A Study of Bubble Behavior and Boiling Heat Transfer ...

Heat Transfer Engineering, 21:33–45, 2000 Copyright ° C 2000 Taylor & Francis 0145–7632/00 $12.00 + .00

A Study of Bubble Behavior and Boiling Heat Transfer Enhancement under Electric Field SI-DOEK OH and HO-YOUNG KWAK Mechanical Engineering Department, Chung-Ang University, Seoul, Korea

The effect of a D.C. electric eld on nucleate boiling heat transfer for refrigerants, R11, R113, and FC72, was investigated experimentally in a single-tube shell/tube heat exchanger by using the temperature control method of wall superheat. Also the behavior of bubble under nonuniform electric eld produced by wire electrodes was studied by numerical calculation. For R11, the electrohydrodynamic (EHD) enhancement for boiling heat transfer was observed for all ranges of wall superheat tested. However, the enhancement in boiling heat transfer disappeared if the wall superheat exceeded 13±C for R113, and no electric eld effect on the boiling heat transfer was observed for FC72. An application of approximately 5 kV was enough to eliminate the boiling hysteresis for R11 and R113. Numerical study of the electric eld in a single medium has hinted that the bubbles are forced away from the heating surface and toward the electrostatic stagnation point by the dielectrophoretic force. Such modi ed bubble motion turns out to promote the boiling heat transfer if one uses proper electrode con guration.

For utilizing low-temperature waste heat sources, one of the major tasks is to develop a high-performance heat exchanger. An especially compact evaporator is an important thermal component for plants such as organic Rankine cycle engines and large-scale heat pumps. EHD augmentation [1] has been proved to be one of the most appropriate techniques to enhance nucleate boiling heat transfer in dielectric liquids which are suitable working

uids for the evaporators employed in waste heat recovery plants. Previous experiments of EHD enhancement in boiling heat transfer have been done mainly on the lm boiling regime [2 – 4], where dramatic increase in heat transfer rate occurs. Such great enhancement is known to be due to the lm destabilization caused by electrical forces acting on the vapor – liquid interface [5]. However, the electric eld effect on nucleate boiling is not yet fully understood. Considerable fundamental research should be done to investigate the nucleation mechanism in a cavity under an electric eld and how the dielectrophoretic force due to the difference between the dielectric permittivity of the liquid and vapor phases in a nonuniform electric eld [6] affects the bubble behavior near the boiling surface, which in turn promotes the

The authors wish to acknowledge the Korea Science and Engineering Foundation for supporting this work under contract 921-0900-014-2. Si-Doek Oh is now at Industrial Machinery & Engineering Team, Research & Development Institute, Hyosung Corporation, 450, KongdukDong, Mapo-Ku, Seoul, 121-020, Korea. Address correspondence to Ho-Young Kwak, Mechanical Engineering Department, Chung-Ang University, Seoul, 156-756, Korea. E-mail: [email protected]

33

heat transfer rate. Recently, EHD enhancement by up to a factor of 10 has been obtained from a low- n tube with complete elimination of boiling hysteresis [1]. Disturbance of the heat transfer layer due to the buoyancydriven motion of bubbles trapped in the weak eld region of the low- n tube [1] has been reported to bring about dramatic enhancement in nucleate boiling heat transfer. Bubble behavior under electric elds and temperature gradients was studied by Ogata and Yabe [7]. One of the common ndings from the previous investigations on EHD enhancement in nucleate boiling is that the number of bubble increases while the diameter of bubble decreases as the electric eld increases [1, 8– 10]. Another interesting observation from previous experiments is the bubble coalescing on the lower part of heat transfer tube surrounded by six wire electrodes with equal spacing [9 – 11]. However, most experimental works on EHD enhancement in boiling heat transfer have been done by electrically controlling the heat ux generated. Nucleate boiling heat transfer data obtained by heat ux control, which show better heat transfer curves, do not provide appropriate data for the design of evaporators utilized in waste heat recovery plants, where waste heat rather than electricity is a main source of evaporation of working uid. The heat ux control method is easily achieved by controlling the input voltage to the heater which is in contact with the boiling surface. However, the temperature of the boiling surface is controlled by means of heat transfer to the surface. It has been found that the two methods provide different slopes of boiling curves for the cases without [12] and with [13] electric eld. Only a systematic experiment on EHD enhancement by the temperature control method has been reported for the small range of the wall superheat less than 10 K [1]. In this study, the effect of a d.c. electric eld on nucleate boiling heat transfer for refrigerants, R11, R113, and FC72, was investigated experimentally by using a single-tube shell/tube heat exchanger. The broad range of the wall superheat for boiling action was controlled by the temperature of the water owing inside tube. The wire electrodes employed in this study, which turn out to be an appropriate electrode type for tube bundle evaporators, are simpler ones, so that the bubble behavior under nonuniform electric eld can be easily tested and analyzed. Also the electric eld strength around the heating surface including the electrodes was calculated by solving the Poisson equation numerically, to study the bubble motion in a nonuniform electric eld. Detailed phenomena related to nucleate boiling under an electric eld, such as effect of the wall superheat and the forces affecting the bubble motion, were investigated in this study. 34

