Tetrafluoroethane (R134a) hydrate formation within variable volume reactor

Tetrafluoroethane (R134a) hydrate formation within variable volume reactoraccompanied by evaporation and condensationK. Jeong, Y. S. Choo, H. J. Hong, Y. S. Yoon, and M. H. Song Citation: Review of Scientific Instruments 86, 035102 (2015); doi: 10.1063/1.4913650 View online: http://dx.doi.org/10.1063/1.4913650 View Table of Contents: http://scitation.aip.org/content/aip/journal/rsi/86/3?ver=pdfcov Published by the AIP Publishing Articles you may be interested in An improved film evaporation correlation for saline water at sub-atmospheric pressures AIP Conf. Proc. 1440, 1085 (2012); 10.1063/1.4704324 Slow evaporation and condensation on a spherical droplet in the presence of a noncondensable gas Phys. Fluids 22, 067101 (2010); 10.1063/1.3432130 On the negative mass flows in evaporation and condensation problems Phys. Fluids 17, 127106 (2005); 10.1063/1.2147608 Flow Visualization within the Evaporator of Planar Loop Heat Pipe AIP Conf. Proc. 746, 195 (2005); 10.1063/1.1867135 Development and Test Results of a Multi‐Evaporator‐Condenser Loop Heat Pipe AIP Conf. Proc. 654, 42 (2003); 10.1063/1.1541275

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

203.250.105.121 On: Fri, 06 Mar 2015 15:32:05

REVIEW OF SCIENTIFIC INSTRUMENTS 86, 035102 (2015)

Tetrafluoroethane (R134a) hydrate formation within variable volumereactor accompanied by evaporation and condensation

K. Jeong, Y. S. Choo, H. J. Hong, Y. S. Yoon, and M. H. Songa)

Department of Mechanical, Robotics, and Energy Engineering, Dongguk University,Seoul 100-715, South Korea

(Received 11 December 2014; accepted 15 February 2015; published online 4 March 2015)

Vast size hydrate formation reactors with fast conversion rate are required for the economic imple-mentation of seawater desalination utilizing gas hydrate technology. The commercial target produc-tion rate is order of thousand tons of potable water per day per train. Various heat and mass transferenhancement schemes including agitation, spraying, and bubbling have been examined to maximizethe production capacities in scaled up design of hydrate formation reactors. The present experimentalstudy focused on acquiring basic knowledge needed to design variable volume reactors to producetetrafluoroethane hydrate slurry. Test vessel was composed of main cavity with fixed volume of140 ml and auxiliary cavity with variable volume of 0 ∼ 64 ml. Temperatures at multiple locationswithin vessel and pressure were monitored while visual access was made through front window.Alternating evaporation and condensation induced by cyclic volume change provided agitation dueto density differences among water and vapor, liquid and hydrate R134a as well as extended interfacearea, which improved hydrate formation kinetics coupled with latent heat release and absorption. In-fluences of coolant temperature, piston stroke/speed, and volume change period on hydrate formationkinetics were investigated. Suggestions of reactor design improvement for future experimental studyare also made. C 2015 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4913650]

I. INTRODUCTION

Gas hydrates are nonstoichiometric inclusion solid com-pound formed from guest gas molecules and host water mole-cules under sufficiently high pressure condition at a giventemperature. Application areas of gas hydrate technologyinclude seawater desalination and carbon dioxide capture fromflue gas taking advantage of distinct stable thermodynamicconditions corresponding to different gases, as well as naturalgas storage and transportation and hydrogen storage utilizingrelatively high gas compressibility.

