Subcooled flow boiling heat transfer of R-134a and bubble characteristics in a horizontal annular duct

Subcooled ¯ow boiling heat transfer of R-134a and bubble

characteristics in a horizontal annular duct

Chih-Ping Yin

_{, Yi-Yie Yan}

_{, Tsing-Fa Lin}

_{*, Bing-Chwen Yang}

b_{Energy and Resources Laboratories, Industrial Technology Research Institute, Hsinchu, Taiwan}

Received 8 June 1999; received in revised form 6 September 1999

Abstract

Experiments were carried out to investigate the subcooled ¯ow boiling heat transfer and visualize the associated bubble characteristics for refrigerant R-134a ¯owing in a horizontal annular duct having inside diameter of 6.35 mm and outside diameter of 16.66 mm. The e€ects of the imposed wall heat ¯ux, mass ¯ux, liquid subcooling and saturation temperature of R-134a on the resulting nucleate boiling heat transfer and bubble characteristics were examined in detail. In the experiment signi®cant hysteresis was noted in the boiling curves during the onset of nucleate boiling (ONB) especially at low saturation temperature and high subcooling. The temperature undershoot at ONB is rather large for most cases. The boiling heat transfer was slightly higher for a lower saturation temperature and was little a€ected by the mass ¯ux. However, for a higher subcooling of the refrigerant better heat transfer results. Furthermore, the ¯ow visualization revealed that at higher imposed wall heat ¯ux the heated surface was covered with more bubbles and the bubble generation frequency is higher. But the size of the bubbles departing from the heated surface was only slightly a€ected by the imposed heat ¯ux. At high mass ¯ux and subcooling the bubble generation was suppressed to a noticeable degree. Besides, the bubbles are much smaller at a higher subcooling. Finally, empirical correlations for the heat transfer coecient and bubble departure diameter in the subcooled ¯ow boiling of R-134a were proposed. # 2000 Elsevier Science Ltd. All rights reserved.

1. Introduction

In view of the quick destruction of the ozone layer in the outer atmosphere around the earth the pro-duction of certain CFC and HCFC refrigerants was prohibited recently or will be limited in the near future. New types of refrigerants with zero ozone-depleting potential such as HFC and HC refrigerants have been developed and some are widely used. The heat transfer

and pressure drop data for boiling, evaporation and condensation of these new refrigerants are desired to facilitate the design of various air conditioning and re-frigeration systems. Recently, e€orts had been made by a number of research groups to establish the design database for some new refrigerants for di€erent enhanced surfaces. Moreover, R-134a is currently con-sidered to be the main replacement to many CFC and HCFC refrigerants. In the present study the subcooled ¯ow boiling of R-134a in an annular duct was exper-imentally investigated by measuring the boiling curves and visualizing the bubbles on the heated surface.

An updated comprehensive review of the literature on boiling of new refrigerants was recently

con-0017-9310/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved. PII: S0017-9310(99)00278-1

* Corresponding author. Tel.: 3-7712-121; fax: +886-3-5720-634.

ducted by Thome [1]. In the following the relevant literature on the present study is brie¯y reviewed.

In the past few decades some studies were reported in the literature on the subcooled ¯ow boiling of water and the associated bubble departure size. Gunther [2] performed a photographic study of boiling in a rectangular channel with a heated strip in the middle of the channel for the liquid subcool-ing rangsubcool-ing from 15 to 658C, mean ¯ow speed from 1.5 to 12.2 m/s and pressure from 1 to 11 bars. He found that at high subcooling …DTsub>

388C† bubbles were hemispherical which grew and then collapsed while sliding along the heater but did not detach from the heated wall. He also noted that the bubble sliding velocity was approximately 80% of the mean ¯ow velocity. In a heated annulus McAdams et al. [3] reported that the moving of the bubbles into the liquid core caused violent agitation of the liquid near the heated surface and enhanced the heat transfer rate. Recently, Bibeau and Salcu-dean [4] visualized the bubble cycling in a vertical annular pipe with a high speed photography. They observed that the maximum bubble diameter varied between 0.8 and 3.0 mm and at ejection the bubbles were smaller than the largest bubbles in the liquid, since the ejecting bubbles slided and con-densed on the heated wall.

