Hydrodynamics of an inclined gas–liquid cocurrent upflow packed bed

Hana Bouteldja, Mohsen Hamidipour, Faïçal Larachi n

Department of Chemical Engineering, LAVAL University, Québec, Canada G1V 0A6

H I G H L I G H T S

� Hydrodynamic of inclined gas–liquid cocurrent upflow packed beds was studied.� Liquid saturation, bed pressure drop and gas–liquid segregation were measured.� Short circuits of gas phase formed along upper wall due to inclination.� Inception of transition from bubble to segregated flow regime was identified.� Segregation developed along bed with minimum impact close to entrance.

a r t i c l e i n f o

Article history:Received 29 April 2013Received in revised form17 July 2013Accepted 17 August 2013Available online 25 August 2013

Keywords:Packed bed hydrodynamicsInclinationSegregationLiquid saturationElectrical capacitance tomography (ECT)

a b s t r a c t

The effects of inclination on the hydrodynamic behavior of a packed bed operating under gas–liquidcocurrent upflow were experimentally investigated in terms of liquid saturation, bed overall pressuredrop and gas–liquid segregation. The non-invasive electrical capacitance tomography (ECT) imagingtechnique was applied to scrutinize local and axial phase distribution pattern and cross-sectionallyaveraged liquid saturation. The results indicate that bed inclination creates short circuits for the gasphase along the upper wall where it can flow in a segregated manner. Inception of transition from bubbleto segregated flow regime was identified through monitoring a defined uniformity factor for ECT images.Phase segregation developed along the bed with minimum impact in the region close to the entrance.The removed bubbles were replaced by liquid phase resulting in higher liquid saturation values ascomplete segregation state was approached. The effect of operating conditions on axial profile of liquidsaturation was examined.

& 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Bubble columns are used in petrochemical, chemical, pharmaceu-tical, biochemical and metallurgical industries as multiphase contac-tors and reactors (Degaleesan et al., 2001). Processes which requiregood contact between gas and liquid phases can be performed inbubble columns (Prakash and Briens, 1990) where a discontinuous gasphase, in the form of bubbles, circulates upward while accompanyingthe continuous liquid phase. Bubble columns may also consist ofthree-phase systems and contain inert, reactive or catalytic particleseither in suspension or in a packed bed form. Wastewater treatment,hydrogenation, oxidation, chlorination, polymerization and alkylationare among the processes that have long been performed in bubblecolumn reactors. Other applications such as absorption, catalytic slurryreactions, and coal liquefaction have been also performed in thesereactors (Joshi et al., 1990; Blenke, 1979; Chisti, 1989; Saez et al., 1998).Simplicity of operation, low operating costs, large interfacial area, good

inter-phase heat and mass transfers, and ease of adjustment of liquidresidence time are the main advantages of this configuration(Kantarcia et al., 2005). However, reliable design and scale-up areknown to be restricted by the complex hydrodynamics and itsinfluence on transport characteristics. During the past decades,scientific interest in bubble column reactors has increased consider-ably (Kantarcia et al., 2005). Research on bubble column covers a widerange of subjects such as gas holdup, bubble properties, flow regimes,back mixing, interfacial area, pressure drop, and heat and masstransfer (Ruzicka et al., 2001; Luther et al., 2004; Majumder et al.,2006; Majumder, 2008; Dudley, 1995).

Currently, areas of fossil-fuel off-shore extraction and processingare vividly interested on problems linked with inclined multiphaseflows. As a matter of fact, the hydrodynamic behavior of floatingreactors and separators on embarked boats such as in FLNG (floatingliquefied natural gas) and FPSO (floating production, storing and off-loading) systems is a crucial aspect of their design for the prediction ofthe floating unit performances (Gu and Ju, 2008; Zhao et al., 2011).Obviously, most hydrodynamic researches in multi-phase reactorshave dealt with vertical columns. In contrast, inclined configurationswith fixed or slurry catalyst phase are rather on the fringes of studies

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/ces

Chemical Engineering Science

0009-2509/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.ces.2013.08.042

n Corresponding author. Tel.: þ418 656 2131x3566; fax: þ1 418 656 5993.E-mail address: faical.larachi@gch.ulaval.ca (F. Larachi).

Chemical Engineering Science 102 (2013) 397–404

and are reported very sparsely in the literature. From a pragmaticstandpoint, column inclination may have detrimental or beneficialeffects on the performance of the reactor depending on its projectedutilization. There are a few studies available which investigate theeffect of inclination angle of the reactor (O'Dea et al., 1990; Sarkar et al.,1991; Del Pozo et al., 1992; Hudson et al., 1996; Yakubov et al., 2007;Valverde et al., 2008; Atta et al., 2010; Schubert et al. 2010).