heat transfer engineering

EXPERIMENTAL APPARATUS AND PROCEDURES Experimental Apparatus A schematic diagram of the EHD augmentation boiling heat transfer unit is shown in Figure 1. The experimental unit consists of a single-tube shell/tube evaporator ° 1 , condensers ° 2 , ° 3 , constant-temperature bath circulator ° 4 , hot water storage tank ° 5 , and high-voltage supply ° 6 . The evaporator shell was made of stainless steel pipe of 150 mm inside diameter with two 125-mmdiameter sight glasses mounted at its midsection to facilitate visual observation of the bubble behavior on the boiling surface. The test tube in this study was made of brass with a smooth surface. The outside diameter of tube is 19.0 mm with 1.2-mm wall thickness. The tube length available for heat transfer is about 687 mm. A detailed crosssectional view of the test tube is shown in Figure 2. The tube wall temperatures were measured at ve axial stations with equal intervals of 150 mm as shown in Figure 3. Thermocouple junctions 3, 4, and 5 were used for the measurement of circumferential temperatures of the tube. Thermocouple junctions used were magnesium oxide insulation T-type sheathed with stainless steel tubing. Figure 4 shows the method of embedding of the junctions in the test tube wall. The stainless steel sheathed thermocouple of 1.6 mm diameter with bare junction was passed into the hole of 1.5 mm diameter in the wall and welded at the tube wall to maintain good thermal contact and to prevent leaking of working uid into the tube. Two additional T-type thermocouples were placed at the top and bottom of the shell to monitor the saturation temperature of working uid as shown in Figure 2. The pressure inside the shell was measured by a Bourdon-type pressure gauge. Hot water to the tube was supplied from an aluminum water tank. The water was heated by two 1-kW and four 100-W immersion heaters adjusted by variacs. Fine control of the water temperature in the tank was done by a temperature controller connected to four 100-W heaters. A friction pump ° 7 with a mechanical seal delivered the hot water from the tank to the tube. Hot water ow rate modi ed by the bypass line was determined by means of a calibrated turbine ow meter ° 8 (Kobold Co., DF-48 ) in the liquid line. A diaphragm-type accumulator ° 9 was installed in the liquid line to minimize the uctuation in the ow rate. The vapor generated in the shell-side tube was condensed and returned to the shell by gravity. Cold water supplied from a constanttemperature bath ° 4 circulated the cooling coil of the condenser. vol. 21 no. 4 2000

Figure 1

Schematic diagram of experimental loop for boiling experiment.

A schematic of the electrode con gurations used in this study is shown in Figure 5. Two electrode con gurations were tested, except for visualization. First, a set of six electrodes, oriented at 60± intervals, are located 5 mm away from the heating surface. In the second electrode con guration, two electrodes are placed just

Figure 2

below and above the heating surface. The rst electrode con guration can be made by rotating the second one by 30±. The electrodes were made of 0.9-mm-diameter carbon steel wire coated with copper. As can be seen in Figures 1 and 2, these wires were supported and insulated from the shell and tube by six ring-type Te on

Sectional drawing of test section.


vol. 21 no. 4 2000

35

Figure 3

Con guration of test section.

supporters tted to the tube. This arrangement with supporters resulted in a constant electric eld around the test tube circumference. A high-voltage generator (Kilovolt Co., model KV 30-20 ) was utilized to supply high voltage up to 30 kV to the electrodes. Also, a voltage regulator was used to prevent D.C. ripples due to the change in the input A.C. voltage to the generator. The high voltage was fed into the EHD evaporator through a specially modi ed spark plug tted in the Te on shell ange. The voltage and current to the electrode were measured by a voltmeter (Fluke, 80K-40 ) and a multimeter (Barnet Instrument Co., TS 352B/U ), respectively. Experimental Procedures For each run, degassing of the test liquid was performed by following procedure. After charging the shell with working uid, the wire heater winding around the

Figure 4

36

Thermocouple instrumentation of evaporator tube.


