Saturated flow boiling heat transfer of refrigerant R-410A in a horizontal annular finned duct

(1)

Saturated ﬂow boiling heat transfer of refrigerant R-410A

in a horizontal annular ﬁnned duct

Y.Y. Hsieh

a

, Y.M. Lie

b

, T.F. Lin

b,*

a_{Department of Mechanical Engineering, Nan Kai Institute of Technology, Nantou, Taiwan, ROC} b_{Department of Mechanical Engineering, National Chaio Tung University, Hsinchu 30010, Taiwan, ROC}

Received 28 June 2006; received in revised form 8 September 2006 Available online 14 November 2006

Abstract

An experiment is conducted here to investigate the saturated flow boiling heat transfer characteristics of ozone friendly refrigerant R-410A in a horizontal annular finned duct. Meanwhile the associated bubble characteristics in the duct are also inspected from the flow visualization. The experimental data are presented in terms of saturated flow boiling curves, boiling heat transfer coefficients and flow photos. In addition, empirical correlation equations for the saturated flow boiling heat transfer coefficient and mean bubble departure diameter are proposed. The saturated flow boiling curves show that boiling hysteresis is insignificant in the flow and the wall superheat needed for the onset of nucleate boiling is slightly affected by the refrigerant mass flux. Besides, the boiling curves are mainly affected by the imposed heat flux and refrigerant mass flux. Moreover, the measured saturated flow boiling heat transfer coefficient increases with the imposed heat flux and refrigerant mass flux. Furthermore, at a higher refrigerant mass flux the departing bubbles are smaller. Ó 2006 Elsevier Ltd. All rights reserved.

1. Introduction

It is well known that refrigerant R-22 is still one of the most widely used HCFCs (hydrochloroﬂuorocarbons) employed as a working ﬂuid in air conditioning systems. But it will be phased out in a short period of time since the chlorine it contains has high ozone depletion and global warming potentials. Recently, various non-chlorine alter-native refrigerants such as R-134a, R-407C and R-410A were proposed to replace R-22 for various applications. In particular, R-410A which is the most widely used ozone depletion free refrigerant is recognized as the main replace-ment to R-22.

In order to properly use these new refrigerants in the air conditioning and refrigeration systems, the detailed under-standing of the two-phase heat transfer and associated bubble characteristics for these new refrigerants is essential.

The main purpose of this study is to investigate the satu-rated flow boiling of refrigerant R-410A in a horizontal annular finned duct by measuring the boiling heat transfer coefficient and visualizing the motion of bubbles in the boiling flow.

In the following the relevant literature on the present study of R-410A saturated flow boiling is reviewed. It should be mentioned here that refrigerant R-410A is a mix-ture of R-32 and R-125 (50% by mass) which exhibits aze-otropic behavior with a small temperature glide of about 0.1°C. A recent comprehensive review of the literature on the nucleate pool and flow boiling of new refrigerants was conducted by Thome [1]. He assessed the strength and weakness in the existing boiling correlations and in the maps for two-phase flow patterns and identified the further research requirements and objectives. Sami and Poirier [2] compared the evaporation and condensation heat transfer data for several refrigerant blends pro-posed as substitutes for R-22, including R-410A, R-410B, R-507 and the quaternary mixture R-32/125/143a/134a, inside an enhanced-surface tubing. They showed that the

*

Corresponding author. Tel.: +886 35 712121x55118; fax: +886 35 726440.

E-mail address:tﬂ[email protected](T.F. Lin).

(2)

two-phase heat transfer coefficients and pressure drops increased with the refrigerant mass flux for all alternative refrigerants and R-22. In a continuing study[3], they pre-sented the data for R-410A and R-507 in a double fluted tube indicating that for the refrigerant Reynolds number higher than 4.2 106, R-410A had higher heat transfer rate than R-507. Wang et al. [4] tested nucleate boiling on several commercially available enhanced-surface tubes to assess the pool boiling heat transfer performance for R-22, R-123, R-134a, R-407C, and R-410A. The heat transfer coefficient of R-410A was found to be higher than that of R-22 for most enhanced tubes. This outcome is attributed to the higher latent heat, thermal conductivity and specific heat and lower liquid viscosity for R-410A. Later, they[5]