Vast size hydrate formation reactors with fast conversionrate are required for the economic implementation of seawaterdesalination utilizing gas hydrate technology. The commercialtarget production rate is order of thousand tons of potable waterper day per train. HFCs (Hydro-fluorocarbons) are consideredas suitable guest gas for seawater desalination since they formhydrates at mild (relatively low pressure) conditions. In thepresent study, tetrafluoroethane (R134a) is chosen as hydrateformer whose formation/dissociation kinetics and thermody-namics have drawn many research attentions.1–4 A detailedreview of experimental devices and methods for measuringhydrate formation kinetics is available.5

However, hydrate formation kinetics are disadvanta-geously slow as HFCs and water barely dissolve into eachother. In order to achieve rapid hydrate formation, drasticincrease in gas/water interface area is imperative as wellas heat and mass transfer enhancement. Several schemes to

a)Author to whom correspondence should be addressed. Electronic mail:[email protected].

increase interfacial area during hydrate formation have beendemonstrated besides typical simple stirring6 such as staticmixer,3 jet reactor,7,8 bubbling column,9,10 and spray.11,12

Recently, Ngema et al.13 demonstrated a variable volumetest cell to rapidly determine hydrate dissociation pressureunder isothermal condition. Other researches also employedvariable volume reactors for various purposes. Lee et al.14

employed variable volume view cell for easy and quickmeasurement of bubble point pressure and the critical point.Licence et al.15 introduced transparent sapphire piston drivenby hydraulic ram within the cell body for the determinationof phase-equilibria and cloud point measurement. Direct P-V-T measurements using the variable volume view cell withhandle bar has been conducted and help a visual and intuitiveunderstanding of critical phase behavior of CO2.16 The test cellfabricated by operation by Silva et al.17 contained a movablepiston driven by syringe pump, which permits operation ofthe pressure cycling.

The present experimental study focused on acquiringbasic knowledge needed to design variable volume reactorsto produce R134a hydrate slurry. Test vessel was composedof main cavity with fixed volume and auxiliary cavity withvariable volume with brass piston driven by electric motor andcontroller. Temperatures at multiple locations within vesseland pressure were monitored while visual access was madethrough front window. Alternating evaporation and conden-sation induced by cyclic volume change provided agitationdue to density differences among water and vapor, liquidand hydrate R134a as well as extended interface area, whichimproved hydrate formation kinetics coupled with latent heatrelease and absorption. Influences of coolant temperature,piston stroke/speed, and volume change period on hydrate

0034-6748/2015/86(3)/035102/8/$30.00 86, 035102-1 © 2015 AIP Publishing LLC This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

203.250.105.121 On: Fri, 06 Mar 2015 15:32:05

035102-2 Jeong et al. Rev. Sci. Instrum. 86, 035102 (2015)

formation kinetics were investigated, and the results are re-ported below.

II. EXPERIMENTAL APPARATUS AND PROCEDURES

Formation of tetrafluoroethane (R134a) hydrate was madeto occur within temperature and volume controlled high pres-sure vessel. The motivation of employing variable volume testvessel was to induce alternating evaporation and condensa-tion and thus to provide agitation due to density differencesamong vapor, liquid, hydrate R134a, and water together withextended vapor R134a and water interface area, which im-proves hydrate formation kinetics coupled with latent heatrelease and absorption. Schematic diagram and photograph ofexperimental apparatus are presented in Figure 1.

The test vessel is composed of main and auxiliary cavities.The main cavity is a cylindrical space with horizontal axishaving inner diameter of 70 mm and axial length of 30 mm.The front wall of the main cavity serves as a window forvisual observation as it is sealed with 20 mm thick transparentpolycarbonate polymer plate. The rear end of main cavity isconnected to the front end of auxiliary cavity, which sharesthe same horizontal axis with main cavity. The inner diam-eter of the auxiliary cavity is 30 mm, and the axial lengthcan vary between 0 and 90 mm as the rear end of auxiliarycavity is sealed with reciprocating piston made of brass. Thevolume of main cavity is approximately 140 ml and that ofauxiliary cavity is 0 ∼ 64 ml. The circumferential wall of maincavity is 30 mm thick SUS 304 and has a port for watersupply and drain and another port for gas supply, vent, andpressure monitoring. Three additional ports were deployed toallow for the insertion of temperature probes from the front.The circumferential wall of auxiliary cavity is 5 mm thickSUS 304. The pressure of test vessel was monitored with apressure transducer (Wika, Type A-10), and temperatures at

various locations were measured with type T thermocouples.Data were processed and recorded with data logger (Agilent,34970A) and software program HP-VEE. After calibration,the uncertainties of measured temperature and pressure were0.1 ◦C and 0.1% of readings, respectively.