shell was turned on and the hot water in the storage tank was circulated until the temperature of the test liquid reached a predetermined value, which was continued until the pressure inside the shell reached 3 bar gauge due to the ebullition on the test tube surface. The liquid level in the shell was maintained at 55 mm above the test tube. Then cold water was circulated in the condenser coil until the vapor generated in the shell was condensed to reduce the system pressure to 1 bar gauge. Next, the noncondensable gases were purged by opening valve ° 6 . The above procedures were repeated several times for degassing. The wall superheat of the tube was adjusted by changing the power supply to the immersion heaters installed in the storage tank. Detailed adjustment of the system pressure was done by controlling the cooling-water ow rate and the temperature at the inlet of the condenser coil. If the temperature change of the vapor and liquid inside the shell are in the range of § 0.2±C, the system is assumed to have reached steady state. The variation

Figure 5 Electrode con gurations for electrostatic analysis and experiments.

vol. 21 no. 4 2000

of system pressure during the test was less than §2%. Once the steady state was established, the temperature and pressure of the system, the temperatures of the hot water at the inlet and outlet, the temperatures of the tube wall, and the voltage and current to the electrode were recorded. For all cases, the boiling experiment at a particular wall superheat without electric eld was performed rst. Next, the high voltage was applied to the electrode. In each case experiments were repeated twice to ensure the validity of data obtained. The saturation temperature at the experimental conditions were 25.5±C for R11, 48.5±C for R113, and 57.5±C for FC72. The ow rates of hot water, which are very important for achieving stable experimental condition, were 8 liters/min for R11 and R113 and 10 liters/min for FC72. For completion of one set of experiments, 10 h or so were required. The heat transfer rate data without electric eld were compared with an existing correlation by Stephan and Abdelsalam [14]. Data Reduction The heat transfer coef cient, h, was determined by h =

QÇ / A Tas ¡ Tsat

(1 )

where A is the heat transfer area of the tube, which is approximately 0.037 m2 . The heat transfer rate to the working uid inside the shell was assumed to be equal to the rate of energy loss of the hot water owing inside the tube. This is given by QÇ = mC Ç p (Twi ¡ Two )

(2 )

where mÇ is the mass ow rate of the water inside the tube. The value of the speci c heat of water, C p , was taken at the average value of the bulk inlet and outlet temperatures of the water, i.e., (Twi + Two )/ 2. The average temperature of the tube wall, Tas , was obtained by taking the arithmetic mean of the temperatures at the ve side-wall stations along the tube length with a weighting factor. That is, Tas =

(T1 + T2 + T3 + T6 + T7 ) ¢ Rm 5

(3 )

The weighting factor Rm is just the ratio of the arithmetic mean of the temperatures measured at the three circumferential stations to the temperature at the side wall, which is given by Rm =

T3 + 2T4 + T5 4T3

(4 ) heat transfer engineering

Table 1 Calibration of the thermocouples to measure the temperatures at the inlet and outlet of the water stream Temperature (±C) 35.0 40.0 Correction values (±C)

0.0

45.0

50.0

55.0

60.0

65.0

¡ 0.1 ¡ 0.1 ¡ 0.1 ¡ 0.1 ¡ 0.2 ¡ 0.3

The same circumferential temperature pro le was assumed at all cross sections relative to the corresponding tube side-wall temperature to obtain this equation. It is noted that the temperature variation along the tube is less than 1±C for all cases tested. A typical way of obtaining the heat transfer coef cients is rst to determine the local h based on local D T and then take an average of all the local h’s to determine the heat transfer coef cient. However, the local heat transfer coef cient or the local heat ux is not de ned here, which justi es our method to determine the heat transfer coef cients. All T-type thermocouples used were calibrated with a D.C. voltage/current standard generator (Yokokawa, 2553 ) and semiconductor probe. Especially careful calibration was done for the thermocouples to measure the temperatures at the inlet and outlet of the water stream. In this case, the maximum calibrated temperature obtained was ¡ 0.3±C at 65±C as shown in Table 1. The acquisition of data obtained from the T-type thermocouples was done by a Yokokawa recorder (HR 1310 ) connected to a PC. The data collected from each thermocouple in 2 min at 2-s intervals were averaged separately to make a data set. The uncertainties in the temperature measurements are less than §0.1±C, which yields a relative error of § 3.5% for the measurements of the wall superheat, (Tas ¡ Tsat ). The uncertainties in the ow rate measurements are approximately § 2.5%, so the calculated magnitudes of the heat uxes are accurate within §5%. Consequently the uncertainty in measurement of the heat transfer coef cient given in Eq. (1 ) was calculated to be within §8.5%. CALCULATION OF ELECTRIC FIELD IN A SINGLE MEDIUM Numerical calculation of the electric eld around the heating tube including electrodes in a single medium has been done in order to estimate the eld strength for dielectric breakdown and the direction of the dielectrophoretic force affecting test bubble behavior in liquid, possibly leading to a boiling heat transfer mechanism. Of course, the calculation of electric eld with consideration of the bubble distribution around the heater, which needs another extensive study, is not the vol. 21 no. 4 2000