Numerous studies have been reported on the issues associatedwith gas–liquid and steam–water flows in inclined ducts andchannels which are in particular encountered in cooling circuitsof pressurized water reactors in nuclear power plants. Singh andGriffith (1970) investigated slug flow of air and water at smallupward inclination angles and developed simple correlations forpressure drop and holdup. Slug flow in inclined pipes was alsoexamined by Bonnecaze et al. (1971) who reported data for air–water system. Barnea et al. (1985) reported data on flow patterntransitions for upward gas–liquid flow in pipes at inclinationangles from 01 to 901. Mathematical models previously presentedfor vertical and horizontal configurations were extended to coverthe full range of pipe inclinations.

Flow regimes, fluidization heterogeneity along with the corre-sponding heat and mass transfer characteristics have been studiedin inclined fluidized beds. Arai et al. (1973) theoretically investi-gated heat exchange between particles and gas in a multistageinclined fluidized bed. A fairly good agreement was recognizedbetween the theoretical and experimental results in a three-stageinclined fluidized bed. Furthermore, effectiveness of multiplyingstages was confirmed on heat efficiency both theoretically andexperimentally. A chart was made up on the relation between therequired number of stages and optimum conditions of heatexchange. Masliyah et al. (1989) studied the enhancement ofseparation of light and heavy particles from suspensions usinginclined channels. They observed that at a fixed set of operatingconditions, the increase of inclination from vertical position resultsin a greater degree of separation. O'Dea et al. (1990) studiedinclined fluidized beds, between 451 and 901, using four differenttypes of powders and air as the fluidizing medium. Flow regimesand transition condition have been identified experimentally andverified by a theoretical model. Sarkar et al. (1991) studied theflow of solid particles from a vertical fluidized bed to a receivingvessel through an inclined downward pipe. They investigated theeffects of the connecting pipe length, diameter, and inclinationangle and fluidizing agent velocity on the flow rate of solidparticles from the fluidized bed to the receiving vessel. Hudsonet al. (1996) studied the effects of small inclination angles (up to101 from the vertical position) in liquid–solid fluidized beds. Theymeasured local holdup and circulation pattern of the liquid andsolid phases, and developed a simple model that predicted thesolid circulation pattern. Yakubov et al. (2007) studied the struc-ture of a fluidized bed in inclined columns. Experiments wereconducted in two glass columns which could be positioned in thewhole range of inclination angles, from horizontal to vertical. Theresults showed that the expansion process of the fluidized beddepends strongly on the inclination angle. The column length wasfound to have a minor effect on the involved phenomena. Valverdeet al. (2008) experimentally examined how bed inclination affectsfluidization and sedimentation behavior of fine cohesive particles.They found that the main effect of inclination is to inducefluidization heterogeneity. The local gas velocity increases in theadjacent region to the upper wall at the expense of diminishedvelocity in the region adjacent to the lower wall. This situationcaused early onset of local bubbling in the region adjacent to theupper wall.

The packed bed operated with descending gas–liquid cocurrentflows in slanted configuration was investigated by Schubert et al.(2010) as an extension of studies of horizontal concurrent gas–liquid

flows in porous media (Iliuta et al., 2003). In addition to liquidsaturation and pressure drop, they studied the transition fromsegregated/trickle regime to pulse flow regime with respect toinclination angle. The Grosser et al. (1988) flow regime transitionmodel was modified by considering only axial components of the gasand liquid superficial velocities to predict the slant-dependent shiftsin transition from segregated/trickle to pulse flow and was found toagree with experimental data. An Eulerian computational fluiddynamics (CFD) framework was implemented by Atta et al. (2010)to simulate an inclined cocurrent downflow packed bed. Twoconfigurations were used (a) a straight tube with an artificiallyinclined gravity and (b) an inclined geometry with straight gravity.The comparison between electrical capacitance tomography data andthe predictions of the liquid saturation showed that there is aconsiderably quantitative deviation from the experimental data.However, the trends could be satisfactorily predicted. They recom-mended formulating appropriate drag force closures which shouldbe incorporated in the CFD model for quantifiable approximation ofthe inclined packed bed hydrodynamics.

Any experimental study analyzing the effect of inclination inpacked beds for ascending two-phase flows has been so far fullydisregarded. It can be anticipated that inclination would consider-ably affect the flow patterns and consequently segregated flowmay appear in the bed. Thus, knowledge of the basic hydrody-namics for such inclined packed-bed configuration is essential.This study presents the results of an experimental investigation ofthe hydrodynamic behavior of gas–liquid cocurrent upflowinclined packed bed columns. The goal is to explore the effect ofcolumn inclination angle on the flow characteristics. The hydro-dynamic response to different operating conditions is studied.

