reflected wave travels back through the inci-dent medium at the original sound velocity.The trans-mitted acoustic wave continues to move through the new medium at the sound velocity of the new medium.In addition,while the ultrasound travels through different media,it is partially scattered and absorbed.This accordingly leads to the decrease of the amplitude of the sound wave,which is called attenuation.2,6-7An accurate measurement of the variations in sound speed and amplitude can therefore be used to detect different media or discontinuities in a system.

The ultrasonic technique has been widely applied in two-phase systems.In a liquid-liquid two-phase sys-tem,Smith8applied the difference of the transmitted acoustic velocities in two liquids to simultaneously measure the volume fraction and the average drop size of one liquid dispersed in another one.Havlicek and Sovava9applied the ultrasonic technique to monitor instantly the local phase fraction by measuring the velocity of sound in the liquid dispersion.Their ap-proach,however,required mounting the probes through holes in the vessel walls.Later,Tavlarides and Bonnet10 demonstrated the nonintrusive use of this technique, which was used for monitoring a local phase fraction in a multistaged pilot extractor and for extractor con-trol.11,12Afterward,Yi and Tavlarides13obtained the dispersed-phase holdups by using a simple volume average of sound velocity through the dispersion as compared to that through the liquid only.In liquid-solid systems,the acoustic properties can be used to

*To whom correspondence should be addressed.E-mail: yzheng@unb.ca.5681

Ind.Eng.Chem.Res.2004,43,5681-5691

10.1021/ie034184c CCC:$27.50©2004American Chemical Society

Published on Web05/27/2004determine particle size,concentration,and the mechan-ical properties of the constituents such as elastic modulus and rheology.Urick14and Hampton15investi-gated the systems of fine powders.Atkinson and Kytomaa16studied the system of bigger particles.More-over,the application of the ultrasonic technique was also found in gas-liquid systems to determine the properties of the gas bubbles such as the void fraction of the system,17specific interfacial area,18and bubble diam-eters.19

Attempts have been made to apply the ultrasonic measurement techniques to study the phase holdups in gas-liquid-solid three-phase fluidized beds as well as to correlate the transmission time and the attenuation of ultrasound with gas bubble and particle concentra-tions.2,20-22Uchida et al.2,20proposed an indirect method to measure the solid holdup by analyzing the shape and the phase lag or lead of the ultrasonic wave transmitted through the three-phase system.An accurate measure-ment of the variations in sound speed and amplitude could therefore be used to detect the different media or the presence of discontinuities in a system.The same method was applied by Maezawa et al.23to measure the longitudinal distribution of gas holdup.Warsito et al.24 used the transmission time difference,which is based on the difference between the velocity of an ultrasound signal in the dispersion and that in the pure liquid,to investigate the gas and solid holdups.This variation of sound speed as well as the difference of the attenuation was applied to correlate the solid volume fraction in the slurry reactor using nitrogen as a gas phase.4-6Later, Warsito et al.25applied the ultrasonic computed tomog-raphy to obtain the cross-sectional distributions of gas and solid holdups.

All the existing ultrasonic measurement methods have been developed on the basis of the phase difference

of acoustic waves,transit time variation,and change in amplitude of ultrasounds with certain simplifications and assumptions.The assumption that gas bubbles have no effect on sound velocity was commonly used.4,5,26 Lately,some researchers have shown that the speed of ultrasound was independent of gas holdups,due to the great distortion of ultrasound around bubbles.22,29This discrepancy challenged the approach of determining phase holdups in a multiphase system.Therefore,there is a need for a new approach to investigate the phase holdup in three-phase systems.

A new analysis approach was proposed in recent publications.28,29,31This analysis technique was used to determine the phase holdups for a system using500-µm glass beads as the solid phase.The method was based on the fact that the variations of signals are caused by the reflection and refraction when a sound beam encountered a boundary between two media with different characteristic impedances.Their observation showed that the fluctuations of sound waves,in terms of the standard deviations of transmission time and amplitude ratio,were highly correlated with gas and solid holdups.A new mathematical analysis of these fluctuations in two-phase flows can be employed to differentiate simultaneously the contributions of gas bubbles and solid particles in three-phase fluidized beds.30However,this approach has been applied only to a three-phase flow with500-mm glass beads as the solid phase.

In addition,most of these studies were performed with small solid particles(below1mm).Only Macchi et al.22presented results with1.3-mm glass beads.They found that the ultrasonic system did not provide a good measurable signal and that this technique was better suited for smaller sizes of particles.

The goal of this work is to assess the validity of the application of the ultrasound device and the new statistical approach for large particle sizes.The local dispersed phase holdups,the distribution of gas bubbles and solid particles,and the influences of operating conditions on the phase holdups were also investigated in the GLSCFB.