37

purpose of this work. However, the qualitative behavior of bubble motion under an electric eld may be obtained by studying test bubbles in a system. Such numerical analysis to investigate the resulting dielectrophoretic plus buoyancy force distribution for a simple electrode system was done by Snyder et al. [15]. The governing equation which is appropriate for the system we tested is a 2-D Poisson equation. It is given by

³

¶

¶ u e x ¶ x

¶ x

´

+

¶ ¶ y

³

¶ u e y ¶ y

´

= ¡ q

f

(5 )

with boundary conditions at S1

= V1 u

and ¶ u =0 ¶ n

at S2

where S1 and S2 denote boundaries as shown in Figure 6. The components of the electric eld can be obtained from the electric potential, u . Ex = ¡

¶ u ¶ x

(6a )

Ey = ¡

¶ u ¶ y

(6b )

In this numerical calculation, we assume that there is no free charge (q f = 0 ) and the electric permittivity of the medium is constant in the domain considered. The computer program utilized in this study is MagNet 5, developed by In oytica Co., Canada, in 1994. The program employs the nite-element method (FEM) for electric eld calculation. Grid generation for the rst electrode con guration is shown in Figure 6. The number of nodes and triangular elements are 789 and 1,452, respectively, for this con guration. The same numerical scheme has been used for the various elec-

Figure 6 Gride systems of the rst and second electrode con gurations for FEM analysis.

38


Table 2

Electrical and thermodynamic properties of test liquids Liquid

Description

R11

Average molecular weight 137.37 Typical boiling point (±C) 23.77 Density, 25±C (g/cm3 ) 1.476 ¡ 111 Pour point, 1 atm (±C) Kinematic viscosity, 25±C 0.42 (C p ) Critical temperature (±C) 198.0 Critical pressure (atm) 43.2 Surface tension (dyn/cm) 22 Speci c heat (cal/g ±C) 0.208 Heat of vaporization (cal/g) 43.51 Thermal conductivity 0.074 (kcal/mhr ±C) Volume resistivity, 1.6414 £ 1011 (\ ¡ m) Dielectric strength (kV/mm) 11.6 Dielectric constant, 1 atm, 2.28 25±C Electrical charge 1.3 relaxation time (s)

R113

FC72

187.38 47.57 1.456 ¡ 35 0.66

340 56 1.68 ¡ 90 0.4

214.1 33.7 19 0.218 35.07 0.057

178 18.1 12 0.25 21 0.049

2.2

0.01

12.2 2.41

14.96 1.76

0.97

156.0

trode gaps. Electrical and thermodynamic properties of liquids needed in this calculation are shown in Table 2. Typical equipotential lines for the rst electrode con guration are shown in Figure 7. The arrows in this gure show the direction of the electric eld. In Table 3, the calculated values of the maximum electric eld strength between an electrode and the surface of the tube given applied voltage are shown. Thus, the maximum voltage applied to the electrode for dielectric breakdown

Figure 7

Equipotential lines for the rst electrode con guration.

vol. 21 no. 4 2000

Table 3 Predicted values of dielectric strength for the rst electrode con guration at different applied voltages Electrode gap

Applied voltage 5 kV 10 kV 15 kV 20 kV

5 mm

8 mm

3.3 kV/mm 6.6 kV/mm 9.9 kV/mm 13.2 kV/mm

2.8 kV/mm 5.6 kV/mm 8.4 kV/mm 11.2 kV/mm

can be estimated as shown in Table 4, since the dielectric strengths of liquids are known values. However, the actual values for the dielectric breakdown have been found to be lower than those calculated, because of the presence of impurities and thermally induced bubbles in the liquid [16]. The other important result obtained from this numerical calculation is detailed information on the action of the dielectrophoretic force in such a nonuniform eld. The electrohydrodynamic (EHD ) force by an electric eld acting on a unit volume of dielectric uid may be expressed as ³ ´ ´ ³ 1 2 1 2 ¶ e (7 ) fe = q f E ¡ E r e + r q E ¶ q T 2 2 where q is the uid density. The rst term represents Coulomb force, which is the force acting on the free change in an electric eld. The Coulomb force is negligible in this study because the current measured is of the order of microamperes, which means the absence of free charge in the uid. The second term stands for the dielectric force exerted on a dielectric uid due to spatial gradient on the permittivity. The resulting dielectrophoretic force for a vapor bubble is given by [17] as FE =