The coolant jacket and circumferential fins surroundedouter surfaces of the reactor to provide the passage for 10 wt. %glycol aqueous solution, which was circulated from constanttemperature bath (Jeio Tech., RC-10 V) to maintain the me-dium in test cavities at desired temperatures. The coolantjacket was insulated at the outer surface with 30 mm thickpolyethylene foam insulation except for the visual access areaas shown in photograph at lower right corner of Figure 1.A stereo microscope (Nikon, SMZ1000) equipped with CCDcamera was connected to a computer recording photographicand moving images during experimental runs through thiswindow.

The position of reciprocating piston was controlled by“power cylinder” (Changwongijeon, CWPSW-1000) drivenby 0.75 kW electrical motor with custom made programmablecontroller. Piston speed during compression and expansion canbe set, respectively, and piston position can be manipulated inmanual mode. In automatic mode, stroke and stay times at highand low pressure positions as well as the number of repetitioncan be specified to define series of volume changing cycle.Figure 2 shows the photograph of test vessel assembly andplumbing with insulation together with piston driving mecha-nism, two constant temperature baths (chiller and heater), andFreon tank.

In order to prepare for each test run, supply tank havingthe volume of 530 ml was heated to the temperature of 40 ◦Cand was charged with high purity (99.9%) industry gradeR134a to the pressure of 500 kPa. The pneumatic controlvalve installed between gas tank and supply tank ensured theexact supply tank charge pressure. After piston was moved tothe zero volume position of auxiliary cavity, the test vessel

FIG. 1. Schematics and photographs of experimental apparatus focused on volume variable test vessel assembly. This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

203.250.105.121 On: Fri, 06 Mar 2015 15:32:05

035102-3 Jeong et al. Rev. Sci. Instrum. 86, 035102 (2015)

FIG. 2. Photograph of insulated test vessel assembly, piston driver withcontroller, chiller, heater, and R134a tank.

was fully filled with de-ionized water, and exactly the sameamount of water at every run was withdrawn to secure about3 ml void space needed for liquid R134a addition. This spacewas flushed with Freon vapor from supply tank at least threetimes to be filled with ambient pressure Freon. Then, the testvessel assembly started to be chilled, and the supply tank wasrecharged to the pressure of 600 kPa while it was maintainedat constant temperature of 40 ◦C. Even though test vessel pres-sure was higher than liquid water–vapor Freon–hydrate phaseequilibrium pressure (Peq), hydrate formation did not occurbecause partial pressure of R134a was lower than Peq.

When the test vessel assembly had reached the predeter-mined initial temperature, the Freon was fed to test cavitiesuntil the supply tank pressure was reduced to 500 kPa. Anotherpneumatic control valve installed between supply tank and testvessel ensured the exact amount of R134a to be added to thetest vessel. Without void space, R134a could not be added.Despite multiple flush, unknown amount of impurity gas hasremained or was introduced within void and access ports.Owing to the inclusion of impurity gas, the initial pressurewas larger than saturation pressure. As soon as the feedingwith R134a was over, pre-programmed cyclic motion of pis-ton started. To aid in the discussion of experimental results,thermo-physical properties of relevant phases are summarizedin Table I.

TABLE I. Density and formation enthalpy of relevant phases.18,19

Phase ρsat at 4 ◦C(kg/m3)

∆H

(kJ/kg)LW 1000 0LR 1281 −175.4VR 16.56 0H ∼1047 −342.6

FIG. 3. The transient changes in pressure and temperatures obtained duringthe representing experiment (Case 8) showing (A) gas feeding stage, (B)induction stage, and (C) hydrate formation stage and time instances forFigs. 4, 5, and 6.

III. RESULTS AND DISCUSSIONS

The transient changes in vessel pressure and temperaturesmeasured during the representing experiment (Case 8 in Ta-ble II) are presented in Figure 3. Coolant temperature was 4 ◦C.Piston speed and stroke were 10 mm/s and 40 mm. The volumeof test vessel varied between 140 ml and 168 ml. The typicalexperimental run can be divided into three time stages: (A)gas feeding stage, (B) induction stage, and [(C) in Figure 3,respectively] hydrate formation stage.