compared the two-phase heat transfer and pressure drop characteristics of refrigerants R-22 and R-410A flowing in a horizontal smooth tube. Their results indicated that the heat transfer coefficients for R-410A were about 10– 20% higher than that for R-22 and the pressure drops of R-410A were about 30–40% lower than that of R-22. Ebisu and Torikoshi [6]measured the evaporation heat transfer coefficients and proposed empirical correlations for R-410A, R-407C and R-22 flowing in a horizontal smooth tube. Their results showed that the evaporation heat trans-fer coefficient of R-410A was about 20% higher than that of R-22 up to the vapor quality of 0.4, while the heat trans-fer coefficients for R-410A and R-22 became almost the

same at the quality of 0.6. Furthermore, the pressure drop for R-410A was about 30% lower than that of R-22 during evaporation. The quantitative differences in the pressure drops between R-410A and R-22 were mainly attributed to the differences in the vapor densities of two refrigerants. The larger the vapor density, the smaller the pressure drop of a refrigerant. Similar study was carried out by Wijaya and Spatz[7]for refrigerants R-22 and R-410A in a hori-zontal smooth copper tube. Their data showed that the evaporation heat transfer coefficient for R-410A was much higher (about 23–63%) than that for R-22, while the R-410A pressure drop was 23–38% lower than that for R-22. The advantageous heat transfer and pressure drop characteristics for R-410A were ascribed to the better transport properties for R-410A. Shen et al. [8] provided the data for the pool boiling heat transfer coefficient of binary mixture R-32/R-125 with different mole fractions of R-32. The results indicated that the pressure and heat flux dependences of the boiling heat transfer coefficient for the binary mixture did not significantly differ from those of pure components. The frictional pressure charac-teristics for R-410A flowing in a 5-mm smooth tube reported by Chang et al.[9]showed the substantial increase of the pressure drop gradient with the refrigerant mass flux and vapor quality. Furthermore, they analyzed the fric-tional pressure drop using the two-phase fricfric-tional multi-plier. The averaged heat transfer coefficients for forced Nomenclature

As, Acs surface area and cross-sectional area, m2

b1, b2 ﬁn size deﬁned inFig. 3, m2

Bo boiling number, Bo¼_Giq_fg, dimensionless cp speciﬁc heat, J/kg°C Co convection number, Co¼ qg qf 0:5 1x x 0:8 , dimen-sionless

dp mean bubble departure diameter, m

dp mean non-dimensional bubble departure

diame-ter, dimensionless

D diameters of annular duct deﬁned in Fig. 3, m Dh hydraulic diameter, m

Fr Froude number, Fr¼ G2

q2

fgDh, dimensionless g acceleration due to gravity, m/s2

G refrigerant mass ﬂux, kg/m2s

hr,f liquid convection heat transfer coeﬃcient,

W/m2°C

hr,sat saturated boiling heat transfer coeﬃcient,

W/m2°C ifg enthalpy of vaporization, J/kg I current, A k thermal conductivity, W/m°C l length, m p pitch of ﬁns, m Pr Prandtl number, Pr¼lfcp;f kf , dimensionless

q, qt average net heat ﬂux and total imposed heat

ﬂux, kW/m2

Q, Qt convection heat transfer rate and total power

input to test section, kW Re Reynolds number, Re¼GDh

lf , dimensionless T temperature,°C

V voltage, V

x vapor quality

z downstream coordinate for annular duct ﬂow, m

Greek symbols

Dq density diﬀerence between liquid and vapor phases, kg/m3 l dynamic viscosity, N s/m2 q density, kg/m3 r surface tension, N/m Subscripts ave average f, g liquid/vapor phase i, o inlet/inner, outlet/outer sat saturated state

(3)

convection condensation of R-410A and R-22 inside smooth horizontal tubes were experimentally determined by Chitti and Anand [10]. Their results showed that R-410A had about 15–20% higher regionally averaged heat transfer coefficient when compared to R-22 at a given refrigerant mass flux. They also studied the effects of lubri-cating oil on the condensation heat transfer characteristics. Recently, Hsieh and Lin [11]measured the saturated flow boiling and pressure drop of R-410A in a vertical plate heat exchanger. Their results indicate that the heat flux exhibits significant effects on the saturated flow boiling heat transfer coefficient.