2p Db3 e f (e v ¡ e e v + 2e f

f

)

r E2

(8 )

where Db is bubble diameter. The third term in Eq. (7 ), which describes the electrostriction force, has no practical in uence on hydrodynamics for an incompressible uid [15]. Table 4 Predicted values of dielectric breakdown voltages for the rst electrode con guration for various liquids Electrode gap Liquid

5 mm

8 mm

R11 R113 FC72

17.57 kV 18.48 kV 22.66 kV

20.71 kV 21.78 kV 26.71 kV


Figure 8

Directions of dielectrophoretic force (FE ).

The direction of this force acting on a bubble is always opposite to the direction of r E because the permittivity of vapor inside a bubble, e v , is less than that of the medium, e f . The direction of dielectrophoretic force on a bubble is given in Figure 8. As can be clearly seen in this gure, only a bubble which exists in the center lines between the tube and electrodes moves toward the heat transfer surface. On the other hand, bubbles in other places are forced away from the heat transfer surface and toward the electrostatic stagnation point between electrodes, provided that no other forces are acting on the bubble. Possible motions of bubbles generated on the heating surface under the buoyancy and electrophoretic forces combined are shown clearly in Figure 9. The bubbles generated on the lower side of the tube move toward and coalesce together near the electrostatic stagnation point, where the buoyancy force acting on bubbles is balanced with the EHD one. Consequently, the rst electrode con guration which provides a electrostatic stagnation point for bubble motion may reduce the heat transfer rate. On the other hand, bubbles generated in other places move radially rst and quickly rise up due to buoyancy. Similar conclusions for the bubble motion depending on the dieletrophoretic and buoyant forces were obtained in pool boiling under heat ux control [11]. However, quite different bubble behavior under the nonuniform electric eld generated by the meshtype electrode is plausible [1]. RESULTS AND DISCUSSION Experiments of nucleate boiling phenomena in a nonuniform electric eld have been done to investigate the vol. 21 no. 4 2000

39

Figure 10 Variation of heat transfer coef cients for R11 with six electrodes (second con guration) at Ts = 25.5±C.

Figure 9 Expected bubble behavior for different electrode con gurations.

nucleate boiling heat transfer depending on the eld strength, the wall superheat, as well as electrode con gurations employed. Some experimental results obtained are as follows. All the data obtained in this study were taken during the heating-up mode. For R11, boiling heat transfer coef cients as a function of the superheat with different applied voltages are shown in Figure 10. The electrode type employed in this case is the second con guration. The heat transfer coef cients increase steadily as the applied voltage increases for all ranges of wall superheat tested. The maximum heat transfer enhancement for R11 is about 40


130% at a wall superheat of 10.3 K. This gure also shows that the enhancement magnitude decreases if the wall superheat exceeds 12 K, regardless of the applied voltage. This may be due to the fact that boiling action by buoyant force is suppressed in the presence of electric eld. The variation of the circumferential temperatures at the midsection of the tube for different heat uxes at zero voltage and an applied voltage of 5 kV are shown in Figures 11 and 12. As shown in Figure 11, the temperature drop at the tube wall, especially at the bottom of the tube, is noticeable as nucleate boiling starts. The arrows in Figures 11 – 14 indicate the onset of boiling. However, such temperature drop due to boiling onset is eliminated completely when the moderate voltage of 5 kV is applied. This may be due to the action of the dielectrophoretic force, which helps detachment of small bubbles from the boiling surface early in their process of growth [15]. It is also noted that once nucleate boiling occurs on the heating surface, the circumferential temperature variation of the test tube appears distinctly, which indicates that the electric eld induces and promotes nucleate boiling on the upper surface. Similar circumferential temperature pro les were observed at high voltages with slight decrease in wall temperatures. Somewhat different results have been obtained for R113. For R11, steady increase in the heat transfer coef cient depending on the wall superheat has been vol. 21 no. 4 2000

Figure 11 Variation of tube wall and inlet water temperature for different heat uxes in R11 with six electrodes (second con guration) at Ts = 25.5±C (the arrow indicates the onset of boiling).

Figure 12 Variation of tube wall and inlet water temperature for different heat uxes in R11 with six electrodes (second con guration) at Ts = 25.5±C (the arrow indicates the onset of boiling).


Figure 13 Variation of heat transfer coef cients for R113 with six electrodes ( rst con guration) at Ts = 48.5±C (the arrow indicates the onset of boiling).