During gas feeding stage [(A) in Figure 3], vessel pressureincreased to about 450 kPa (Pini), which is larger than liquidFreon–vapor Freon–hydrate phase equilibrium pressure (Psat).The amount of Freon introduced into test vessel was 2.64 g,equivalent to 2.06 ml of liquid Freon considering the densityof 1281 kg/m3. Initially, most Freon existed as liquid phaseforming a droplet adjacent to the lowest inner wall of maincavity bounded with water as shown in the upper photographof Figure 4(a). The rest of Freon existed as vapor within voidnear the top of main cavity mixed with impurity gas (N2). Eventhough the pressure and temperature condition belonged tohydrate stable region in phase equilibrium diagram, noticeablehydrate formation did not occur until the end of inductionstage.

Induction stage [(B) in Figure 3] began with the start ofcyclic volume change of auxiliary cavity. One volume changecycle is composed of “expansion” ( A - B ), “low pressure stay”( B - C ), “compression” ( C - D ), and “high pressure stay” [ D - A ’in Figure 5].

During expansion, liquid level was lowered and vesselpressure decreased due to void volume increase caused byincrease of auxiliary cavity volume. When the vessel pressuredecreased below Psat, evaporation of Freon is driven. However,the evaporation rate is not large enough to hold the vesselpressure near Psat when piston speed was 10 mm/s (7 ml/s involume). The decrease of vessel pressure occurred almostat constant rate throughout entire expansion. During the

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

203.250.105.121 On: Fri, 06 Mar 2015 15:32:05

035102-4 Jeong et al. Rev. Sci. Instrum. 86, 035102 (2015)

FIG. 4. Photographs taken during the representing experiment (Case 8); (a) at the beginning of induction stage, t=−481 s and −477 s; (b) at early hydrateformation stage, t= 276 s and 280 s; and (c) at late hydrate formation stage, t= 2932 s and 2936 s.

succeeding low pressure stay, vessel pressure was restoredtoward Psat. In the following discussion, the pressure at theend of low pressure stay is referred to as Psl (stabilized lowpressure). As evaporation continued during the low pressurestay, the liquid level is further lowered slightly. Separate oraggregated column like bubbles emanating from liquid Freondroplet near the bottom ascended resulting in simultaneousagitation and extension of vapor/liquid interface. Small Freondroplet still remained at the end of low pressure stay whichbelonged to induction stage [lower photograph of Figure4(a)]. During compression, the phenomena are reversed fromexpansion. The liquid level was raised and the vessel pressureincreased rapidly driving Freon condensation. Again, thecondensation rate is not large enough that vessel pressureincreased almost linearly with time during entire compression.The pressure at the end of compression belonged to induction

FIG. 5. Detailed change in vessel pressure during volume change cyclesindicating stabilized high and low pressures.

stage was well above Pini. During the succeeding high pressurestay, vessel pressure was restored toward Pini. In the followingdiscussion, the pressure at the end of high pressure stay isreferred to as Psh (stabilized high pressure). The liquid levelis further raised slightly through high pressure stay due tocondensation. During compression and high pressure stay,liquid Freon which condensed near the void/liquid interfacedescended either along the inner surface of circumferentialwall or along the front window of main cavity without causingsignificant fluid motion. One volume change cycle took 40 s,as each of expansion and compression lasted 4 s and eachof low and high pressure stay continued for 16 s. Duringinduction stage, Psh and Psl did not deviate much from Pini

and Psat, respectively, since hydrate had not formed yet.Hydrate formation stage [(C) in Figure 3] initiated with

the observation of hydrate crystals. Experimental time wasset to zero at the beginning of the first expansion strokeafter hydrate appearance. Operation of cyclic volume changehad been continued since the beginning of induction stage.In early hydrate formation stage [Figure 4(b)], crumbs andclusters of hydrate crystals formed mainly along the wallsand interface bounding the void during expansion and lowpressure stay. Crystal growth within liquid pool could beobserved.20 Some of hydrate crystals remained aggregatednear the top. Rest of crystals were separated from walls andwere drifting around within liquid pool due to fluid motion.Flow was induced by piston motion, bubble floating, liquidFreon sinking, and crystal descending. Hydrate rested nearbottom of main cavity as flow was mitigated during low andhigh pressure stays. Overtime, amount of Freon converted intohydrate gradually increased. Psl started to decrease noticeably(t = 333 s in Figure 3) when no more liquid Freon was left atthe end of low pressure stay. Psh started to decrease due to thevolume shrinkage caused by larger density of R134a hydrate(∼1047 kg/m3) compared to that of water. The difference