Some studies also exist in the literature on the sub-cooled ¯ow boiling of refrigerants. Hasan et al. [5]

measured the subcooled ¯ow boiling of R-113 in a ver-tical annular channel with an electrical heated inner pipe and an outer Pyrex glass pipe. They found that the boiling heat transfer coecient based on the imposed heat ¯ux divided by the di€erence of the wall temperature and time averaged mixed-mean tempera-ture of the liquid was lower for a higher pressure and for a higher subcooling. Moreover, for a higher mass ¯ux the boiling heat transfer coecient was mildly lower. Their data agreed well with some correlations [6,7] but showed poor agreement with the modi®ed Chen correlation [8]. Subcooled ¯ow boiling of R-11, R-123 and two R-123/alkybenzene lubricant mixtures in a horizontal circular quartz pipe was investigated by Kedzierski [9]. The pipe was heated by a brass strip of 3 mm wide and 0.25 mm thick placed horizontally along the pipe bottom with its length aligned with the ¯ow direction. He noted that the boiling heat transfer coecient of R-123 was about 22% higher than R-11 because the boiling in R-123 had approximately 10 more active nucleation sites/cm2_{. Moreover, adding a}

small amount of lubricant into R-123 was found to produce a signi®cant number of new and active nuclea-tion sites and hence can enhance the boiling heat trans-fer. But a further addition of the lubricant resulted in an opposite trend. Furthermore, he found that the bubble diameter was independent of the Reynolds number of the mean ¯ow and the imposed heat ¯ux. The measured bubble diameter was in good agreement Nomenclature

Bo boiling number …ˆ q00 w=Gifg†

dp bubble departure diameter, m

Dh hydraulic diameter …ˆ Doÿ Di†, m

Di inner heated pipe outside diameter, m

Do outer Pyrex inside diameter, m

g gravitational acceleration, m/s2

G mass ¯ux, kg/m2_s

hf speci®c enthalpy of liquid at saturation

tem-perature, J/kg

hi speci®c enthalpy of subcooled liquid at inlet,

J/kg

hl all-liquid nonboiling heat transfer coecient,

W/m2_8C

hr boiling heat transfer coecient, W/m28C

ifg latent heat of vaporization at saturation

tem-perature, J/kg

k thermal conductivity, W/m 8C L length of heated pipe, m Nsub subcooling number ‰ˆ …hfiÿhfgi†…

vgÿvf

vf †Š

P system pressure, mm Hg q00

w imposed wall heat ¯ux, W/m2

T temperature, K u ¯ow velocity, m/s

vf speci®c volume of liquid, m3/kg

vg speci®c volume of vapor, m3/kg

Greek symbols

DTsat wall superheat (=Twÿ Tsat), 8C

DTsub subcooling ……ˆ Tsatÿ Tf†, 8C

b dynamic contact angle (8) r density, kg/m3 m viscosity, Ns/m2 s surface tension, N/m Subscripts f liquid g vapor i inlet/ inner r refrigerant sat saturated sub subcooling w heated wall

with the Fritz equation [10]. Klausner et al. [11] measured the bubble departure diameter for the satu-rated ¯ow boiling of R-113 in a rectangular channel with an electrically heated nichrone plate of 20 mm wide and 0.127 mm thick on the channel bottom. They found that the bubble departure diameter is smaller for a higher mass ¯ow rate and for a lower imposed heat ¯ux. They also noted that before lifting o€ from the heated wall, the bubbles would slide a ®nite dis-tance along the surface. They concluded that not only the surface tension force but also the asymmetrical bubble growth acting in the direction opposite to the ¯uid motion were important in holding the bubbles at the nucleation sites before departure.

Subcooled ¯ow boiling heat transfer of heptane on a resistance-heated coiled wire of diameter 1.25 mm con-tained in a concentric annulus was examined by Mul-ler-Steinhagen et al. [12]. Their results indicated that the boiling heat transfer coecient increased with increasing heat ¯ux but decreased with increasing sys-tem pressure and subcooling, while independent of the mass ¯ux. The hysteresis in the boiling curves was small. Similar study carried out by Bland [13] showed that the hysteresis decreased at a higher ¯ow velocity.