Based on the available experimental data from the open literature, some correlating equations for flow boiling heat transfer were proposed [12–15]. An early general correla-tion model for the two-phase flow boiling, which is still widely used, is that of Chen [12]. He divided the boiling heat transfer coefficient into two parts: a microconvective (nucleate boiling) contribution estimated by some pool boiling correlations and a macroconvective (non-boiling forced convection) contribution estimated by some single-phase correlations. In order to account for the diminished contribution of nucleate boiling as the macroconvective effects increased at a higher vapor quality, he introduced the enhanced factor E and suppression factor S to respec-tively accommodate the forced convective and nucleate boiling contributions. Gungor and Winterton[13]modified the Chen’s correlation and proposed the correlations for the enhanced and suppression factors. An enhanced model was developed by Kandlikar [14] to predict the saturated boiling heat transfer coefficients for refrigerants inside hor-izontal and vertical tubes. A new correlation from Liu and Winterton [15]introduced an asymptotic function to pre-dict the heat transfer coefficients for vertical and horizontal flows in tubes and annuli.

The above literature review clearly reveals that although R-410A is one of the most possible substitutes for R-22, the two-phase boiling heat transfer and associated bubble characteristics for R-410A remain poorly explored. In the present study experiments are conducted to investigate the heat transfer characteristics in the saturated ﬂow boil-ing of R-410A in a horizontal annular ﬁnned duct and to visualize the motion of bubbles in the duct.

2. Experimental apparatus and procedures

The experimental apparatus established in the previous study[16]is modified slightly here to investigate the satu-rated flow boiling heat transfer of R-410A in a horizontal annular finned duct and is schematically shown inFig. 1. The apparatus consists of three main loops, namely, the refrigerant, water–glycol and water loops, along with a data acquisition system. Refrigerant R-410A is circulated in the refrigerant loop. We need to control the temperature and flow rate in the water–glycol loop to have enough cool-ing capacity for condenscool-ing the R-410A vapor and sub-cooling the R-410A liquid to a preset temperature. We

further utilize the water loop installed before the test sec-tion to procure the preset condisec-tion of refrigerant R-410A at the test section inlet. A DC power supply is used for heating the inner pipe in the test section. To obtain the data of the bubble departure size in the R-410A satu-rated ﬂow boiling in the annular duct, a camera connected to a photographic microscope is set up beside the test sec-tion to visualize the boiling ﬂow.

2.1. Refrigerant loop

The refrigerant loop contains a variable-speed refriger-ant pump, an accumulator, a mass flow meter, a test section, a condenser, a sub-cooler, a receiver, a filter/dryer and four sight glasses. A rotational DC motor is used to drive the refrigerant pump which then circulates the R-410A liquid from the receiver to the test section. Through changing the DC current in the motor, the liquid flow rate of R-410A can be varied. By adjusting the gate open-ing 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 flow rate and pressure should be adjusted simultaneously in order to control them at the required levels. Moreover, a high-pressure nitrogen tank is installed in the loop to dampen the fluctuations of the flow rate and pressure. The refrigerant flow rate is mea-sured by a mass flow meter with a reading accuracy of 1%. The filter/dryer is used to filter the impurities and non-condensable gas possibly existing in the loop. Meanwhile, the condenser and subcooler are used to condense the R-410A vapor leaving the test section and then the liquid R-410A is subcooled by the cold water–glycol loop. Vary-ing the temperature and flow rate of the water–glycol mix-ture moving through the condenser and subcooler allows us to control the bulk temperature of R-410A leaving the subcooler. After subcooled, the liquid R-410A flows back to the receiver at the bottom of the system.