Figure 14 Variation of heat transfer coef cients for R113 with six electrodes (second con guration) at Ts = 48.5±C (the arrow indicates the onset of boiling).

vol. 21 no. 4 2000

41

observed. On the other hand, abrupt increase when an electric eld is applied and thereafter no appreciable dependence on the wall superheat in the heat transfer coef cients at lower heat ux levels have been found for R113. This is similar to the boiling behavior observed for R123 under an electric eld [13]. Also, the observed heat transfer coef cient value for R113 is comparable to the value measured for R123, which is about 1,600 W/m2 K with application of 19 kV with the temperaturecontrol method [13]. However, the boiling heat transfer coef cient with electric eld applied is lower than that for zero eld if the wall superheat exceeds 13 K, as shown in Figures 13 and 14. This can be attributed to the fact that boiling action due to the buoyancy-driven force is suppressed by the EHD force in this region, where many bubbles form on the heating surface. In other words, the paths available for the bubble to depart are restricted [11] more severely if many bubbles are present. It is well known that the superheat required to nucleate from a vapor- lled cavity is dependent on the following nucleation parameter [18]: Nucl =

r Tsat q v h fg

(9 )

The ratio of this parameter for R11 to R113 is 1.1, which means that the superheats required for R11 are slightly greater than those required for R113, or less superheat needed for nucleation produces more bubble for R113. Much more bubble formation than prediction at lower superheat was observed for R113 in this experiment. Less superheat for nucleation and shorter charge relaxation time may be the main reasons why the heat transfer coef cient for R113 is quite different from the other uids in the lower wall superheat region under electric eld. The maximum heat transfer enhancement obtained for R113 is about 180%. Similarly as for R11, the magnitude of heat transfer enhancement decreases very much with increasing wall superheat for R113, with the same reason. In the case of the rst electrode con guration, the boiling heat transfer coef cient with an applied voltage of 10 kV decreases even when the wall superheat is greater than 10 K. This may be due to the fact that the dielectrophoretic force makes bubbles move to the electrostatic stagnation region just below the lower side of the tube so that free movement of the bubbles is hindered very much, which shows clearly how the nonuniform electric eld around the boiling surface affects the bubble motion and, in turn, the heat transfer mechanism. However, the heat transfer coef cient increases again when the buoyancy-driven force dominates the EHD one at high heat uxes. Similar results were obtained for 42


the EHD system employing eight wire electrodes along the heater surface [11]. A static vapor line of bubble accumulation was observed between the three bottom electrodes, which caused lower heat transfer coef cient at higher voltage and higher heat ux [11]. On the other hand, the bubbles generated at the lower side of the tube move outward radially and rise by buoyancy for the case of the second electrode, as shown graphically in Figure 9. With zero eld, a temperature drop after boiling onset was also observed for R113. Again, moderate applied voltage is enough to eliminate the temperature drop for the second electrode con guration. However, there was no appreciable difference in temperature at the bottom, side, and upper surfaces of the tube, which suggests that the boiling is active for all circumferences of the tube at relatively low heat uxes. It should be noted that the slope of the heat transfer coef cients obtained for R11 and R113 at zero eld in this study are lower than those obtained with the heat ux control method [13, 19]. As shown in Figure 15, the boiling heat transfer coef cients for FC72 are insensitive to the applied electric voltages. In this case boiling starts below the wall superheat of 6±C. This can be attributed to the longer charge relaxation time than the bubble departure one, so that the bubble behavior is no longer in uenced by the electric eld [9]. The fact that the electric charge relaxation time is the time taken for an electric eld to affect a

Figure 15 Variation of heat transfer coef cients for FC72 with six electrodes (second con guration) at Ts = 57.5±C (boiling starts below the wall superheat of 6±C in this case).

vol. 21 no. 4 2000

Figure 16

Effect of applied voltage on bubble behavior in R113 with 12 electrodes at Ts = 48.5 ±C.

bubble in the eld may be deduced by Coulomb’s law and charge conservation [20]. Even though the FC72 requires lower superheats to nucleate than R113, much longer charge relaxation time yields the results mentioned above. In fact, for FC-72 the charge relaxation time of 1.56 £ 102 s as shown in Table 2 is much longer than the bubble departure time of 2 £ 10 ¡ 2 s. On the other hand, remarkable boiling heat transfer enhancement has been obtained up to threefold increase for heat transfer engineering

HCFC-123, which has an electrical charge relaxation time of 0.89 £ 10 ¡ 3 s. More detailed discussion on the experimental results for the subject was given by Han et al. [21]. Flow visualization of nucleate boiling for R113 at a surface heat ux of about 30 kW/m 2 and several voltage levels is shown in Figure 16. From the results in Figure 16, it is clear that as the applied voltage increases, the number of bubbles increases while the diameter of vol. 21 no. 4 2000

43

NOMENCLATURE

bubbles decreases, which has been observed previously for R11 [9] and for R123 [10]. The electrode con guration employed in this experiment is the combination of the rst electrode con guration with 5-mm gap and the second one with electrodes 8 mm away from the heating surface. At higher voltage, the bubbles are pulled out from the heating surface radially to form a sheet of bubbles between the lower side of the tube and the electrodes. This observation, referred as coalescing effect [9], results from the dielectrophoretic force acting on the bubble at the bottom part of the boiling surface [15], as discussed before.