Experimental measurements and empirical corre-lations for the size of the departure bubbles in the saturated boiling were carried out and proposed by Cole and Shulman [14]. They measured the bubble departure size in various liquids at subatmospheric pressure. Speci®cally, they obtained data for toluene at 48 mm Hg, acetone at 222 and 461 mm Hg, carbon tetrachloride at 138 mm Hg, n-pentane at 524 mm Hg, methanol from 134 to 540 mm Hg and water from 50 to 360 mm Hg. Their results indicated that the Fritz equation [10] dpˆ 0:0208b  _s g…rfÿ rg† 1=2 …1† was applicable only at the atmospheric pressure. A modi®ed correlation was proposed in their study as

dp  _s g…rfÿ rg† 1=2 ˆ 1000 P …2†

The above literature review clearly indicates that the experimental data for the subcooled ¯ow boiling heat transfer in channels and the associated bubble charac-teristics for the widely used new refrigerant R-134a are not available. In this study, the subcooled ¯ow boiling of R-134a in an annular duct was explored experimen-tally. The e€ects of the imposed heat ¯ux, mass ¯ux, subcooling and saturation temperature (pressure) of R-134a on the subcooled boiling heat transfer character-istics were investigated. Meanwhile, ¯ow visualization

was conducted to examine some boiling characteristics such as the distribution of the bubble size and the den-sity of the nucleation sites. Moreover, the hysteresis existing in the subcooled nucleate boiling was inspected.

2. Experimental apparatus and procedures

The experimental apparatus established in the pre-sent study to investigate the subcooled boiling of R-134a, schematically shown in Fig. 1, consists of two main loops, namely, the refrigerant and water±glycol loops, and a data acquisition system. Refrigerant R-134a is circulated in the refrigerant loop. We need to control the temperature and ¯ow rate in the water±gly-col loop to have enough cooling capacity for conden-sing the R-134a vapor and subcooling the R-134a liquid to a preset subcooled temperature. A DC power supply is used for heating the inner pipe in the test sec-tion. For measuring the bubble size and nucleation density in the R-134a subcooled ¯ow boiling on the heated surface, a camera connected to a photographic microscope is set up beside the test section to observe the boiling ¯ow.

2.1. Refrigerant loop

The refrigerant loop contains a refrigerant pump, an accumulator, a mass ¯ow meter, a test section, a con-denser, a sub-cooler, a receiver, a ®lter/dryer and three sight glasses. A rotational DC motor is used to drive the refrigerant pump. Through changing the DC cur-rent in the motor the liquid ¯ow rate of R-134a can be varied. By adjusting the gate opening of the valve at the downstream of the test section, the pressure of the refrigerant in the test section can be regulated. Note that the refrigerant ¯ow rate and the pressure should be further adjusted simultaneously in order to control them at the required levels. Moreover, the accumulator is installed to dampen the ¯uctuations of the ¯ow rate and pressure. The refrigerant ¯ow rate is measured by a mass ¯ow meter with an accuracy of 1%. Mean-while, a condenser and a subcooler are used to con-dense the R-134a vapor leaving the test section and then subcool the liquid R-134a. Varying the tempera-ture and ¯ow rate of the water±glycol mixtempera-ture moving through the condenser and subcooler allows us to con-trol the bulk temperature of R-134a leaving the sub-cooler. After subcooled, the liquid R-134a ¯ows back to the receiver.

2.2. Test section

a horizontal annular duct with the outer pipe made of pyrex glass to facilitate the visualization of boiling pro-cesses. The outer pipe has an inside diameter of 16.66 mm and is 4 mm thick. A circular, electrically heated SS-304 inner pipe is used as the heating surface and the subcooled R-134a liquid enters the annular passage

¯owing over the heating surface. This heated inner pipe has an outside diameter of 6.35 mm and it is 21 cm long and 1 mm thick. This stainless steel pipe was soldered with two pure copper pipes of the same size at their ends. Careful procedures have been taken to insure that these three pipes are connected intactly and

Fig. 1. Schematic diagram of the experimental system.

are like a single straight pipe. The copper pipes allow the DC current to deliver from the power supply to pass through the SS-304 pipe.

To measure the temperature of the heating surface, the inner surface of the SS-304 pipe is covered with a thin high conductivity thermal bond (k = 200 W/m 8C), as indicated in Fig. 2. Then 12 T-type thermo-couples are ®xed onto the thermal bond so that the voltage signals from the thermocouples are not inter-fered by the DC current passing through the heating surface. The thermocouples are positioned at three axial stations. At each axial station four thermocouples are placed at the top, bottom and two sides of the cir-cumference with 908 apart. Moreover, the thermal bond is further covered by a te¯on pipe of 0.3 mm thick to reduce the heat loss from the SS-304 pipe. Note that the temperature of the boiling surface can be evaluated from the thermocouple data.