2. Experimental setup

The experimental set-up built to study the effect of inclinationangle on two-phase flow hydrodynamics in cocurrent upflow packedbed is illustrated in Fig. 1. Experiments were performed at atmo-spheric pressure and room temperature. The main element of theset-up is a Plexiglas cylindrical columnwith an inner diameter (ID) of0.057 m and a height of 1.50 m. Glass beads (3 mm, ε¼0.395) wereused as packing. The packed bed was kept in place using a metallicgrid inserted from top to avoid bed fluidization during the experi-ments. Kerosene and air were used as the liquid and gas phases,respectively. A peristaltic pump was used to feed the liquid from thereservoir to the bottom of the column. The two phases wereseparated at the top of the column by means of a gas/liquid separator.The liquid was returned to the reservoir and the gas was vented tothe atmosphere.

The column was supported on a steel frame capable of varyingthe angle of inclination continuously in the full range from horizontalto vertical position. A wide range of gas and liquid superficialvelocities were explored to cover the bubble and segregated/bubbleflow regimes. Table 1 summarizes the range of operating conditions.The inclination angle was varied in small steps in the range of 0–551.The angular precision was 0.51. Pressure drops through the bed wasmeasured using a differential pressure transmitter.

A twin-plane 12-electrode per-plane sensor with a ModelDAM200E data acquisition system (PTL300E type) was utilized(Process Tomography, Ltd.) to perform electrical capacitancetomography (ECT) measurements. Tomographic imaging is helpfulto capture, without interference with the actual hydrodynamics,the evolving tomography at depths inside the bed that areotherwise inaccessible from mere wall scrutiny. ECT is compatiblewith oil-like non-polar liquids such as kerosene, as used in ourexperiments. The ECT system produces instantaneous informationwith sampling frequencies of 50 Hz. A Tikhonov reconstruction

H. Bouteldja et al. / Chemical Engineering Science 102 (2013) 397–404398

algorithm was chosen to generate normalized permittivity imagesfrom the measured capacitances. Selection of this algorithm, withregard to our application was discussed in our previous works(Tibirna et al., 2006; Hamidipour et al., 2007). The sensor, with aninner diameter of 0.0635 m, can be pushed over the outer wall ofthe column and fixed at different axial positions. The sensor hasbeen placed at three different axial positions to identify thevertical flow dynamics profiles. ECT calibration enabled settinglowest and highest limits of the normalized permittivities intervalso that intermediate per-pixel normalized permittivity values inimages during flow conditions can be obtained for rendition of theper-pixel liquid saturations (Hamidipour et al. 2009).

2.1. ECT calibration

Liquid saturation measurements in the packed bed were carriedout after the calibration of the ECT sensor at flooded (100%) anddrained (0%) bed conditions. The average mixture permittivities canbe expressed as

Flooded bed mixture permittivity

e 1½ � ¼ ð1�εÞeSþεeL ð1ÞDrained-bed mixture permittivity

e 0½ � ¼ ð1�εÞeSþ ε�εresL

� �eGþεresL eL ð2Þ

The overall permittivity of a mixture in two-phase operation canbe expressed as

Gas–liquid (–solid) mixture permittivity

e GL½ � ¼ ð1�εÞeSþ ε�εresL �εf dL

� �eGþ εf dL þεresL

� �eL ð3Þ

where ε, εresL and εf dL represent, respectively, the bed porosity, theresidual liquid holdup retained due to capillary forces and free-draining liquid holdup, and eS, eL, and eG, are the packing, liquid,and gas permittivities, respectively. Furthermore, the free-drainingliquid holdup normalized by the effective bed porosity (afterresting the residual liquid holdup) or the free-draining liquidsaturation (βf dL ), can be estimated using the normalized permittiv-ity, NoP

NoP ¼ e GL½ ��e 0½ �

e 1½ ��e 0½ � ¼εf dL

ε�εresL¼ βf dL ð4Þ

Instantaneous local free-draining saturation values for all the32�32 pixels of the ECT image were determined, βf dL;i, to providethe cross-sectionally averaged free draining liquid saturation, βf dL ,

βf dL ¼ 1NP

∑NP

i ¼ 1βf dL;i and βf dL ¼ 1

NF∑NF

j ¼ 1

1NP

∑NP

i ¼ 1βf dL;i

!ð5Þ

Time-averaged liquid saturation values, βf dL , were obtained forsteady-state flow conditions from �300 successive cross-sectionalimages. Number of pixels and frames applied are denoted by NPand NF, respectively. Afterward, using free-draining liquid satura-tions (Eq. (4)), the free-draining liquid holdup, εf dL , could becalculated knowing the experimentally determined bed porosity,ε, and the residual liquid holdup, εresL .