Experimental Setup

The experiments were performed in a fluidization loop,as shown in Figure1.This loop consists of two main Plexiglas columns,called riser and downer,re-spectively.The riser is7.6cm in diameter and2.0m in height.The downer is10.2cm in diameter.The liquid pumped from the reservoir was divided into two streams and then fed into the bed.The primary liquid stream entered at the base of the riser as a continuous fluid to carry gas bubbles and solid particles up along the riser column.The second liquid stream was the auxiliary liquid stream entering at the side of the downer column to push and to control the amount of solid particles to the riser via the L-valve.After the three phases moved concurrently upward to the top of the riser and entered the downer,gas bubbles left the top exit of the downer while solid particles were separated by a centrifugal filter.This centrifugal filter is a cylindrical tube5

cm Figure 1.Model gas-liquid-solid circulating fluidized bed column:(a)auxiliary liquid line,(b)primary liquid line,(c)air line,(d)gas distributor,(e)conductivity probes,(f)butterfly valve, (g)ultrasonic probes,(h)ultrasonic controller,(i)heat controller, (j)riser,(k)downer,(l)centrifugal filter,(m)heater,(n)flow meters,(o)cooling coil,(p)L-valve,(q)pressure transducers,(r) water reservoir

5682Ind.Eng.Chem.Res.,Vol.43,No.18,2004in diameter,which is located on the top of the downer. Solids were removed from the liquid stream by a radial centrifugal force exerted on the particles and were pushed downward in the downer.The butterfly valve was used for measuring the solid circulating rate,which is the mass flow rate of solids circulating within the fluidization loop.The solid circulating rate was deter-mined by measuring the height of solid particles for a given time period.This solid circulating rate and overall liquid flow rate could be controlled independently by regulating the flow ratio between the two liquid streams. In addition,the superficial liquid velocity mentioned in this work was the sum of the primary and auxiliary liquid flow rates.

Tap water was the continuous liquid phase.Glass beads with average diameters of433,700,and1300µm and a density of2500kg/m3were used as the dispersed solid phases.Oil-free air bubbles generated by a gas distributor were the dispersed gas phase.The gas distributor was a perforated stainless steel plate with small pores of35µm diameter,located close to the base of the riser.

A two-channel ultrasonic pulser-receiver system(model FUI2100,Fallon Ultrasonics Inc.)was used to measure the transmitted sound signals through the column in terms of the transmission time and the amplitude of the selected echoes within the gates.The system had a0.1-30MHz amplifier bandwidth and high-power HP mod-ules.The ultrasonic probes(13mm in diameter),a3 MHz transmitter,and a3MHz receiver were mounted along the riser column and operated in a pulse mode. The gain of sound wave was continuously adjusted in order to maintain the signal above the instrument threshold limit.The instantaneous ultrasonic signals were analyzed to obtain the standard deviation of sound speed and attenuation in order to determine the gas and solid holdups.

The horizontal space between the two ultrasonic probes was set at7or2cm apart.The local phase holdups were measured at two different axial positions of0.5and1.2m from the gas distributor.Besides,the transmitter and receiver were moved simultaneously at seven radial positions with the constant probe spacing of2cm.The local holdups are the mean of the phase holdups within the measuring gap.The water temper-ature is precisely maintained at25°C by a digital temperature controller during the experiments in order to achieve an accurate measurement of the ultrasonic system.

Moreover,the phase holdups in the liquid-solid and gas-liquid systems were also determined by differential pressure transducers.To validate the ultrasonic results in the three-phase flow system,the pressure transduc-ers and wall conductivity probes were used concurrently to determine the volume fractions of gas bubbles and solid particles.Since both gas and solid phases are nonconductive,the average value of conductivity data can be directly related to the liquid holdups.27Conduc-tivity probes were supplied with a1kHz ac current. Data from pressure transducers and the conductivity probes were taken at a data acquisition frequency of150 Hz for30s at the same time.During the actual test,at least two repeated sets of data were obtained for each operating condition to ensure accuracy.

Analysis of Ultrasonic Signals.The transmitted ultrasonic signals,the transmission time,and amplitude are recorded from a multiphase flow system.The amplitude ratio or the attenuation is determined from the variation of system sound amplitude compared with the amplitude in pure liquid system.Vatanakul et al.28,29studied the characteristics of the acoustic signals in the air/water and water/500-µm glass bead systems. From their work,it can be seen that the presence of dispersed gas and solid phases clearly had different effects on the fluctuation of the waveforms.Gas bubbles were apparently the significant factor in generating high fluctuations of sound signals,especially the transmis-sion time within the multiphase flows.Besides,the standard deviations of transmission time and amplitude ratio were well-functioned with the concentrations of gas bubbles and solid particles.In addition,the determina-tion of signals in three-phase flow showed that the fluctuations of the sound signal resulted from three factors,gases,solids,and the interaction between gases and solids.30

In this investigation,the statistical approach for the three-phase flow developed by Zheng and Zhang31was applied.The gas and solid holdups of three-phase systems can be statistically estimated through the relationship between phase holdups and the sequence of the signal’s fluctuation obtained in gas-liquid and liquid-solid systems.