2.2. Test section

As schematically shown in Fig. 2, the test section is a horizontal annular finned duct with the outer pipe made of pyrex glass to permit the visualization of boiling pro-cesses. The outer pyrex glass pipe is 160 mm long, having a 20-mm inside diameter and 2.5-mm thick wall. Both ends of the pipe are connected with copper tubes by means of flanges and are sealed by O-rings. The inner finned pipe is manufactured from a copper pipe. It is 180 mm long and has 12.7-mm nominal outside diameter. There are heli-cal integral low fins (26 fins/in.) on the outside surface of the inner pipe. The detailed geometrical data of the finned pipe are given inFig. 3 andTable 1. An electric cartridge heater of 160-mm in length and 7.0-mm in diameter with a maximum power output of 800 W is inserted into the inner finned pipe. Furthermore, the finned pipe has a 10-mm inactive heating zone at each end and is insulated with teflon blocks and the thermally non-conducting epoxy to

(4)

minimize heat loss from it. Thermal contact between the heater and the finned pipe is improved by coating a thin layer of heat-sink compound on the heater before installat-ing the heater. Then, 12 T-type calibrated thermocouples with their beads covered with the electrically non-conduct-ing thermal bond are fixed on the inside surface of the finned pipe so that the voltage signals from the thermocou-ples are not interfered by the DC current passing through the cartridge heater. The thermocouples are positioned at five axial stations along the finned pipe. At each axial sta-tion, two to four thermocouples are placed at top, bottom, or two sides of the pipe circumference with 180° or 90° apart. The outside surface temperature of the finned pipe is then derived from the measured inside surface tempera-ture by taking the radial thermal conduction through the pipe wall into account.Fig. 4 shows the detailed thermo-couple locations at each axial station and the arrangement of the cartridge heater.

2.3. DC power supply

As described above, the ﬁnned pipe is heated by a 800-W cartridge heater. A 50V-25A DC power supply delivers the required electric current to the cartridge heater. A

Yokog-awa DC meter is used to measure the dc current through the cartridge heater with an accuracy of 1%. Then the volt-age drop across the heater is measured by a Yokogawa multimeter. Thus the power input to the heater can be calculated.

2.4. Photographic system

The photographic apparatus established in the present study to unravel the bubble characteristics in the saturated flow boiling in the annular finned duct consists of a Kodak Motion Corder Analyzer and a Nikon 85-mm microlens camera. The high-speed motion analyzer system can take photographs up to 10,000 frames/s. Here, a recording rate of 1000 frames/s is used to obtain the images of bubble ebullition processes in the flow boiling. The bubble dynam-ics data are collected at three different axial locations along the flow direction (z/Dh= 6, 12 and 18). After the system

reaches a statistically steady state, we record the boiling activity. The high-speed motion analyzer stores the images which are later downloaded to a tape to facilitate the play-back at slower motion (15 frames/s) on a CRT. Then, the bubble departure diameter is calculated by viewing more than 500 frames at each location. In order to achieve the

(5)

highest possible resolution and eliminate errors in calibra-tion, the camera lens is ﬁxed at a constant focal length, resulting in a ﬁxed viewing area of approximately

5-ﬁn-pitch length (corresponding to an image ﬁeld area of 5.5 4.5 mm). Typically, a total of over 150 bubble diam-eter measurements is used to construct the present data.

Fig. 2. Schematic diagram of the test section for saturated ﬂow boiling in horizontal annular ﬁnned duct.

b1 p b2 h R D D i D o

(6)

2.5. Water–glycol mixture loop

The water–glycol loop designed for condensing the R-410A vapor and for subcooling the liquid R-R-410A contains a 125-l constant temperature bath with a water cooled refrigeration system. The cooling capacity is 2 kW for the water–glycol mixture at 20 °C. The cold water–glycol mixture at a specified flow 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 flow rate. By adjusting the mixture temperature and flow rate, the bulk temperature of the R-410A in the subcooler can be controlled at a pre-set level.