3. Results and discussion

3.1. Effect of inclination angle on local liquid distribution patternand overall bed pressure drop

The study was realized through examination of liquid satura-tion and bed pressure drop as a function of inclination angle andfluid throughputs. The inclination angle was initiated from uni-form liquid distribution (i.e., 01) and changed up to 551 to covera wide range of inclinations. To characterize the flow patterns thatevolve from bubble to segregated flow regime, the non-invasiveECT method was applied. Each ECT image represents a two-dimensional tomogram averaged over 5 cm thick slices in the

Fig. 1. Schematic diagram of the experimental setup.

Table 1System proprieties and range of operating conditions.

Parameter Value/range

Gas superficial velocity, uG 1.55–46.5 mm/sLiquid superficial velocity, uL 0.7–2.8 mm/sLiquid phase density 792 kg/m3

Liquid viscosity 0.002 kg/m/sLiquid surface tension 0.025 N/mGas density 1.2 kg/m3

Glass beads diameter 3 mmSphericity factor 1.0Bed porosity 0.395Bed length 1.5 mReactor diameter, ID 57 mmInclination angle 0–551

H. Bouteldja et al. / Chemical Engineering Science 102 (2013) 397–404 399

reactor corresponding to the electrodes' heights. Systematicexperimental comparisons between vertically aligned packed bedreactor and inclined bed in two-phase flow operation were carriedout to determine at which inclination angle deviations in flowpatterns start to occur. Therefore, the conventional straight reactoroperated in a stable bubble flow regime was gradually inclined.Fig. 2 shows the development of the liquid saturation over thecross section of the packed bed 55 cm away from bed entrance forinclinations varying from θ¼01 (vertical) up to 551. Vessel inclina-tion inevitably affects the gas–liquid distribution and flow patternsinside the packed bed. At 101, the segregation has already startedto develop becoming more pronounced the higher the inclination.The appearance of blue color in ECT images represents theformation of gas channels along the upper wall consisting of largebubbles as a function of increased inclination angle resulting inless gas–liquid contact. The red color illustrates that liquid ismoving toward a gas-free state, i.e., complete segregation, along-side the lower wall.

The segregation state can be identified based on the lack ofcrosswise uniformity of the liquid saturation distribution. A uniformityfactor, χ, was defined based on the deviation of pixel saturation withrespect to the average cross-sectional liquid saturation

χ ¼ 1NF

∑NF

j ¼ 1

1NP

∑NP

i ¼ 1

βf dL;i�βf dL

βf dL

!20@

1A ð6Þ

Where NP is the number of pixels in the image, NF is the number ofsuccessive cross-sectional images (�300 frames) after establishmentof a steady state. Also, βf dL;i and βf dL are the ith pixel and average cross-sectional liquid saturations for each image, respectively. While close tozero value is an indication of a uniform distribution, higher uniformityfactor values describe occurrence of mal-distribution. The uniformityfactor values for different operating conditions as a function of

inclination angle at 55 cm away from the bed entrance are shown inFig. 3. For the presented range of operating conditions, the effect ofinclination angle on the uniformity factor starts at �101. All operatingconditions reveal a peak of uniformity factor where the worstdistribution pattern takes place. As the bed is tilted the gas phasehas the tendency to escape toward the upper wall. At a critical angle,complete phase segregation starts to develop andmore cross-sectionalarea is uniformly covered by either liquid or gas phases. Hence, theuniformity factor diminishes. Therefore, the peak characterizes thetransition from bubble to segregated flow regime. Fig. 3 depicts thatthe location of peak depends on the gas and liquid velocities. At lowphase interaction (i.e., low superficial velocity) segregation requiressmaller inclination angle to establish (�251). At high phase interaction(i.e., high superficial velocity) more deviation from vertical position isrequired for segregation to evolve (�351). Liquid phase appearsto play an important role. Fig. 3 shows that higher liquid velocity

0º

0 1

5º 10º 15º 25º 35º 55º

Fig. 2. Exemplary illustration of the inclination effect on the flow texture (uG¼15.5 mm/s; uL¼0.7 mm/s, 55 cm away from the bed entrance). (For interpretation of thereferences to color in this figure legend, the reader is referred to the web version of this article.)

0

0.5

1

1.5

2

0 10 20 30 40 50 60

Uni

form

ity F

acto

r (-

)

Inclination Angle (deg)

uG=1.55 mm/s; uL=0.7 mm/suG=15.5 mm/s; uL=0.7 mm/s

uG=1.55 mm/s; uL=2.8 mm/suG=15.5 mm/s; uL=2.8 mm/s

Fig. 3. Effect of inclination angle on uniformity factor (55 cm away from the bedentrance).