This statistical approach is briefly presented.Only the signals of the transmission time determining the gas holdup are exhibited here in an example.A typical sequence of transmission time recorded in a three-phase flow is denoted as T gls.To differentiate a subset repre-senting the individual contributions of phase holdups from the standard deviation of T gls,a separation coef-ficient,δ,is defined

and

where T gls(t n)is the transmission time in the three-phase system recorded at time t n.T gls,max is the maxi-mum element of T gls.T h gls is the average T gls value. The sequence T gls is then arbitrarily partitioned by the defined separation coefficient,δ,into two subsets. The first subsets of the series are called the upper portion,defined as T U(δ).This subset consists of the data higher than or equal to T h glsδ.Under similar circumstances,the lower portion of T gls,presented as T L(δ),involves the elements less than or equal to T h gls/δ. An increase in the value ofδconsequently leads to a decrease in the standard deviation of the two subsets. Whenδreached its optimum value,presented asδ*, T U(δ*)of a gas-liquid-solid three-phase flow was to statistically represent the contribution of gas bubbles to the standard deviation of T gls.Moreover,this subset of T gls was also highly correlated with the sequence of ultrasonic signal recorded in a two-phase flow system that had the same gas holdup.

To achieveδ*the correlation coefficient based on canonical analysis,F,is used.This coefficient can be mathematically described as follows:

T

gls,max

/T h

gls

>δg1(1)

T

gls

)[T

gls

(t

1

),T

gls

(t

2

),T

gls

(t

3

),...,T

gls

(t

n

)]′(2) F)

Cov[L′T

U

(δ),M′T

gl

]

[Var(L′T

U

(δ))Var(M′T

gl

)]1/2

)

Cov(W,V)

[Var(W)Var(V)]1/2

(3)

Ind.Eng.Chem.Res.,Vol.43,No.18,20045683

It is noted that |F |e 1.L and M are the coefficient vectors.The absolute value of F is maximized to achieve the highest correlation between W and V .

Following the statistical procedure mentioned above,the canonical correlation coefficient between two sub-sets,F ,is consequently a function of δ.

Therefore,if the separation coefficient is adjusted to its optimal value,δ*,the maximum canonical correlation coefficient,F *,is also reached.From a statistical point of view,the T (δ*)is an appropriate substitute for T gl ,since both of the two sequences represent the flow having the same gas holdup and reflect the effect of dispersed gas phase on the ultrasonic signal.Finally,the gas holdup of a three-phase system can be estimated by inserting the standard deviation of T (δ*)in to the relationship of the standard deviation and gas holdup that is experimentally obtained in a gas -liquid flow.The fluctuation of the amplitude ratio of an ultrasonic signal can be analogously analyzed by this approach.An appropriate portion is extracted from the amplitude ratio sequence obtained in a three-phase system and,then,replaced by the amplitude ratio of liquid -solid flow having the same solid holdup.Finally,the solid holdup of a three-phase system can be estimated using the relationship between the standard deviation of the amplitude ratio and solid holdup observed in liquid -solid flows.

The statistical approach presented earlier provided a reliable analysis of gas and solid holdups in the 500-µm glass bead system.31The goal of this work is to apply the aforementioned method to determine the phase holdups in three-phase systems for three different sizes of glass beads (433µm,700µm,and 1.3mm).Results and Discussion

Characteristics of Sound Signals in Different Flow Systems.Experiments were carried out in four different systems:a pure liquid flow,a gas -liquid flow,a liquid -solid flow,and a gas -liquid -solid flow.To study the influences of dispersed phases on the char-acteristics of the ultrasonic signals,all the experiments

took place at 0.5m above the gas distributor.In a two-phase fluidized bed,the solids or gas bubbles were uniformly distributed.Glass beads were in a narrow size distribution.Air bubbles were uniform in size.The average bubble size,observed through a video camera,increased from 2to 6mm as the gas velocity increased.Figures 2and 3show the ultrasound signals,in terms of transmission time difference (∆T )and amplitude ratio measured in systems with gas -liquid,solid -liquid,gas -liquid -solid flows and liquid one-phase flows,respectively.The solid phase was 1.3-mm glass beads.∆T represents the difference between the transmission time of the ultrasound in a multiphase flow and that in a pure water system.The characteristics of acoustic signals will be discussed separately for each flow system:gas -liquid,liquid -solid,and gas -liquid -solid flows.

In the gas -liquid system,the waveform of the trans-mission time difference caused by gas bubbles demon-strated a greatly fluctuating distribution (Figure 2).It is because the impedance of air and water is very different.When the sound beam strikes at the interface of these two phases,the sound wave is reflected/reradiated.Thus,the traveling time of the ultrasound is longer,which gives a positive ∆T.The same charac-teristic was also reported in the time signal of a single bubble.28,29From Figure 3,it is noted that the system of air -water also introduced the highest fluctuation of amplitude ratio compared to the other water -glass bead and air -water -glass bead fluidized systems.The at-tenuation of ultrasound is exponentially proportional to the interfacial area of suspended inhomogenities as follows 18,32

and the absorption coefficient,R ,is generalized to

A and A 0are the amplitudes of the acoustic signals with and without the presence of dispersed phases,gas bubbles,and/or solid particles.νis the volumetric interfacial area,Z is the traveled distance of the incident sound wave,R is the scattering coefficient,k is

the

Figure 2.Transmission time difference signals recorded in the pure water,gas -liquid,liquid -solid,and gas -liquid -solid flows of 1.3-mm glass beads.