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

203.250.105.121 On: Fri, 06 Mar 2015 15:32:05

035102-5 Jeong et al. Rev. Sci. Instrum. 86, 035102 (2015)

between Psh and Psl steadily increased with hydrate formation.Figure 5 shows the detailed transient change in vessel pressureenlarged from Figure 3 for three volume change cycles in themidst of hydrate formation stage. At later times of hydrateformation stage [Figure 4(c)], the region of aggregated crystalsnear the top extended to the maximum void volume, while theregion of accumulated crystals near bottom occupied lowerhalf of main cavity. Gradually less amount of liquid Freon wasfound near the bottom until it became indistinct. In the longrun, evolution of vapor phase occurred simultaneously frommultiple locations. Operation of cyclic volume change stoppedwhen the decrease in Psl slowed down sufficiently, i.e., no morehydrate conversion occurred.

The upper diagram of Figure 3 displays transient tempera-ture changes during the representing experiment. Threetemperature measuring locations are indicated in Figure 4(a);T1 was inserted 8 mm into the main cavity along 40◦ counter-clockwise direction from the top, T2 was near the center, andT3 was inserted 8 mm into the main cavity along the oppositedirection of T1 insertion. Temperature at T1 location, which be-longed to void region during low pressure stay, fluctuated withthe amplitude larger than 3 ◦C during induction stage influ-enced by evaporation and condensation. The extent of fluctua-tion gradually reduced as hydrate formation proceeded until T1temperature converged to coolant temperature. Temperaturemeasured at T2, which belonged to condensed phase all thetime, was maintained close to coolant temperature. Data fromT3 location are omitted from Figure 3 because temperaturereadings from T2 and T3 differed little from each other. It is notclear how well T1 temperature represented vapor temperaturebecause the tip of thermocouple T1 could be still wet afterthe liquid level was lowered below T1. Aggregated hydratecrystals might append to thermocouple sheathing tube, whichprovided preferred precipitation site for hydrate formation.

Measured pressure and temperature pairs at T1 and T2locations are plotted in phase equilibrium diagram (Figure 6)to investigate hydrate formation driving force. During volumechange cycles which belonged to induction stage [Figure 6(a)]vessel pressure dropped from 434 kPa to 269 kPa duringexpansion, restored to Psat during low pressure stay, increasedto 461 kPa during compression and returned to Pini afterhigh pressure stay. During expansion T1 temperature startedto decrease from coolant temperature (near 4 ◦C) as liquidlevel was lowered below thermocouple tip indicating that voidtemperature was lowered by vapor expansion. At the end ofexpansion ( B ), pressure and temperature noticeably deviatedfrom equilibrium saturation. Void temperature was restoredto coolant temperature during low pressure stay implyingthat cooling caused by evaporation little influenced eithervoid or condensed phase temperatures. The agitation inducedby ascending bubbles provided sufficient heat offsetting theabsorption of evaporation latent enthalpy. T1 temperaturestarted to increase during compression. When liquid level wasraised past thermocouple tip, T1 temperature was more than2 ◦C higher than the coolant temperature. Media near interfacebetween void and liquid pool were relatively warmer dueto compression and latent energy release caused by conden-sation. Heat removal was relatively inefficient because va-por has poor thermal conductivity and fluid motion caused

FIG. 6. Vessel pressure and temperatures measured at T1 and T2 locations onP-T diagram during volume change cycle; (a) at induction stage, t=−160∼−124 s, and (b) at hydrate formation stage, t= 920∼ 956 s.

by descending liquid Freon was weak. T1 temperature wasrestored to coolant temperature during high pressure stay. T2temperature remained near coolant temperature indication thatmedium temperature at lower portion of liquid pool was notinfluenced significantly by vapor expansion/compression andFreon evaporation/condensation.