0

0.2

0.4

0.6

0.8

1

0 10 20 30 40 50 60

Liq

uid

Satu

ratio

n (-

)

Inclination Angle (deg)

uG=1.55mm/s; uL=0.7mm/suG=15.5mm/s; uL=0.7mm/suG=1.55mm/s; uL=2.8mm/suG=15.5mm/s; uL=2.8mm/s

0

4

8

12

16

20

0 10 20 30 40 50 60

Pres

sure

Dro

p (k

Pa/m

)

Inclination Angle (deg)

uG=1.55mm/s; uL=0.7mm/s

uG=15.5mm/s; uL=0.7mm/suG=1.55mm/s; uL=2.8mm/suG=15.5mm/s; uL=2.8mm/s

Fig. 4. Effect of inclination angle on (a) on cross-sectional average liquid saturation(55 cm away from the bed entrance) and (b) overall bed pressure drop.

H. Bouteldja et al. / Chemical Engineering Science 102 (2013) 397–404400

prevents facile escape of gas bubbles toward the upper wall, therefore,segregation occurs at higher inclination angle.

The cross-sectionally averaged liquid saturation (Eq. 5, at 55 cmaway from the bed entrance) data were analyzed as well as thebed overall pressure drop (Fig. 4a, b). The liquid saturation (Fig. 4a)is clearly affected by the inclination angle for all superficial fluidvelocities investigated in this study. At vertical position, increasinggas velocity at constant liquid throughput results in higherpresence of gas phase which in turn causes lower liquid saturation.For the selected range of operating conditions, at around 101deviation from vertical position, the average liquid saturationstarted to increase. As the bed is inclined, gas and liquid start tosegregate. The space originally occupied by gas is filled by liquidforcing the gas phase to squeeze close to the upper wall. Higherinclination angles, approaching complete segregation state, causemore coverage of cross-sectional area by the liquid. Therefore,higher liquid saturation is observed. As mentioned earlier, liquidvelocity has an important role to maintain bubbles in the liquidphase. Consequently, at same inclination angle, higher liquidvelocity results in lower liquid saturation.

In a bubble column, two main factors contribute to the overallbed pressure drop: (i) the interaction between gas, liquid andsolid phases (i.e., drag forces) and (ii) the static head of liquidphase. Fig. 4b shows the variation of bed overall pressure drop as

a function of inclination angle. Till around 101 inclination, as thedistribution pattern does not change significantly, a minor effect isobserved on pressure drop. Further inclination creates an easypath for the gas phase close to the upper wall resulting in lowerphase interactions. Therefore, as the bed is tilted toward completesegregation, pressure drop continues to decline. At lower liquidsuperficial velocity gas holdup is lower (Fig. 4a) and the contribu-tion of higher static liquid head results in greater overall pressuredrop. The results show that for the selected range of velocities atvertical position, for a constant gas superficial velocity, pressuredrop decreases by increase of liquid velocity. Higher liquidthroughput causes more phase interaction when the two phasescompete for flow path. However, a greater drag force is imposedon the gas phase by a low velocity liquid phase. The net outcomefor the range of our experiments (i.e., low velocities to preventpulse flow regime) is less overall bed pressure drop.

3.2. Effect of inclination angle and operating conditionson the axial phase distribution

An Eulerian slice representation of ECT images is used tovisualize the axial development of the liquid-flow field as shownin Fig. 5a, b (Hamidipour and Larachi, 2010). In our case, Eulerianslices were plotted to study the morphology of two-phase flow

0 2.85 (cm)-2.85Side view of the bed Top view of the bed

A A

A A

A At0t0+Δtt0+2 Δt

t0+(n -1) Δtt0+n Δt

t0+(n -2) Δt

f = 50 HzΔt = 1/50 = 0.02 s

0 1Fig. 5. (a) Schematic of Eulerian slice construction, and (b) ECT snapshots of bubble and segregated flow regimes; first row at vertical position, 120 cm away from the bedentrance, uL¼1.4 mm/s; uG¼1.55 mm/s; second row 55 cm away from the bed entrance, inclination angle¼251, uL¼1.4 mm/s; uG¼1.55 mm/s.