F)f (δ)(4)

A /A 0)exp(-R Z )

(5)

R)

(kd h s 2)(νZ 8θ

)

(6)

5684Ind.Eng.Chem.Res.,Vol.43,No.18,2004

wavenumber of ultrasonic wave,and d h s is the Sauter mean diameter of suspended objects.In the gas bubble -water system,the average size of bubbles is big (2-6mm).This results in a small value of the volumetric interfacial area,which in turn gives a small absorption coefficient (eq 5).This leads to a high amplitude ratio in the flow system (eq 6).

In the liquid -solid system,it was evident that transmission time slightly oscillated compared to that in gas -liquid flow.The partial penetration and the scattering of the acoustic beam through solid particles generated smaller fluctuations in the signals.For the waveform of amplitude ratio,the small variation gener-ated by solid particles was observed as well.It can be attributed to the small difference in impedance at the liquid -solid boundary.

Further observation was made in the three-phase system where both gases and solids were present.The variation of transmission time difference in three-phase flow was higher than those in two-phase flows (Figure 2).It can be explained that these variations of acoustic signals were created by discontinuous phases (gases and solids)and interaction between phases.Gas bubbles were clearly a dominant factor for the fluctuation of ultrasonic speed,which was further promoted by the interaction between gas bubbles and suspended solid particle in a three-phase system.

The observations of the waveform characteristics for 1.3-mm glass bead system are in agreement with the study of the system using 500-µm glass beads reported by Vatanakul et al.28,29In addition,in this work,the systems of glass beads with an average size of 433and 700µm also provided similar results.

The contributions of gas bubbles and solid particles to the fluctuation of ultrasonic signals are measured in two-phase fluidizations.It can be observed that the volume fractions of gas bubbles and solid particles are highly correlated with the standard deviations of the transmission time and amplitude ratio of the ultrasound (Figure 4a,b).

Figure 4a shows that the fluctuations of transmission time increase when concentrations of air bubbles or glass particles increase.The transmission time’s fluc-tuations in the gas -liquid system are much higher than that in the solid -liquid system.This finding demon-strates that gas bubbles are certainly the major factor that generates the fluctuations of the transmission time.In the liquid -solid system,the glass beads with average sizes of 433and 1300µm were used as the solid phase.It can be seen that the particle size affected the standard deviation of transmission time.The system with bigger particles generated a higher standard deviation of transmission time.For a given solid holdup,the system with smaller beads has a higher number of particles.This means that the sound wave travels through a more consistent medium and therefore the transmission time presents less fluctuation than in the case of bigger beads.

Figure 4b shows the standard deviation of the am-plitude ratio of the gas -liquid system and the liquid -solid systems using three different sizes of glass beads.The standard deviation of the amplitude ratio in water -air flow increased gradually as gas holdup increased.The system of water and glass beads was operated in two different regimes,which were the expanded bed and solid circulating bed regimes.The results from these two regimes can be fitted with the same curve.Apparently,liquid velocity has no influence on the ultrasonic signals.Therefore,only one calibration line was needed to describe the solid holdup in terms of the standard deviation of amplitude ratio,independent of the system regime.The standard deviation of the amplitude ratio exponentially decreased with increasing solid holdup.The relationship between the fluctuation of amplitude ratio and the solid concentration for three particle systems demonstrated a similar profile.However,for a given solid holdup,the standard deviation of amplitude ratio increased with increasing particle size.This can be explained by eqs 5and 6.The system with bigger beads has a smaller volumetric interfacial area,which leads to a smaller scattering coefficient and then to a higher amplitude ratio.This characteristic results in the higher fluctuation observed in a bigger particle

system.

Figure 3.Amplitude ratio signals recorded in the gas -liquid,liquid -solid,and gas -liquid -solid flows of 1.3-mm glass beads.

Ind.Eng.Chem.Res.,Vol.43,No.18,20045685

Comparison of Experimental Phase Holdups and Model Predictions.The accuracies of the pro-posed analysis method were evaluated in a gas -liquid -solid expanded fluidization system.The solid and gas holdups,determined by pressure transducer and a pair of conductivity probes,were considered as the experi-mental results,whereas the predicted phase holdups were the results obtained from the ultrasonic and the statistical approach.The distributions of solid particles and gas bubbles were assumed to be uniform under a wide range of operating conditions.The predicted phase holdups were compared with the experimental ones,as shown in parts a,b,and c of Figure 5for the systems of air,water,and glass beads with three different bead sizes of 433,700,and 1300µm,respectively.It can be seen that the results of gas and solid holdups estimated from these two techniques are in very good agreement for every particle size.Thus,it can be concluded that the statistical analysis of the fluctuation of an ultra-sound wave is very practical and accurate to predict values of gas and solid holdups in a three-phase fluid-ized bed.Moreover,the size of the solid particles has no influence on the accuracy of the new proposed procedure.This pair of 3MHz ultrasonic transducers is proved to be a reliable application for the three-phase systems of solid particles with diameters in the range of 433-1300µm,which is close to or larger than the wavelength of the sound wave (477µm)emitted from the transmitter.If the particle size is smaller than the acoustic wavelength,the sound wave may pass around and have no change to be reflected at the solid particle boundary.33Therefore,in a system of a smaller particle size,ultrasonic probes with a lower frequency are suggested.