A DC power supply with a maximum rating of 20 kW provides the electric current passing through ®rst to the copper pipe and then to the stainless steel pipe. The DC current passing through the heating pipe is measured by a Yokogawa DC meter with an accuracy of 1%. Moreover, the voltage drop across the heating pipe is measured by a Yokogawa multimeter and the power input to the heating pipe can then be calculated. The refrigerant R-134a pressure is measured by a pressure transducer with an accuracy of 1%.

2.3. Water±glycol mixture loop

The water±glycol loop designed for condensing the R-134a vapor and for subcooling the liquid R-134a contains a 125-litre constant temperature bath with a water cooled refrigeration system. The cooling capacity is 2 kW for the water±glycol mixture at ÿ208C. The cold water±glycol mixture at a speci®ed ¯ow rate is driven by a 0.5 hp pump to the condenser as well as to the subcooler. A by-pass loop is provided to adjust the ¯ow rate. By adjusting the mixture temperature and ¯ow rate, the bulk temperature of the R-134a in the subcooler can be controlled at a preset level.

2.4. Data acquisition

The data acquisition system for recording and pro-cessing the signals from the thermocouples and press-ure transducers includes a recorder and a controller. The recorder is used to record the temperature and voltage data. The IEEE488 interface is used to connect the controller and the recorder, allowing the measured data to transmit from the recorder to the controller and then analyzed by the computer immediately.

2.5. Experimental

In the test the subcooled R-134a liquid at the inlet of the test section is ®rst maintained at a speci®ed sub-cooling by adjusting the water±glycol temperature and ¯ow rate. Then, we regulate the R-134a pressure at the test section inlet by adjusting the gate opening of the valve locating right after the exit of the test section. Meanwhile, by changing the current of the DC motor connecting to the refrigerant pump, the R-134a ¯ow rate can be adjusted. The imposed heat ¯ux from the heater to the R-134a is adjusted by varying the electric current delivered from the DC power supply. By measuring the current and voltage drop across the hea-ter, we can calculate the amount of the heat transfer rate to the refrigerant.

Generally and also in the present study, the boiling heat transfer coecient is de®ned as the average imposed heat ¯ux divided by the wall superheat Twÿ

Tsat: For some cases the subcooled ¯ow boiling heat

transfer coecient is de®ned here as the heat ¯ux divided by the di€erence of the wall and bulk tempera-tures, especially when discussing the subcooling e€ects [5].

2.6. Uncertainty analysis

The uncertainties of the experimental results are analyzed by the procedures proposed by Kline and McClintock [15]. The results from this analysis indicate that the uncertainties associated with the temperature and pressure measurements are respectively 0.38C and 0.002 MPa. While for the imposed heat ¯ux and mass ¯ux of R-134a the uncertainties are about 3 and 2%, respectively. The uncertainty for the boiling heat trans-fer coecient is about 16%. The detailed results from the present uncertainty analysis for the experiments conducted here are summarized in Table 1.

Table 1

Parameters and estimated uncertainties

Parameter Uncertainty Length, width and thickness (m) 20.5%

Area (m2_{) 21%}

Temperature, T (8C) 20.28C

DT (8C) 20.38C

Pressure, P (MPa) 20.002 MPa Mass ¯ux of refrigerant, G 22% Heat ¯ux, q00

w 23%

3. Results and discussion

In the present study experiments were conducted for the refrigerant mass ¯ux varying from 100 to 300 kg/ m2_{s, imposed heat ¯ux from 0 to 20 kW/m}2_,

subcool-ing from 6 to 138C and saturation temperature from ÿ6 to 18C. In what follows selected data are reported to illustrate the characteristics of the subcooled ¯ow boiling heat transfer for R-134a in an annular duct by presenting the nucleate boiling curves for various cases. E€ects of the mass ¯ux, heat ¯ux, subcooling and saturation temperature on the heat transfer data will be examined in detail. In addition, the associated bubble characteristics in the subcooled boiling on the heated surface will be inspected based on the results from the ¯ow visualization.