H. Bouteldja et al. / Chemical Engineering Science 102 (2013) 397–404 401

under segregated/bubble flow regime (Fig. 6a–c). Pixelized liquidsaturations reconstructed along a selected diametrical line (e.g.,A–A line in Fig. 5a) was shot at 50 Hz pace as recorded and plottedone after the other from individual images recorded during two-phase flow operation. In the current instance, the perpendicularline A–A crossing the segregated liquid area from the bottommostto the uppermost wall areas at a given axial position is the mostrepresentative line. Evolving this line time-wise from bottom totop as in Fig. 5a, would depict upstream events for the flowdirection but delayed in time until they hit the tomograph sensingplane. This gives, in an approximate sense, virtual local axialtomogram representation of the liquid-flow field.

Fig. 5b shows typical images used to build the Eulerian slices.Each row shows different instants of flow under specific operatingconditions. The first row (Fig. 5b) was obtained at vertical position,120 cm away from the bed entrance (uL¼1.4 mm/s; uG¼1.55 mm/s).The second row presents different moments of flow, 55 cmaway from of the bed entrance while bed was inclined at 251(uL¼1.4 mm/s; uG¼1.55 mm/s). Several images (21 s�50 Hz)similar to these two rows were used to construct Fig. 6a–c (i.e.,Eulerian slices). Liquid and gas superficial velocites were keptconstant (uL¼1.4 mm/s; uG¼1.55 mm/s) whilst the inclinationangle was increased stepwise (01, 251, and 451) for different axialpositions of ECT, 20 cm, 55 cm and 120 cm away from bedentrance.

Maintaining the bed vertically revealed no effect of height. Underbubble flow regime and vertically positioned bed (10) slightly lowerliquid saturations are observed close to the walls. This is attributed to

the higher porosity in this region. As inclination angle was increased,the gravity force acting on the liquid increased resulting in moreaccumulation of liquid in the bottom wall region. Lower liquidsaturation nearby the upper wall area is in accordance with thepassage of gas bubbles. It is noted that at inclination angle equal to251 middle of the column shows less phase segregation compared tothe top of the column. At 451, segregation is less close to the entranceof the column; however, the distribution pattern from middle to thetop of column is similar. In fact, segregation requires a minimumlength to develop which is a function of inclination angle. At higherinclination angles less distance is necessary to come close to acomplete phase segregation state. At the top of the column, a wavypattern is observed at an inclination angle equal to 251. The wavestend to disappears with the increase of inclination angle (Fig. 6c). Thewaves are an indication of the region where gas–liquid disengage-ment occurs. At higher inclination angles this region moves towardthe bed entrance, therefore, a relativly calm pattern is establisheddownstream of the disengagement zone.

Fig. 7a shows the effect of inclination angle on the axial profileof cross-sectionally averaged liquid saturation for uG¼1.55 mm/sand uL¼2.8 mm/s. At vertical position (i.e., 01) liquid saturation isindependent of bed height. An increasing trend of liquid saturationis observed along the bed with deviations from vertical position.Development of phase segregation along the bed creates morespace for liquid presence equivalent to higher liquid saturation. Forthe selected operating conditions, at inclination angle¼451, finalphase distribution pattern was approached around midway of thebed resulting in an almost invariant liquid saturation afterwards.

0

0.2

0.4

0.6

0.8

1

Liquidesaturation

0

4

8

12

16

20

t(s)0 2.85-2.85 0 2.85-2.850 2.85-2.85

Sensor plane

0º 25º 45º

0

0.2

0.4

0.6

0.8

1

Liquidesaturation

0

4

8

12

16

20

t(s)0 2.85-2.85 0 2.85-2.85 0 2.85-2.85

Sensor plane

0º 25º 45º

0

0.2

0.4

0.6

0.8

1

Liquidesaturation

0

4

8

12

16

20

t(s)0 2.85-2.850 2.85-2.85 0 2.85-2.85

Sensor plane

0º 25º 45º

Fig. 6. Eulerian slice representation for uL¼1.4 mm/s; uG¼1.55 mm/s, (a) 20 cm, (b) 55 cm, and (c) 120 cm away from the bed entrance.

H. Bouteldja et al. / Chemical Engineering Science 102 (2013) 397–404402

It is important to note that even for small inclination anglescomplete segregation might take place in a long bed.

Fig. 7b depicts the effect of gas superficial velocity on the axialprofile of liquid saturation for uL¼2.8 mm/s and inclinationangle¼451. Similar to the vertical bed behavior, higher gas velocityboosts the gas–liquid interaction while seeking flow space. Thus,part of liquid is pushed out and lower liquid saturation is observedalong the bed. This effect is less pronounced in the region close tothe entrance due to higher static head of liquid phase whichimposes higher resistance.

The influence of liquid superficial velocity on the axial profile ofliquid saturation for uG¼1.55 mm/s and inclination angle¼451 ispresented in Fig. 7c. At low liquid velocity, no significant effect isobserved along the bed. Increase of liquid velocity results in lowerliquid saturation close to bed entrance. The axial shear imposed onthe gas bubbles by the liquid phase retards bubble removal to

higher elevations which in turn lowers the area occupied by liquid(i.e., less liquid saturation).