Application of the Ultrasonic Technique in GLSCFB.The ultrasonic technique was set up in the riser of a GLSCFB to measure cross-sectional averaged solid and gas holdups simultaneously.The location of transmitter and receiver probes was at 0.5and 1.2m from the gas distributor.The distance between the two probes was set at 7cm apart.For the phase holdup’s measurement,the pressure transducers and conductiv-ity probes were also employed and located at about the same axial positions.Experiments with the fluidized bed,consisting of air,water,and 433-µm glass beads,were performed under circulating operating conditions,in which the liquid velocity is higher than the particle terminal velocity (6.3cm/s for 433-µm glass beads).The predicted phase holdups,measured by the ultra-sonic technique and the statistical approach,and the experimental phase holdups,determined by the pres-sure gradients and conductivity voltages,are reported in Table 1.At a low liquid velocity of 9.7cm/s with constant gas and solid circulating rate,the experimental and the predicted volume fraction of 433-µm glass beads and air bubbles were very similar.However,from Table 1,at a higher liquid velocity of 15.2cm/s,the experi-mental cross-sectional averaged axial solid holdups were smaller than the predicted ones for various gas velocities and solid circulating rates,since the experimental gas holdup was calculated from the fact that the summation of overall cross-sectional averaged phase holdup is equal to unity.Hence,the experimental value of gas holdup was expectably higher compared to the gas holdup obtained from the ultrasonic technique under the same operating conditions.Moreover,similar results were observed in the system with 1.3-mm glass beads as well.The difference between experimental and predicted phase holdups was obtained at the liquid velocity higher than 15cm/s,where 1.3-mm glass beads started to be circulated in the fluidization loop.The high liquid velocity may affect the disagreement of the phase holdup’s results between two methods.The accuracy of both measurement techniques applied for the circulating fluidized bed,operated under high liquid velocities,needed to be investigated.The testing experiments took place in a liquid -solid two-phase circulating fluidized bed.The pressure transducer,conductivity probes,and ultrasonic analysis system were applied to simulta-neously measure the cross-sectional averaged solid holdup at the same axial location in the riser.The two-phase fluidized bed was operated at a liquid velocity higher than 12cm/s with varying solid circulating rates,and 1.3-mm glass beads were used as a solid phase.Figure 6shows a comparison of the results from these three different measurement methods.It can be seen that the solid holdups computed from every technique were fairly similar at liquid velocities less than ap-proximately 14cm/s.However,at the liquid velocity greater than 14cm/s,the solid holdups measured by the ultrasonic technique and the pressure transducer were still in close agreement.In contrast,the

values

Figure 4.Standard deviation of (a)transmission time and (b)amplitude ratio as a function of gas and solid holdups in two-phase flows.

5686Ind.Eng.Chem.Res.,Vol.43,No.18,2004

determined by the wall conductivity probes were lower than the data from the other two techniques.The disagreement between the results of the

conductivity

Figure 6.Comparison of solid holdups measured by three different techniques in the liquid -solid circulating fluidized

bed.

Figure 5.Comparison of experimental and predicted phase holdups for the three-phase systems of (a)434-µm,(b)700-µm,and (c)1.3-mm glass beads.

Table 1.Comparison between the Predicted and Experimental Phase Holdups

solid holdup

gas holdup l (m)U l (cm/s)Q g (mL/s)G s

(kg/m 2‚s)pred exptl pred exptl 0.5

9.709.20.07310.07270

9.735.19.20.08720.07200.02650.02229.7014.10.07720.07450

9.735.114.10.09470.09020.02220.02481.2

9.709.20.07440.07080

9.735.19.20.07790.07520.02150.02129.7014.10.06830.07290

9.735.114.10.09820.09080.02140.02530.5

15.209.20.03490.03080

15.235.19.20.03730.02140.01140.020615.2014.10.05090.04000

15.235.114.10.08600.04920.01060.03031.2

15.209.20.03120.02850

15.235.19.20.03720.02120.01090.021915.2014.10.05140.04580

15.2

35.1

14.1

0.0952

0.0434

0.0108

0.0269

Ind.Eng.Chem.Res.,Vol.43,No.18,20045687

probes and those of the ultrasonic technique and the pressure transducer increases when the fluidization system applied a higher liquid velocity.Nevertheless,the results obtained with the three methods showed a similar profile.It can be concluded that the conductivity probes underestimated the solid holdups at relatively high liquid velocities.The variation of the conductivity data may be created by the high turbulence of the flow structure in the circulating system.The wall conductiv-ity probes were not sensitive enough to detect the small changes of the cross-sectional phase holdups of the circulating fluidized system.The data shown in Figure 6also revealed that the ultrasonic technique was a reliable measurement device.Consequently,the ultra-sonic technique should give an accurate measurement of the phase holdups in three-phase circulating flow for the particle sizes and operating conditions studied.The explanation given in the previous paragraph could account for the difference between the phase holdups computed by the ultrasonic analysis technique and those measured by the pressure transducer and conductivity probes presented in Table 1,especially at liquid velocity higher than 14cm/s.