3.1. Boiling curves

E€ects of the mass ¯ux on the R-134a subcooled ¯ow boiling heat transfer are manifested ®rst in Figs. 3 and 4 for di€erent subcooling and saturated tempera-tures. An inspection of a given boiling curve for a ®xed mass ¯ux reveals that as the imposed heat ¯ux is raised gradually from an unheated state, the tempera-ture of the heated wall Tw increases slowly. This

increase in Twis almost linear, suggesting that no

boil-ing takes place in the R-134a liquid ¯ow and the heat transfer in the ¯ow is due to single phase forced con-vection. It is of interest to note that this single phase heat transfer region extends signi®cantly to the wall temperature well above the saturated value Tsat. The

wall superheat, Twÿ Tsat, can be as high as 188C for G

= 300 kg/m2_{s in Fig. 3 with DT}

subˆ 10 8C and Tsat=

ÿ28C. It is further noted that at a certain high wall superheat a very small raise of the imposed heat ¯ux causes the heated wall temperature to drop substan-tially, obviously due to the sudden appearance of the boiling on the wall. Thus, we have onset of nucleate boiling (ONB) at this wall heat ¯ux. This temperature undershoot during ONB is rather signi®cant in the subcooled ¯ow boiling of R-134a according to the present data. Note that at a high liquid subcooling for DTsubˆ 10 8C the temperature undershoot is larger for

a high R-134a mass ¯ux (see Fig. 3). But at a lower subcooling for DTsubˆ 6 8C the situation is di€erent

(see Fig. 4). More speci®cally, at the lowest G of 100 kg/m2 _{s the temperature undershoot is the largest.}

While at the higher G of 200 and 300 kg/m2 _{s the}

undershoot is nearly at the same magnitude. In the single phase region the wall heat ¯ux at a given wall superheat is higher for a higher mass ¯ux. This indi-cates that an increase in the mass ¯ux results in better heat transfer. But in the nucleate boiling region beyond ONB the R-134a mass ¯ux shows negligible e€ects on the boiling heat transfer from the heated wall. It should be pointed out that when the imposed heat ¯ux is lowered gradually from a high level at which the boiling on the heated wall is rather intense, the nu-cleate boiling can be maintained at a very low wall superheat. We have boiling even at a wall superheat of 28C (see Fig. 3). This clearly illustrates the existence of the signi®cant hyteresis in the boiling curves.

Next, the R-134a subcooled ¯ow boiling curves for various saturation temperatures with Tsat = ÿ55, ÿ2

and 18C shown in Fig. 5 for G = 200 kg/m2 _{s and}

Fig. 4. Comparison of the subcooled ¯ow boiling curves for G = 100, 200 and 300 kg/m2_{s at T}

satˆ 18C and DTsubˆ 68C.

Fig. 3. Comparison of the subcooled ¯ow boiling curves for G = 200 and 300 kg/m2_{s at T}

DTsubˆ 6 8C indicate that the temperature undershoot

is much larger for the low Tsat of ÿ58C. But for Tsat

= ÿ2 and 18C the corresponding temperature under-shoots are nearly the same. While in the nucleate boil-ing region the e€ects of the saturation temperature on the boiling heat transfer is slight. The boiling heat ¯ux for Tsat = ÿ58C is only slightly higher than those for

Tsat= ÿ2 and 18C.

The e€ects of the subcooling are shown in Fig. 6 for DTsubˆ 6 and 138C at G = 200 kg/m2 s and Tsat =

18C. The results indicate that at a higher subcooling

the temperature undershoot is larger. Besides, the single phase and boiling heat transfer is substantially higher for a higher subcooling.

Then, the present data were compared with those evaluated from the Shah correlations [6] for the sub-cooled ¯ow boiling in annuli. In that study Shah has proposed correlations for saturated and for subcooled ¯ow boiling mainly for water, methanol and R-113. Among which the correlation proposed for the high subcooling is compared with the present data here. His correlation is hrˆ hl  230
Bo0:5_‡DTsub DTsat  …3† here hlis the all-liquid nonboiling heat transfer

coe-cient and is calculated from the McAdams equation as hlDh kf ˆ 0:023 _GD h m 0:8 Pr0:4 _4†

Fig. 5. Comparison of the subcooled ¯ow boiling curves for Tsatˆ ÿ5, ÿ2 and 18C at G = 200 kg/m2s and DTsubˆ 68C.

Fig. 6. Comparison of the subcooled ¯ow boiling curves for DTsubˆ 6 and 138C at G = 200 kg/m2s and Tsatˆ 18C.