During volume change cycles which belonged to hydrateformation stage [Figure 6(b)], the magnitude of vessel pres-sure change gradually increased. Temperatures measured atT1 were slightly higher than those at T2. It is not clear thatT1 temperature correctly represents vapor temperature becausethe upper portion of main cavity was occupied by aggregatedhydrate crystals which were wet even during low pressure stay.

In order to examine the influence of coolant temperature(Tc), high pressure stay time (tH), low pressure stay time(tL), piston stroke (S), and velocity (Vp) on induction time(t I) and hydrate formation kinetics, series of experimentswere conducted under various combinations of experimentalconditions. The cases for which results are discussed in thepresent report are summarized in Table II. The change in

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

203.250.105.121 On: Fri, 06 Mar 2015 15:32:05

035102-6 Jeong et al. Rev. Sci. Instrum. 86, 035102 (2015)

TABLE II. Summary of experimental conditions.

Tc S tH tL τ tI Vp

Case (◦C) (mm) (s) (s) (s) (s) (mm/s)

17

80 2 10 20 787

20

2 40 2 10 16 160

34

80 2 10 20 154 40 2 10 16 40

5 7 80 10 2 20 1546 4 80 10 2 20 54

74

40 20 12 40 848108 40 16 16 40 478

9 40 10 20 40 256

induction time is known to show stochastic nature.21 However,Table II clearly reveals that induction time significantly de-creases at lower coolant temperature and larger piston speedwithin tested ranges of experimental conditions. When Tc was4 ◦C and Vp was 20 mm/s, hydrate formation could be initiatedwithin a few volume change cycles. The favored conditionsfor short induction time need not concur with those for fasterhydrate formation kinetics.

In the following discussions, the trace of stabilized lowpressure (Psl) demonstrated in Figure 5 is used to quantifyhydrate conversion rate. Disadvantages of employing Psl asan index for hydrate formation kinetics are as follows. (1)Psl starts to decrease sensitively only after no liquid Freon isavailable for evaporation at the end of low pressure stay. There-fore, information regarding the hydrate formation kinetic inthe very beginning of hydrate formation stage is confined tothe time required for the amount of liquid Freon at the end oflow pressure stay to become zero. The amount is equivalentto 77% and 55% of total Freon mass supplied during feedingstage when stroke was 40 mm and 80 mm, respectively. (2) Theresults for different strokes cannot be directly compared. Thetransient change in Psl for different coolant temperature andpiston strokes is shown in Figure 7. The influence of coolanttemperature is evident. The lower coolant temperature resultsin faster hydrate formation kinetics as well as shorter induction

FIG. 7. Influence of Tc and S on transient change in Psl.

FIG. 8. Transient changes in Psl for insufficient low pressure stay time.

time. Since void volume is the larger for the larger stroke,decrease in Psl started earlier. When coolant temperature ishigher, the starting Psl is larger because Psat is larger.

The transient changes in Psl when low pressure stay timeis insufficient are shown in Figure 8. Experimental conditionsfor Cases 5 and 6 are identical with Cases 1 and 3, respectively,except that low pressure stay time and high pressure stay timeare switched. In analyzing Cases 5 and 6, it was found that staytimes need to be longer than 10 s when stroke is 40 mm. Whenlow pressure stay time was only 2 s, Psl was not stabilizedproducing inappropriate fluctuations.