Local Phase Holdups and the Axial Variation of the Radial Phase Distribution.Figure 7shows the local gas and solid holdups obtained at seven different radial positions in GLSCFB of 433-µm and 1.3-mm glass beads.The measurements were taken at a fixed liquid velocity and solid circulating rate with five gas flow rates.It was clear that nonuniform distributions of solid and gas holdups existed in both solid systems after the fluidized bed was completely operated under the three-phase circulating fluidization,U l )9.7and 15.2cm/s for 433-µm and 1.3-mm glass beads,respectively.Local solid concentration slightly increased from the center to the wall of the riser.The radial distribution of gas bubbles presented the opposite trend.The gas holdup was highest at the bed center.As a liquid flowed through a cylindrical pipe,the liquid velocity distributed nonuniformly across the pipe with higher liquid velocity in the center but lower at the wall.The difference in local liquid velocity generated an inward pressure,which forced gas bubbles to move toward the center of the riser.Solid particles in the riser were accelerated by the flow of the fluid phase (liquid and gas).In the center of the bed,a higher fluid velocity carried more solids upward,which therefore was responsible for a higher solid concentration at the wall and the radial nonuniform profiles of the solid holdup.It was also seen that the radial distributions of solid and gas holdups were symmetric about the axis of the riser for both solid systems (Figure 7).The radial distribution of gas holdup was more uniform in the 1.3-mm glass bead system comparing to the 433-µm glass bead system.

In the three-phase circulating system of 1.3-mm glass beads,the local gas and solid holdups were also inves-tigated at a higher axial position,located at 1.2m from the gas distributors.Figure 8shows the comparison of the radial distributions of phase holdups at two different axial elevations.Similar nonuniform distributions of solid and gas holdups were observed at higher location as well.Local solid holdup was a maximum at the column wall and a minimum at the center of the riser.The opposite trend was obtained for the radial gas distribution.Moreover,under the same operating condi-tions,the concentrations of gases and solids measured at two locations were apparently similar.The nonuni-formity of radial phase distributions was therefore developed longitudinally within the three-phase circu-lating fluidized bed.This characteristic indicated a uniform axial flow structure.Thus,in the GLSCFB,solid particles typically circulated in the unit system with liquid velocity higher than the particle terminal velocity.Beyond this point,a fully developed flow structure was formed within the GLSCFB riser.

The influences of the operating conditions on the local phase holdups and their profiles were also determined.The effect of gas flow rate on solid holdup (Figures 7and 9)was more significant in the 1.3-mm glass bead system.When the gas flow rate increased,the volume fraction of 1.3-mm glass beads increased;however,the radial profile of the solid phase was not changed.The gas holdups expectedly increased with increasing gas flow rate for all solid

systems.

Figure 7.Symmetrical radial distribution of solid and gas holdups for the systems of (a)433-µm and (b)1.3-mm glass beads.

5688Ind.Eng.Chem.Res.,Vol.43,No.18,2004

Figure 9shows the effect of liquid velocity on phase holdup at higher location (1.2m).A higher liquid velocity resulted in more uniform radial distributions of gas bubbles and solid particles and lower gas and solid holdups.The increase of liquid velocity enhanced the turbulence in the column,resulting in a more uniform distribution of phase holdups.On the other hand,gases and solids stayed at a shorter location

in

Figure 8.Symmetrical radial distribution of (a)solid and (b)gas holdups for the system of 1.3-mm glass beads at two axial

locations.

Figure 9.Effect of liquid velocity on the radial distribution of solid and gas holdups for the system of 1.3-mm glass bead at axial location of 1.2

m.

Figure 10.Effect of solid circulating rate on the radial distribution of solid and gas holdups for the system of 1.3-mm glass beads at axial location of 1.2m.

Ind.Eng.Chem.Res.,Vol.43,No.18,200456

The solid holdup increased with increasing solid circulating rate.However,the solid circulating rate had no observable influence on the uniformity of the solid profile(Figure10).Moreover,the local gas holdup and its radial distribution appeared to be independent of the solid circulating rate.The results of the effects of operating conditions,presented in the last two para-graphs,were in agreement with the work of Vatanakul34 for the system of434-µm and1.3-mm glass beads at lower axial position.