Fig. 7. Comparisons of the present data with the Shah corre-lation for q00

w from about 5000 to 20 000 W/m2 at (a) G =

Fig. 7 shows the comparison between the present data and Shah correlation. Note that the average deviation is about 16.5%. It should be mentioned that our data for the low subcooling which are not shown here are substantially di€erent from the corresponding corre-lation values from Shah.

Finally, the present data for heat transfer coecient in the subcooled ¯ow boiling of R-134a are correlated as hrˆ hl  282
ln…Bo† ‡DTsub DTsat ‡ 27:4  …5† The above equation well correlates the present data with an average deviation of 9.9%, as shown in Fig. 8. 3.2. Bubbles in subcooled boiling

It is well known that the e€ects of various par-ameters on the R-134a subcooled ¯ow boiling heat transfer presented above have to be related to the characteristics of bubbles on the heated surface during the boiling. These bubble characteristics are examined in the following.

Fig. 9 shows the typical photos of bubbles taken from the present ¯ow visualization for various imposed heat ¯uxes for G = 200 kg/m2 _{s, T}

sat = ÿ58C and

DTsubˆ 7 8C. The results indicate that at a higher qw00 a

larger fraction of the heated surface is covered by the bubbles. Besides, the bubble generation frequency and bubble rising velocity in the liquid were found to be also higher. Thus, the corresponding boiling heat

transfer from the surface is better. A close inspection of the photos reveals that most bubbles are elliptic and the largest bubbles are about 0.6 mm in average diam-eter, which is rather small in comparison with those in boiling of water. Moreover, the bubbles are largest when they are about to departure from the surface. Note also that the maximum bubble size changes only slightly with the imposed heat ¯ux.

The results from the visualization of bubbles for var-ious mass ¯uxes shown in Fig. 10 suggest that at a lower mass ¯ux more bubbles are generated at a larger number of active nucleation sites and at a higher fre-quency. This mainly attributes to the fact that the heated surface temperature and hence the wall super-heat are higher for a lower G. Secondly, the liquid re-frigerant R-134a moves faster at a higher G, which tends to sweep the bubbles away from the surface. However, the bubble sizes are not a€ected by the mass ¯ux to a noticeable degree.

Next, in Fig. 11 we show that the subcooling of the refrigerant exhibits profound in¯uences on the bubble size and population. At the higher subcooling of 138C the bubbles are much smaller and are in larger num-ber, compared with those for DTsubˆ 6 8C and

DTsubˆ 0 8C. This outcome results from the fact that

in a highly subcooled liquid the bubbles are signi®-cantly suppressed by the surrounding liquid and grow slowly. For example, the bubbles about to depart from the heated surface for DTsubˆ 13 8C have diameter

about dp 0:3 mm. But for DTsubˆ 6 8C, dp 0:6

mm and for DTsubˆ 1 8C, dp 1:0 mm. According to

the present data, an empirical correlation for the bubble departure diameter modi®ed from that of Cole and Shulman [14] for the saturated ¯ow boiling is pro-posed here for the subcooled ¯ow boiling of R-134a and it is dp  _s g…rfÿ rg† 1=2_{ˆ 2:84 }1000 P exp… ÿ 0:184Nsub† …6† where Nsubis the subcooling number, de®ned as

Nsubˆ _h fÿ hi ifg vgÿ vf vf  …7† The above equation well correlates the present data with an average deviation of 12.6%, as shown in Fig. 12.

4. Concluding remarks

Experiments have been carried out in the present study to investigate the subcooled nucleate ¯ow boiling heat transfer and the associated bubble characteristics

Fig. 8. Comparison of the measured data for heat transfer coecient in the subcooled ¯ow boiling of R-134a with the proposed correlation.