Finally, the influence of low and high and low pressurestay time sharing was examined in Figure 9. Coolant tempera-ture was 4 ◦C, stroke was 40 mm, and Vp was 10 mm/s for Cases7, 8, and 9. The sum of high pressure stay time and low pressurestay time was fixed at 32 s. The hydrate formation was fasterwhen the longer time was assigned for high pressure stay. Twofactors are believed responsible for this enhancement. (1) Thedriving force to form hydrate is stronger during high pressurestay compared to that during low pressure stay as shown inFigure 6. The fugacity difference driving hydrate formation islarger when the vertical distance from Lw-V-H triple phase linein P-T diagram is larger. (2) Removal of latent heat release

FIG. 9. Influence of stay time sharing on hydrate formation kinetics. This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

203.250.105.121 On: Fri, 06 Mar 2015 15:32:05

035102-7 Jeong et al. Rev. Sci. Instrum. 86, 035102 (2015)

caused by hydrate formation is more efficient during highpressure stay due to relatively small void volume having lowthermal conductivity.

IV. CONCLUSIONS

In order to acquire basic knowledge needed to designvariable volume reactors to produce tetrafluoroethane (R134a)hydrate slurry, series of hydrate formation experiments utiliz-ing variable volume reactor have been performed focusedon the influence of coolant temperature, piston stroke/speed,and volume change period on induction time and hydrateformation kinetics. Measured vessel pressure and tempera-tures at different locations together with visual observationsare analyzed to obtain the following conclusions.

(1) The influence of coolant temperature is dominant amonginvestigated test parameters. Lower coolant temperatureis favored for faster hydrate formation kinetics as well asshorter induction time.

(2) Larger piston speed resulted in shorter induction timewithin tested ranges of experimental conditions. When Tc

was 4 ◦C and Vp was 20 mm/s, hydrate formation couldbe initiated within a few volume change cycles.

(3) When piston speed was larger than 10 mm/s (7 ml/sin volume), the evaporation rate during expansion wasnot large enough to hold the vessel pressure near Psat atinduction stage. At the end of expansion, pressure andtemperature noticeably deviated from equilibrium satu-ration. The condensation rate during compression wasalso not large enough so that vessel pressure at the endof compression was well above Pini.

(4) Some of hydrate crystals remained aggregated near thetop. Hydrate particles which were separated from wallswere drifting within liquid pool until rested near bottomof main cavity during low and high pressure stays as flowwas mitigated.

(5) Fluid motion induced by piston motion, bubble floating,liquid Freon sinking, and crystal descending as well asextended interface area improved hydrate formation ki-netics. The favored conditions for short induction timeneed not concur with those for faster hydrate formationkinetics. The 77% of total Freon mass supplied duringfeeding stage could be converted into hydrate within500 s when stroke was 40 mm.

(6) Stay times need to be longer than 10 s under the testconditions examined. Otherwise, vessel pressure is notstabilized as expansion/compression starts while evapo-ration/condensation is still in progress.

(7) The hydrate formation was faster when the longer timewas assigned for high pressure stay because the drivingforce to form hydrate is stronger during high pressurestay compared to that of low pressure stay. Also, theremoval of latent heat release is more efficient duringhigh pressure stay because void volume having poorthermal conductivity is smaller.

Suggestions regarding test reactor improvement forfurther experimental study before the scale up design of R134ahydrate slurry production reactor are as follows.

(1) Test cell having considerable vertical dimension is rec-ommended to fully explore the enhancement caused byascending bubbles.

(2) Apparatus to ensure the migration of formed hydratecrystals into liquid pool needs to be added.

(3) Experimental diagnostics to quantify hydrate conversionrate is needed. For example, addition of salt into liquidwater phase and monitoring the changes in concentrationby measuring electric conductivity will be tried.

(4) Influence of impurity gas needs to be investigated.

Finally, a viable scale up strategy based on the knowledgeobtained from the present study toward 2000 tons of potablewater production from seawater per day commercial plant canbe outlined as follows.

(1) A bench scale assembly is to be fabricated for furtherresearch purpose, composed of a horizontal cylinder,having 160 mm diameter and 1800 mm length, and apiston, located along the middle length of the cylinderand having 200 mm stroke. The bench scale assembly isthe 5:1 reduced size design of a pilot plant facility.

(2) A pilot plant of 10 tons of potable water per day produc-tion capacity is to be built for process development pur-pose. The size of the cylinder, having 800 mm diameterand 9 m length, and piston, having 1000 mm stroke,is already within commercially available range. Theheat transfer feasibility and mechanical problems can berecognized and improved.