Conclusions

The applications of ultrasonic technique and statisti-cal approach were very practical and accurate to inves-tigate the phase holdups in multiphase flows.This method was proved to be reliable to measure simulta-neously the gas and solid holdups under high liquid velocity conditions within the two-phase and three-phase circulating fluidized beds as well.In a GLSCFB, the radial distributions of gas and solid holdups were nonuniform,whereas the profiles were symmetric about the axis of the riser in both solid systems.The solid holdup was low in the center but high toward the column wall of the riser,whereas the gas holdup distri-bution presented an opposite trend.The local solid and gas holdups at two different elevations along the height of the riser column were found to be similar,as were the radial profiles of gas bubbles and solid particles. These observations confirmed the hydrodynamic behav-ior of gas bubbles and solid particles within the GLSCFB studied.There was a uniform distribution in the axial direction and a nonuniform distribution in the radial directions for gas and solid phases.The nonuniform radial distribution of the solid holdup did not vary significantly with the solid circulating rate within the operating range of this study.The radial distributions of both phase holdups became more uniform and the concentrations of gases and solids decreased when the liquid velocity increased.Larger solid particles gener-ated a more uniform radial distribution of gas holdup. Acknowledgment

The authors gratefully acknowledge financial assis-tance from NSERC.Help provided by Frank Collins, Jody Chessie,Keith Rollins,and Dr.Andy Patel is highly appreciated.

Nomenclature

A)amplitude of ultrasonic signals,%

A0)amplitude of ultrasonic signals in pure liquid phase, %

Cov)covariance

e)absorption coefficient

k)wavenumber of ultrasonic wave

d p)particl

e diameter,cm

d h s)Sauter mean diameter,cm

G s)particle circulating rate,kg/m2‚s

l)axial position,m

L)canonical weight

M)canonical weight

Q)flow rate,mL/s

r/R)reduced radius

∆T)transmission time difference

U)superficial velocity,(cm/s)Var)variance

V)weighted canonical variate,defined in eq5

W)weighted canonical variate,defined in eq4

T)Transmission time of the ultrasound

T h)average value of T

Z)traveled distance of the incident sound wave

Greek Symbols

)phase holdup

λ)Lagrange multiplier

)scattering coefficient

F)canonical correlation coefficient

F*)optimum canonical correlation coefficient

R)absorption coefficient

δ)separation coefficient

δ*)optimum separation coefficient

ν)volumetric interfacial area

Subscripts

s)solid phase

l)liquid phase

g)gas phase

gl)gas-liquid two phase

gls)gas-liquid-solid three phase

U)upper portion

L)lower portion

Literature Cited

(1)Fan,L.-S.Gas-Liquid-Solid Fluidization Engineering; Butterworth:Boston,MA,19.

(2)Uchida,S.;Okamura,S.;Katsumata,T.Measurement of Longitudinal Distribution of Solids Holdup in a Three-Phase Fluidized Bed by Ultrasonic Technique.Can.J.Chem.Eng.19, 67,166-169.

(3)Stolojanu,V.;Prakash,A.Hydrodynamic measurements in

a slurry bubble column using ultrasonic technique.Chem.Eng. Sci.1997,52,4225-4230.

(4)Soong,Y.;Gamwo,I.K.;Blackwell,A.G.;Schehl,R.R.; Zarochak,M.F.Measurements of Solids Concentration in a Three-Phase Reactor by an Ultrasonic Technique.Chem.Eng.J.1995, 60,161-167.

(5)Soong,Y.;Blackwell,A.G.;Schehl,R.R.;Zarochak,M.F.; Rayne,J.A.Ultrasonic characterization of three-phase slurries. Chem.Eng.Commun.1995,138,213-224

(6)Soong,Y.;Gamwo,I.K.;Blackwell,A.G.;Harke,F.W.; Ladner,E.P.Ultrasonic Characterizations of Slurries in a Bubble Column Reactor.Ind.Eng.Chem.Res.1999,38,2137-2143.

(7)Goldberg,B.B.;Kotler,M.N.;Ziskin,M.C.;Waxham,R.

D.Diagnostic Uses of Ultrasound;Grune and Stratton:New York, 1975.

(8)Smith,T.N.Measurement of Drop Size in Liquid-Liquid Dispersions.Chem.Eng.Sci.1974,29,583-587.

(9)Havlicek,A.;Sovava,H.Measurement of Liquid Holdup of the Dispersed Phase by an Ultrasonic Methodology.Collect.Czech. Chem.Commun.1984,49,378.

(10)Tavlarides,L.L.;Bonnet J.C.An Ultrasonic Technique for Dispersed Phase Holdup Measurements.U.S.Patent4,726,-221,1988.

(11)Kirou,V.I.;Tavlarides,L.L.;Bonnet,J.C.;Tsouris,C. Flooding Holdup and Drop Size Measurements in a Multistage Column Extractor.AIChE J.1988,34,283-292.

(12)Kirou,V.I.,Tavlarides,L.L.;Bonnet,J.C.An Ultrasonic Technique for Contral of Holdup Profile in a Multistage Column Extractor.Ind.Eng.Chem.Res.1988,27,1084-1086.