Fig. 9. Subcooled ¯ow boiling of R-134a in an annular channel at G = 200 kg/m2_{s T}

satÿ 58C and DTsubˆ 78C: (a) qw00ˆ 5 kW/m2

(b) q00

for R-134a in an annular duct. E€ects of various par-ameters on the hysteresis during the onset of nucleate boiling (ONB) were examined in detail. At lower Tsat

and higher DTsub the temperature undershoots during

ONB are larger. But the mass ¯ux of R-134a shows a nonmonotonic e€ect on the temperature undershoot. It was noted that the boiling heat transfer is not signi®-cantly a€ected by the mass ¯ux, imposed heat ¯ux and

Fig. 10. Subcooled ¯ow boiling of R-134a in an annular channel at q00

wˆ 10 kW/m2Tsatˆ 2:58C and DTsubˆ 78C: (a) G = 100 kg/

Fig. 11. Subcooled ¯ow boiling of R-134a in an annular channel at G = 200 kg/m2 _{s, q}00

wˆ 10 kW/m2 and Tsatˆ 18C: (a)

saturation temperature. But an increase in the subcool-ing results in much better heat transfer. Flow visualiza-tion of the boiling processes revealed that the bubble generation frequency was suppressed by raising the mass ¯ux and subcooling of R-134a. Moreover, only the subcooling showed a large e€ect on the bubble size. Finally, empirical correlations for the boiling heat transfer coecient and for the bubble departure diam-eter in the subcooled ¯ow boiling of R-134a based on the present data were provided.

Some other important phenomena such as the in¯u-ences of the oil contamination on the R-134a ¯ow boil-ing and condensation characteristics remain largely unknown. This will be explored in the future.

Acknowledgements

The ®nancial support of this study by the engineer-ing division of National Science Council of Taiwan, R.O.C. through the contract NSC85-2221-E-009-046 is greatly appreciated.

References

[1] J.R. Thome, Boiling of new refrigerants: a state-of-the art review, Int. J. Refrig. 19 (1996) 435±457.

[2] F.C. Gunther, Photographic study of surface-boiling heat transfer to water with forced convection, ASME Transaction C Journal of Heat Transfer 73 (1951) 497± 499.

[3] W.H. McAdams, W.E. Kennel, C.S. Minden, R. Carl, P.M. Picornell, J.E. Dew, Heat transfer at high rates to water with surface boiling, Indust. Engng. Chem. 41 (1949) 1945±1953.

[4] E.L. Bibeau, M. Salcudean, A study of bubble ebullition in forced-convective subcooled nucleate boiling at low pressure, Int. J. Heat and Mass Transfer 37 (1994) 2245±2259.

[5] A. Hasan, R.P. Roy, S.P. Kalra, Experiments on sub-cooled ¯ow boiling heat transfer in a vertical annular channel, Int. J. Heat and Mass Transfer 33 (1990) 2285± 2293.

[6] M.M. Shah, Generalized prediction of heat transfer during subcooled boiling in annuli, Heat Transfer Engng 4 (1) (1983) 24±31.

[7] K.E. Gunger, R.H.S. Winterton, A general correlation for ¯ow boiling in tubes and annuli, Int. J. Heat and Mass Transfer 29 (1986) 351±358.

[8] J.G. Collier, Convective Boiling and Condensation, McGraw-Hill, New York, 1981, pp. 211±221 (Chapter 7).

[9] M.A. Kedzierski, Simultaneous visual and calorimetric measurements of R11, R123, and R123/alkylbenzene nucleate ¯ow boiling, State of the Art and Emerging Technologies in Enhanced Heat Transfer, A Short Course by the University of Maryland, College of Engineering, April 18/19, 1996.

[10] W. Fritz, Berechnung des Maximalvolumens von Dampfblasen, Physik, Zeitschr 36 (1935) 379±384. [11] J.F. Klausner, R. Mei, D.M. Bernhard, L.Z. Zeng,

Vapor bubble departure in forced convection boiling, Int. J. Heat and Mass Transfer 36 (1993) 651±662. [12] H. Muller-Steinhagen, A.P. Watkinson, N. Epstein,

Subcooled-boiling and convective heat transfer to hep-tane ¯owing inside an annulus and past a coiled wire: part 1Ðexperimental results, ASME Transaction C Journal of Heat Transfer 108 (1986) 922±927.

[13] M.E. Bland, Bubble nucleation in cryogenic ¯uids. Ph.D. thesis, Oxford University, England, 1970. [14] R. Cole, H.L. Shulman, Bubble departure diameters at

subatmospheric pressures, Chemical Engineering Progress, Symposium Series 62 (64) (1996) 6±16. [15] S.J. Kline, F.A. McClintock, Describing uncertainties in

single-sample experiments, Mechanical Engineering 75 (1953) 3±12.

Fig. 12. Comparison of the measured bubble departure diam-eters with the proposed correlation.