(3) A demo production (or a small scale industrial plant) of40 tons of potable water per day production capacity isto be constructed for verification purpose of commercialscale plant. Four of the pilot plant units are combined inthe similar manner as four cylinder internal combustionengine to save the energy required to drive pistons. Theprocess control data can be obtained and analyzed.

ACKNOWLEDGMENTS

This work was supported by the Energy Efficiency &Resources of the Korea Institute of Energy Technology Evalu-ation and Planning (KETEP) grant, funded by the Ministry ofKnowledge Economy, Republic of Korea (No. 2007-M-CC22-P-01-0-000). Authors are grateful for the support.

1J. Li, D. Liang, K. Guo, R. Wang, and S. Fan, Energy Convers. Manage. 47,201 (2006).

2D. Liang, K. Guo, R. Wang, and S. Fan, Fluid Phase Equilib. 187, 61 (2001).3H. Tajima, T. Nagata, Y. Abe, A. Yamasaki, F. Kiyono, and K. Yamagiwa,Ind. Eng. Chem. Res. 49, 2525 (2010).

4F. Nikbakht, A. A. Izadpanah, F. Varaminian, and A. H. Mohammadi, Int.J. Refrig. 35, 1914 (2012).

5E. D. Sloan, Jr. and C. Koh, Clathrate Hydrates of Natural Gases (CRCpress, New York, 2007).

6B. Tohidi, R. Burgass, A. Danesh, K. Østergaard, and A. Todd, Ann. N. Y.Acad. Sci. 912, 924 (2000).

7R. P. Warzinski, D. E. Riestenberg, J. Gabitto, I. V. Haljasmaa, R. J. Lynn,and C. Tsouris, Chem. Eng. Sci. 63, 3235 (2008).

8P. Szymcek, S. D. McCallum, P. Taboada-Serrano, and C. Tsouris, Chem.Eng. J. 135, 71 (2008).

9Y.-T. Luo, J.-H. Zhu, S.-S. Fan, and G.-J. Chen, Chem. Eng. Sci. 62, 1000(2007).

10S. Hashemi, A. Macchi, and P. Servio, Chem. Eng. Sci. 64, 3709 (2009). This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

203.250.105.121 On: Fri, 06 Mar 2015 15:32:05

035102-8 Jeong et al. Rev. Sci. Instrum. 86, 035102 (2015)

11K. Fukumoto, J.-i. Tobe, R. Ohmura, and Y. H. Mori, AIChE J. 47, 1899(2001).

12R. Ohmura, S. Kashiwazaki, S. Shiota, H. Tsuji, and Y. H. Mori, EnergyFuels 16, 1141 (2002).

13P. T. Ngema, W. M. Nelson, P. Naidoo, D. Ramjugernath, and D. Richon,Rev. Sci. Instrum. 85, 045123 (2014).

14J. M. Lee, B. C. Lee, and C. H. Cho, Korean J. Chem. Eng. 17, 510 (2000).15P. Licence, M. P. Dellar, R. G. M. Wilson, P. A. Fields, D. Litchfield, H. M.

Woods, M. Poliakoff, and S. M. Howdle, Rev. Sci. Instrum. 75, 3233 (2004).16M. M. Hoffmann and J. D. Salter, J. Chem. Educ. 81, 411 (2004).

17J. Melo Silva, A. A. Rigo, I. A. Dalmolin, I. Debien, R. L. Cansian, J. V.Oliveira, and M. A. Mazutti, Food Control 29, 76 (2013).

18Y. A. Cengel, M. A. Boles, and M. Kanoglu, Thermodynamics: An Engi-neering Approach (McGraw-Hill, New York, 2005).

19S. Hashimoto, T. Makino, Y. Inoue, and K. Ohgaki, J. Chem. Eng. Data 55,4951 (2010).

20R. Kumar, J. D. Lee, M. Song, and P. Englezos, J. Cryst. Growth 310, 1154(2008).

21R. Ohmura, M. Ogawa, K. Yasuoka, and Y. H. Mori, J. Phys. Chem. B 107,5289 (2003).

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

203.250.105.121 On: Fri, 06 Mar 2015 15:32:05