(13)Yi,J.;Tavlarides,L.L.Model for hold-up measurements in liquid dispersions using an ultrasonic technique.Ind.Eng. Chem.Res.1990,29,475-482.

(14)Urick,R.J.A Sound Velocity Method for Determining the Compressibility of Finey Divided Substances.J.Acoust.Soc.Am. 1947,18,983-987.

(15)Hampton,L.Acoustic Properties of Sediments.J.Acoust. Soc.Am.1967,42,882-0.

(16)Atkinson,C.M.;Kytomaa,H.K.Acoustic Wave Speed and Attenuation in Suspensions.Int.J.Multiphase Flow1992,18, 577-592.

5690Ind.Eng.Chem.Res.,Vol.43,No.18,2004(17)Chang,J.S.;Ichikawa,Y.;Irons,G.A.;Morala,E.C.;Wan, P.T.Void Fraction Measurement by an Ultrasonic Transmission Technique in Bubbly Gas-Liquid Two-Phase Flow.In Measuring Techniques in Gas-Liquid Two-phase Flows;Delhaye,J.M., Cognet,G.,Eds.;Springer-Verlag:New York,1984.

(18)Stravs,A.A.;Stocker,U.V.Measurement of interfacial areas in gas-liquid dispersions by ultrasonic pulse transmission. Chem.Eng.Sci.1985,40,1169-1175.

(19)Bugmann,G.;Lister,J.B.;Stockar,U.V.Characterizing bubbles in bioreactors using light or ultrasound probes:data analysis by classical means and by neural networks.Can.J.Chem. Eng.1991,69,474-480.

(20)Uchida,S.;Okamura,S.;Katsumata,T.;Iida,K.Measure-ment of Solids Holdup in a Three-Phase Fluidized Bed by an Ultrasonic Technique.Chem.Eng.Sci.19,44,196-198.

(21)Anson,L.W.;Chivers,R.C.Ultrasonic Propagation in Suspensions s A Comparison of a Multiple Scattering and an Effective Medium Approach.J.Acoust.Soc.Am.19,85,535-540.

(22)Macchi,A.;Grace,J.R.;Bi,H.Use of ultrasound for phase holdup measurements in multiphase systems.Can.J.Chem.Eng. 2001,79,570-578.

(23)Maezawa A.;Muramatsu,S.;Uchida,S.;Okamura,S. Measurement of Gas Hold-up in Three-Phase Systems by Ultra-sonic Technique.Chem.Eng.Technol.1993,16,260-262.

(24)Warsito;Maezawa,A.;Uchida,S.;Okamura,S.A Model for Simultaneous Measurement of Gas and Solid Holdup in a Bubble Column using Ultrasonic Methodology.Can.J.Chem.Eng. 1995,73,734-743.

(25)Warsito;Ohkawa,M.;Kawata,N.;Uchida,S.Cross-Sectional Distribution of Gas and Solid Holdups in Slurry Bubble Column Investigated by Ultrasonic computed Tomography.Chem. Eng.Sci.1999,54,4711-4728.

(26)Warsito;Ohkawa,M.;Maezawa, A.;Uchida,S.Flow Structure and Phase Distributions in a Slurry Bubble Column Measured by an Ultrasonic Methodology.Chem.Eng.Sci.1997, 52,3941-3947.

(27)Yasser,A.I.;Briens,C.L.;Margaritis,A.;Bergongnou, M. A.Hydrodynamic characteristics of a three-phase inverse fluidized-bed column.AIChE J.1996,42,18-1990.

(28)Vatanakul,M.;Zheng,Y.;Couturier,M.Ultrasonic Mea-surement Technique for Multiphase Systems.In Proceedings of the7th International Conference on Circulating Fluidized Beds; 2002;pp497-504.

(29)Vatanakul,M.;Zheng,Y.;Couturier,M.Ultrasonic mea-surement technique for phase holdups of multiphase systems. Chem.Eng.Commun.,n press.

(30)Horvat,D.;Mozina,J.;Zun,I.;Perpar,M.Laser ultrasonics for bubbly flow detection.Ultrasonics1998,36,565-568.

(31)Zheng,Y.;Zhang,Q.Simultaneous Measurement of Gas and Solid Holdups in Multiphase Systems Using Ultrasonic Technique.Chem.Eng.Sci.,manuscript submitted.

(32)Bensler,H.P.;Delhaye,J.M.;Favreau,C.Measurement of Interfacial Area in Bubbly Flows by Means of an Ultrasonic Technique.In Proceeding of the ANS National Heat Transfer Conference;1987,pp240-246.

(33)Carlin,B.Ultrasonic,1st ed.;McGraw-Hill:New York, 1949.

(34)Vatanakul,M.;Zheng,Y.;Couturier,M.Flow character-ization of a three-phase circulating fluidized bed using an ultra-sonic technique.Can.J.Chem.Eng.,i n press.

Received for review October14,2003

Revised manuscript received March22,2004

Accepted March24,2004

IE034184C Ind.Eng.Chem.Res.,Vol.43,No.18,20045691