2.6. Water loop for preheater

In order to have a better control of the R-410A temper-ature at the test section inlet, a water loop is used to pre-heat the R-410A before it arrives at the test section inlet. This water loop consists of a 20-l water thermostat with a 0.5-kW heater and an air cooled refrigeration unit of 0.5-kW cooling capacity. A 0.5-hp water pump is used to drive the hot water at a specified temperature to the heat exchanger of the preheater at a specified water flow rate. A by-pass valve can also be used to adjust the water flow rate. The reading accuracy of measuring the water flow rate is ±0.5%.

2.7. Data acquisition

The data acquisition system includes a recorder, a 24V-3A power supply and a computer. The recorder is used to record the temperature and voltage data. The water flow-meter and differential pressure transducer need the power supply as a driver to output an electric current of 4– 20 mA. The IEEE 488 interface is used to connect the com-puter with the recorder, allowing the measured data to be transmitted from the recorder to the computer and then analyzed immediately. The system automatically monitors 22 copper–constantan thermocouples, two pressure trans-ducers, and two mass flowmeters.

2.8. Experimental procedures

In the test the R-410A liquid at the inlet of the test sec-tion is first maintained at a specified temperature by adjust-ing the water–glycol temperature and flow rate. In addition, we adjust the thermostat temperature in the water loop to obtain the saturated liquid R-410A at the test sec-tion inlet. Then, we regulate the R-410A pressure at the test section inlet by adjusting the gate valve locating right after the exit of the test section. Meanwhile, the current of the DC motor connecting with the refrigerant pump is selected to obtain the required R-410A flow rate. The imposed heat flux from the heater to the R-410A is adjusted by varying

Table 1

Speciﬁcations of the inner ﬁnned duct

Items Speciﬁcations

Geometry type Helical integral low ﬁns

Finned pipe diameter, Do(mm) 12.7

Nominal outside diameter, D (mm) 11.5

Nominal inside diameter, Di(mm) 8.0

Number of ﬁns, N (ﬁns/in.) 26

Fin pitch, p (mm) 0.98

Fin height, hR(mm) 1.3

Finned thickness (base), b1(mm) 0.45

Finned thickness (tip), b2(mm) 0.25

Helix angle, a (deg) 5.0

Fin included angle, b (deg) 25

Actual heat-transfer area, Aas(m2/m) 0.098

Tube length, l (mm) 160

Fig. 4. Detailed cross-section of the cartridge heater and thermocouple locations.

(7)

the electric current delivered from the DC power supply. By measuring the current delivered to and voltage drop across the heater, we can calculate the heat transfer rate to the refrigerant.

2.9. Uncertainty analysis

The uncertainties of the experimental results are analyzed by the procedures proposed by Kline and McClintock [17]. The detailed results from the present uncertainty analysis for the experiments conducted here are summarized inTable 2.

3. Data reduction

A data reduction analysis is needed to calculate the sat-urated flow boiling heat transfer coefficient from the raw data measured in the horizontal annular finned duct. The data reduction process is described in the following.

The imposed heat ﬂux was calculated on the basis of the total power input and the outside surface area of the inner pipe. The total electrical power input to the test section was computed from the product of the measured voltage drop across the cartridge heater and the electric current passing through it. The imposed heat ﬂux at the outside surface of the inner pipe based on the total power input is evaluated from

q_t¼V I As

ð1Þ where V and I respectively represent the measured voltage drop and current. The outside surface area of the inner ﬁnned pipe Asis used in Eq.(1)and is equal to pDol, where

Do is the outer diameter of the pipe (Fig. 3) and l is the

length of the inner pipe (l = 160 mm). Before the ﬂow boil-ing experiments, the total heat loss from the test section was inspected by comparing the total power input from

the power supply with the enthalpy increase in the single-phase liquid refrigerant ﬂow which can be expressed as Q¼ G Acs cp;f ðTo TiÞ ð2Þ

where Acs is the cross-sectional area of the annular duct,

and Toand Tiare respectively the refrigerant temperatures

at the exit and inlet of the test section. The results from this heat loss test indicated that the relative heat loss from the test section, e = (Qt Q)/Qt, was generally less than ±4%

of the total power input Qt.