4. Conclusion

The hydrodynamic behavior of an inclined gas–liquid cocurrentupflow packed bed was experimentally investigated. The inclina-tion angle was varied from 01 (i.e., vertical position) up to 551 tocover a wide range of inclination. The electrical capacitancetomography (ECT) imaging technique was implemented to moni-tor local and axial distribution patterns. Cross-sectionally averagedliquid saturation, bed overall pressure drop and gas–liquid segre-gation were measured. In addition, the effect of operating condi-tions on axial profile of liquid saturation was examined. Theexperimental results indicated that

– Short circuits of gas phase were formed along the upper walldue to bed inclination. Thus, phase interaction decreasedcausing lower pressure drop values.

– Inception of transition from bubble to segregated flow regimewas identified through monitoring uniformity factor for ECTimages. The worst distribution pattern (i.e., the highest unifor-mity factor) was considered as the flow regime transition point.

– Segregation was developed along the bed and removed bubbleswere replaced by liquid phase resulting in higher liquid satura-tion values as complete segregation state was approached.

– High liquid superficial velocity imposed a greater shear on thegas bubbles preventing segregation especially in the regionclose to the entrance.

Nomenclature

e electrical permittivity, F/mt time, su superficial velocity, mm/s

Subscripts

G gasL liquidS solid

Greek Letters

e bed porosity, dimensionlessb liquid saturation, dimensionlessw uniformity factor, dimensionless

Acknowledgment

Financial support from the Canada Research Chair “Greenprocesses for cleaner and sustainable energy” and the DiscoveryGrant from the Natural Sciences and Engineering Research Council(NSERC) is gratefully acknowledged.

References

Arai, N., Hasatani, M., Sugiyama, S., 1973. Heat transfer in multistage inclinedfluidized beds. Chemical Engineering 37, 379.

Atta, A., Schubert, M., Nigam, K.D.P., Roy, S., Larachi, F., 2010. Co-current descendingtwo-phase flows in inclined packed beds: experiments versus simulations.Canadian Journal of Chemical Engineering 88, 742–750.

0

0.2

0.4

0.6

0.8

1

0 20 40 60 80 100 120 140

Liq

uid

Satu

ratio

n (-

)

Height (cm)

uG = 1.55 mm/suG = 7.75 mm/suG = 15.5 mm/suG = 46.5 mm/s

0

0.2

0.4

0.6

0.8

1

0 20 40 60 80 100 120 140

Height (cm)

0 °15 °25 °45 °

0

0.2

0.4

0.6

0.8

1

0 20 40 60 80 100 120 140

Height (cm)

uL = 0 mm/suL = 0.7 mm/suL = 1.4 mm/suL = 2.8 mm/s

Liq

uid

Satu

ratio

n (-

)L

iqui

d Sa

tura

tion

(-)

Fig. 7. Axial profile of liquid saturation, (a) effect of inclination angle,uG¼1.55 mm/s and uL¼2.8 mm/s, (b) effect of gas superficial velocity,uL¼2.8 mm/s and inclination angle¼451, and (c) effect of liquid superficial velocity,uG¼1.55 mm/s and inclination angle¼451.

H. Bouteldja et al. / Chemical Engineering Science 102 (2013) 397–404 403

Bonnecaze, R.H., Erskine, W., Greskovich, E.J., 1971. Hold-up and pressure drop fortwo-phase slug flow in inclined pipelines. AIChE Journal 17, 1109–1113.

Blenke, H., 1979. Loop reactors. Advances in Biochemical Engineering 13, 121–215.Barnea, D., Shoham, O., Taitel, Y., 1985. Gas–liquid flow in inclined tubes: flow

pattern transitions for upward flow. Chemical Engineering Science 40, 131–136.Chisti, Y., 1989. Airlift Bioreactors. Elsevier Applied Science, London and New York.Del Pozo, M., Briens, C.L., Wild, G., 1992. Effect of column inclination on the

performance of three-phase fluidized beds. AIChE Journal 38, 1206–1212.Dudley, J., 1995. Mass transfer in bubble columns: a comparison of correlations.

Water Research 29, 1129–1138.Degaleesan, S., Dudukovic, M., Pan, Y., 2001. Experimental study of gas-induced

liquid-flow structures in bubble columns. AIChE Journal 47, 1913–1931.Grosser, K., Carbonell, R.G., Sundresan, S., 1988. Onset of pulsing in two-phase

cocurrent downflow through a packed bed. AIChE Journal 34, 1850–1860.Gu, Y., Ju, Y., 2008. LNG-FPSO: offshore LNG solution. Frontiers of Energy and Power

Engineering in China 2, 249–255.Hamidipour, M., Larachi, F., Ring, Z., 2007. Cyclic operation strategy for extending

cycle life of trickle beds under gas liquid filtration. Chemical EngineeringScience 62, 7426–7435.