A Cp Db E fe

CONCLUSIONS

FE h h fg mÇ QÇ Rm T Tas V

Experimental and numerical studies were performed to investigate the bubble behavior and nucleate boiling heat transfer with a wide range of wall superheats with the temperature-control method under a nonuniform electric eld produced by wire electrodes. The conclusions obtained in this study are as follows. 1. When a moderate voltage of 5 kV was applied, the temperature drop disappeared, which indicates that the applied electric eld induces and promotes bubble nucleation on heating surface, although the cause has not been found. 2. EHD augmentation on nucleate boiling heat transfer is crucially dependent on working uids. For R11, EHD enhancement in nucleate boiling heat transfer was observed for all ranges of wall superheat tested. However, the enhancement of the boiling heat transfer disappeared if the wall superheat exceeded 13±C for R113, and no electric eld effect on the boiling heat transfer was observed for FC72. The maximum heat transfer enhancements obtained were about 130% for R11 and 180% for R113 at a wall superheat of around 10 K. 3. Proper electrode con guration is needed to induce buoyancy-driven bubble motion under an electric eld which, in turn, enhances the boiling heat transfer. Numerical study has revealed that the bubbles are forced away from the heating surface and toward the electrostatic stagnation point produced by the dielectrophoretic force due to nonuniform electric eld generated by wire electrode. In conclusion, proper electrode con guration, appropriate electric eld strength, and wall superheat are all important factors to enhance the nucleate boiling heat transfer as well as to prevent reducing the electric eld strength for dielectric breakdown, which should be considered for EHD-augmented evaporator design and operation. 44


e q

q f

r u

heat transfer area, m2 (ft2 ) speci c heat of water, kJ/kg K (Btu/lb m R ) bubble diameter, m (ft ) electric eld strength, V/m electrohydrodynamic force per unit volume, N/m3 (lbf /ft3 ) dielectrophoretic force, N (lbf ) heat transfer coef cient, W/m2 K (Btu/hr ft2 R ) heat of vaporization, kJ/kg (Btu/lb m ) mass ow rate of water inside tube, kg/s (lbm /hr ) heat ow rate, kJ/s (Btu/hr ) weighting factor de ned by Eq. (4 ) temperature, K (R ) average temperature of tube wall, K (R ) electric potential, V permittivity constant, C2 /N2 m2 uid density, kg/m 3 (lbm /ft3 ) free charge density, C/m 3 surface tention, N/m (lbf /ft ) electric potential, V

Subscripts f sat v wi wo

liquid saturated state vapor water stream at the inlet of tube water stream at the outlet of tube

REFERENCES [1] Cooper, P., EHD Enhancement of Nucleate Boiling, ASME J. Heat Transfer, vol. 112, pp. 458 – 464, 1990.

[2] Bochirol, L., Bonjour, E., and Weil, L., Etude de l’action de’

[3] [4] [5] [6] [7]

[8]

champs elestriques sur les transferts de chaleur dans les liquides bouillants, C. R. Hebd, Seances Acad. Sci. (Paris), pp. 76 – 78, 1960. Choi, H. Y., Electrohydrodynamic Boiling Heat Transfer, Ph.D. thesis, Massachusetts Institute of Technology, Cambridge, MA, 1960. Markels, M., and Durfee, R. L., The Effect of Applied Voltage on Boiling Heat Transfer, AIChE J., vol. 10, pp. 106 – 110, 1964. Johnson, R. L., Effect of an Electric Field on Boiling Heat Transfer, AIAA J., vol. 6, pp. 1456 – 1460, 1968. Bonjour, E., and Verdier, J., Mecanisme de l’ebullition sous champ electrique, C. R. Hebd, Seances Acad. Sci. (Paris), pp. 924 – 926, 1960. Ogata, J., and Yabe, A., Augmentation of Boiling Heat Transfer by Utilizing the EHD Behavior of Boiling Bubbles and Heat Transfer Characteristics, Int. J. Heat Mass Transfer, vol. 16, pp. 783 – 791, 1993. Basu, D. K., Effect of Electric Fields on Boiling Hysteresis in Carbon Tetrachloride, Int. J. Heat Mass Transfer, vol. 16, pp. 1322 – 1424, 1973.