The average saturated flow boiling heat transfer coeffi-cient of R-410A in the annular duct from the present exper-iments is defined as

hr;sat¼

q ðTwall TsatÞ

ð3Þ where q is the net imposed heat ﬂux to the R-410A ﬂow and is estimated by q = qt (1 e). Tsatis the saturated

temper-ature corresponding to the test pressure. Note that Twall

Table 2

Summary of the uncertainty analysis

Parameter Uncertainty

Annular ﬁnned duct geometry

Length, diameter (m) ±0.00005

Area (m2) ±7 105

Sensors

Temperature, T (°C) ±0.2

Temperature diﬀerence, DT (%) ±4.5%

System pressure, P (MPa) ±0.002

Pressure drop, DP (Pa) ±200

Water ﬂow rate, Ww(%) ±2

Mass ﬂux of refrigerant, G (%) ±2

Single-phase heat transfer

Heat transfer rate, q (%) ±6.5

Heat transfer coeﬃcient, hr,f(%) ±11

Saturated ﬂow boiling heat transfer

Boiling heat ﬂux, q (%) ±6.5

Heat transfer coeﬃcient, hr,sat(%) ±14.5

Boiling Stops ONB Boiling Stops ONB 0 5 10 15 0 20 40 60 q (kW/m 2) 0 5 10 15 T_wall- T_sat(oC ) 0 20 40 60 q (kW/m 2)

R-410A Saturated Flow Boiling Curves at Tsat= 10

o

C, G= 200 kg/m2_s

Increasing Heat Flux

Decreasing Heat Flux ONB: Onset Nucleate Boiling

o

C, G= 250 kg/m2_s

Increasing Heat Flux

Decreasing Heat Flux ONB: Onset Nucleate Boiling

Fig. 5. Boiling curves showing the degree of hysteresis eﬀect during the saturated ﬂow boiling at (a) Tsat= 10°C, G = 200 kg/m

2

s and (b) Tsat= 15°C, G = 250 kg/m

2

(8)

denotes the average outside surface temperature of the in-ner pipe. The outside surface temperature Twall,o at each

thermocouple location is deduced from the measured inside surface temperature Twall,i of the pipe by subtracting the

radial temperature drop due to the radial heat conduction in the pipe wall. Thus we have

Twall;o¼ Twall;i Q

lnðD0=DiÞ

2pkwall l

ð4Þ In this evaluation of Twall,othe axial conduction in the pipe

wall is assumed to be small compared with the radial con-duction. More speciﬁcally, Twall is then considered as the

arithmetic average of the deduced outside surface tempera-tures of the inner pipe from the thermocouple data. 4. Results and discussion

The present saturated ﬂow boiling experiments are con-ducted for the refrigerant mass ﬂux varying from 150 to

250 kg/m2s, imposed heat flux from 0 to 50 kW/m2 and refrigerant saturated temperature from 5 to 15°C (corre-sponding to the saturation pressure from 0.93 to 1.25 MPa). Selected data from the experiments are pre-sented here to illustrate the characteristics of the heat transfer and bubble motion in the saturated flow boiling of R-410A in the horizontal annular finned duct. Specifi-cally, the effects of the imposed heat flux, refrigerant mass flux and saturated temperature on the measured boil-ing curves and heat transfer coefficients and bubble char-acteristics will be examined in detail. In addition, empirical correlation equations for the flow boiling heat transfer coefficients and bubble departure diameter are proposed.

4.1. Boiling curves

The possible presence of the boiling hysteresis is exam-ined ﬁrst in Fig. 5 by presenting the data for the boiling

ONB (G=150 kg/m2 s) ONB (G=200 kg/m2_s) ONB (G=150 kg/m2_s) ONB (G=200 kg/m2 s) ONB (G=250 kg/m2_s) ONB (G=250 kg/m2 s) 0 5 10 15 0 20 40 60 q (kW/m 2) 0 5 10 15 T_wall- T_sat(oC ) 0 20 40 60 q (kW/ m 2)

o C G= 150 kg/m2_s G= 200 kg/m2_s G= 250 kg/m2_s

ONB: Onset Nucleate Boiling

o C G= 150 kg/m2_s G= 200 kg/m2_s G= 250 kg/m2_s

Fig. 6. Saturated ﬂow boiling curves for various refrigerant mass ﬂuxes (G = 150, 200 and 250 kg/m2s) at (a) Tsat= 5°C and (b) Tsat= 15°C.