Hamidipour, M., Larachi, F., Ring, Z., 2009. Monitoring filtration in trickle beds usingelectrical capacitance tomography. Industrial and Engineering ChemistryResearch 48, 1140–1153.

Hamidipour, M., Larachi, F., 2010. Characterizing the liquid dynamics in cocurrentgas–liquid flows in porous media using twin-plane electrical capacitancetomography. Chemical Engineering Journal 165, 310–323.

Hudson, C., Briens, C.L., Prakash, A., 1996. Effect of inclination on liquid solidfluidized beds. Powder Technology 89, 101–113.

Iliuta, I., Fourard, M., Larachi, F., 2003. Hydrodynamic model for horizontal two-phase flow through porous media. Canadian Journal of Chemical Engineering81, 957–962.

Joshi, J.B., Ranade, V., Gharat, S.D., Lele, S.S., 1990. Sparged loop reactors. CanadianJournal of Chemical Engineering 68, 705–741.

Kantarcia, N., Borakb, F., Ulgena, K.O., 2005. Bubble column reactors. ProcessBiochemistry 40, 2263–2283.

Luther, S., Rensen, J., Guet, S., 2004. Bubble aspect ratio and velocity measurementusing a four-point fiber-optical probe. Experiments in Fluids 36, 326–333.

Majumder, S.K., 2008. Analysis of dispersion coefficient of bubble motion andvelocity characteristic factor in down and upflow bubble column reactor.Chemical Engineering Science 63, 3160–3170.

Majumder, S.K., Kundu, G., Mukherjee, D., 2006. Prediction of pressure drop in amodified gas–liquid down flow bubble column. Chemical Engineering Science61, 4060–4070.

Masliyah, J.H., Nasr-El-Din, H., Nandakumar, K., 1989. Continuous separation ofbidisperse suspensions in inclined channels. International Journal of Multi-phase Flow 15, 815–829.

O'Dea, D.P., Rudolph, V., Chong, Y.O., Leung, L.S., 1990. The effect of inclination onfluidized beds. Powder Technology 63, 169–178.

Prakash, A., Briens, C.L., 1990. Porous gas distributors in bubble columns: effect ofstart-up procedure on distributor performance. Canadian Journal of ChemicalEngineering 68, 204–210.

Ruzicka, M.C., Zahradnık, J., Drahos, J., Thomas, N.H., 2001. Homogeneous hetero-geneous regime transition in bubble columns. Chemical Engineering Science56, 4609–4626.

Saez, A.E., Marquez, M.A., Roberts, G.W., Carbonell, R.G., 1998. Hydrodynamic modelfor gas-lift reactors. AIChE Journal 44, 1413–1423.

Sarkar, M., Gupta, S.K., Sarkar, M.K., 1991. An experimental investigation of the flowof solids from a fluidized bed through an inclined pipe. Powder Technology 64,221–231.

Schubert, M., Hamidipour, M., Duchesne, C., Larachi, F., 2010. Hydrodynamics ofcocurrent two-phase flows in slanted porous media – modulation of pulse flowvia bed obliquity. AIChE Journal 56, 3189–3205.

Singh, G., Griffith, P., 1970. Determination of the pressure drop optimum pipe sizefor a two-phase slug flow in an inclined pipe. Journal of Manufacturing Scienceand Engineering 92, 717–726.

Tibirna, C., Edouard, D., Fortin, A., Larachi, F., 2006. Usability of ECT for quantitativeand qualitative characterization of trickle-bed flow dynamics subject to filtra-tion conditions. Chemical Engineering and Processing 45, 538–545.

Valverde, J.M., Castellanos, A., Quintanilla, M.A.S., Gilabert, F.A., 2008. Effect ofinclination on gas-fluidized beds of fine cohesive powders. Powder Technology182, 398–405.

Yakubov, B., Tanny, J., Maron, D.M., Brauner, N., 2007. The dynamics and structureof a liquid–solid fluidized bed in inclined pipes. Chemical Engineering Journal128, 105–114.

Zhao, W.H., Yang, J.M., Hu, Z.Q., Wei, Y.F., 2011. Recent developments on thehydrodynamics of floating liquid natural gas (FLNG). Ocean Engineering 38,1555–1567.

H. Bouteldja et al. / Chemical Engineering Science 102 (2013) 397–404404