vol. 21 no. 4 2000

[9] Kawahira, H., Kubo, Y., Yokoyama, T., and Ogata, J., The

[10]

[11] [12] [13]

[14] [15] [16]

[17] [18] [19]

Effect of an Electric Field on Boiling Heat Transfer of Refrigerant-11; Boiling on a Single Tube, IEEE Trans. Ind. Appl., vol. 26, pp. 359 – 365, 1990. Ohadi, M. M., Paper, R. A., Ng, T. L., Faani, M. A., and Radermacher, R., EHD Enhancement of Shell-Side Boiling Heat Transfer Coef cients of R-123/Oil Mixture, ASHRAE Trans. Symp., pp. 427 – 434 (BA-92-5-1), 1992. Seyed-Yagoobi, J., Geppert, C. A., and Geppert, L. M., Electrohydrodynamically Enhanced Heat Transfer in Pool Boiling, J. Heat Transfer, vol. 118, pp. 233 – 237, 1996. Incropera, F. P., and DeWitt, D. P., Introduction to Heat Transfer, Wiley, New York, 1985. Singh, A., Kumar, A., Dessiatoun, S., Fanni, M. A., Ohadi, M. M., and Ansari, A. I., Compound EHD-Enhanced Pool Boiling of R-123 in a Liquid-to-Refrigerant Heat Exchanger, ASME Paper 93-WA/HT-40, 1993. Stephan, K., and Abdelsalam, M., Heat Transfer Correlations for Natural Convective Boiling, Int. J. Heat Mass Transfer, vol. 23, pp. 73– 87, 1980. Snyder, T. J., Schneider, J. B., and Chung, J. N., A Second Look at Electrokinetic Phenomena in Boiling, J. Appl. Phys., vol. 79, pp. 6755 – 6760, 1996. Hara, M., Honda, K., and Kaneko, T., D.C. Electrical Breakdown of Saturated Liquid Helium at 0.1 MPa in the Presence of Thermally Induced Bubbles, Cryogenics, vol. 27, pp. 567 – 576, 1987. Baboi, N. F., Bologa, M. K., and Klyukanov, A. A., Some Features of Ebullition in Electric Field, Appl. Electr. Phenom. (USSR), vol. 20, pp. 57– 70, 1968. Bergles, A. E., and Rohsenow, W. M., The Determination of Forced Convection Surface-Boiling Heat Transfer, J. Heat Transfer, vol. 86, pp. 365 – 372, 1964. Marto, P. J., and Lepere, V. J., Pool Boiling Heat Transfer from Enhanced Surface to Dielectric Fluids, ASME J. Heat Transfer, vol. 104, pp. 292 – 299, 1982.


[20] Panofsky, W. K. H., and Phillips, M., Classical Electricity and Magnetism, Addison-Wesley, New York, 1962.

[21] Han, S., Na, M., Oh, S., and Kwak, H., Electrohydrodynamic (EHD) Enhancement of Boiling Heat Transfer with a LoFin Tube, HTD-Vol 351, Proc. ASME Heat Transfer Division, vol. 1, pp. 223 – 233, 1997.

Si-Doek Oh received the B.S., M.S., and Ph.D. degrees in Mechanical Engineering from ChungAng University, Korea, in 1982, 1985, and 1995, respectively. Since 1985, he has been with the Research & Development Institute, Hyosung Cooperation, Seoul, Korea, where he is now a general manager. Also, he is currently an Adjunct Professor in the Department of Mechanical Engineering, Chung-Ang University. He is a consultant to industry in areas related to energy saving. His research interests include optimal planning and design of energy systems such as heat recovery and energy-saving systems, co-generation systems, and CNG fueling systems.

Ho-Young Kwak received the B.S. degree in Physics from Seoul National University in 1971 and the M.A. in Plasma Physics and Ph.D. in Mechanical Engineering from the University of Texas at Austin, USA, in 1977 and 1981, respectively. He joined the Mechanical Engineering Faculty, Chung-Ang University, Korea, in 1981 and is currently a professor there. Now he is serving as Dean of College of Engineering. His research interest are in bubble nucleation, bubble dynamics, sonoluminescene phonemena, boiling heat transfer, electronic equipment cooling, and exergetic and exergoeconomic analysis for thermal systems. He is a member of the KSME, ASME, KPS, APS, and ASA.

vol. 21 no. 4 2000

45