ONB (Tsat=5 ) ONB (Tsat=10 ) ONB (Tsat=15 ) ONB (Tsat=10 ) ONB (Tsat=15 ) ONB (Tsat=5 ) 0 5 10 15 0 20 40 60 q (kW/m 2) 0 5 10 15 T_wall- T_sat( C ) 0 20 40 60 q (kW/ m 2)

R-410A Saturated Flow Boiling Curves at G= 150 kg/m2_s

Tsat= 5 C C C C Tsat= 10 C Tsat= 15 C ONB: Onset Nucleate Boiling

R-410A Saturated Flow Boiling Curves at G= 200 kg/m2_s

Tsat= 5 C

Tsat= 10 C

Tsat= 15 C

C C C

Fig. 7. Saturated ﬂow boiling curves for various saturated temperatures (Tsat= 5, 10 and 15°C) at (a) G = 150 kg/m

2

(9)

curves measured at the middle axial location (z = 80 mm) for the cases with Tsat= 10°C and G = 200 kg/m

2

s and Tsat= 15°C and G = 250 kg/m

2

s. The results show that for each case the wall superheat (Twall Tsat) increases

almost exponentially with the imposed heat flux q. More-over, the boiling curves for the increasing and decreasing imposed heat fluxes almost overlap. Apparently, the boil-ing hysteresis and the temperature undershoot at the onset of nucleate boiling (ONB) are small for the saturated flow boiling in this annular finned duct. The smaller boiling hys-teresis and temperature undershoot are considered to mainly result from the fact that most of cavities are acti-vated on the fin surface in the saturated flow boiling even at the low imposed heat flux of q = 5 kW/m2. It is worth mentioning that the hysteresis and temperature undershoot in the horizontal smooth duct are relatively significant[18]. Besides, the condition for the ONB is difficult to determine from the data in the measured boiling curves since the slope change in the boiling curves during the ONB is very small.

Hence, the ONB is determined here from the ﬂow visualization.

Next, the effects of the refrigerant mass flux and satu-rated temperature on the satusatu-rated flow boiling heat trans-fer of R-410A in the horizontal annular finned duct are illustrated inFigs. 6 and 7by presenting the boiling curves for various G at a fixed Tsatand for various Tsatat a fixed

G. The results inFig. 6manifest that at low wall superheat for DTsat< 5°C the boiling curves are only slightly

affected by the refrigerant mass flux. From the flow visual-ization, we noted that in the initial stage of heating the pipe wall, the wall temperature rose slowly but was below that needed for the ONB. Hence no bubble appears on the heating surface. Heat transfer in the flow results com-pletely from the single-phase forced convection. As the imposed wall heat flux was raised gradually, the wall tem-perature increased correspondingly. At a certain wall superheat a few bubbles started to nucleate from the heat-ing surface and we had onset of nucleate boilheat-ing in the

Onset Nucleate Boiling Single Phase Regime

Saturated Flow Boiling Regime Onset Nucleate Boiling Single Phase Regime

Saturated Flow Boiling Regime

0 20 40 60 0 2000 4000 6000 hr, sa t (W/m 2 oC ) 0 20 40 60 q (kW/m2₎ 0 2000 4000 6000 hr,sa t (W/m 2 oC )

R-410A Saturated Heat Transfer Coefficients at Tsat= 5 C

G= 150 kg/m2_s

G= 200 kg/m2_s

G= 250 kg/m2_s

R-410A Saturated Heat Transfer Coefficients at Tsat=15 C

G= 150 kg/m2_s

G= 200 kg/m2_s

G= 250 kg/m2_s

Fig. 8. Variations of saturated flow boiling heat transfer coefficient with imposed heat flux for various refrigerant mass fluxes (G = 150, 200 and 250 kg/m2s) at (a) Tsat= 5°C and (b) Tsat= 15°C.

Onset Nucleate Boiling Single Phase Regime

Saturated Flow Boiling Regime Onset Nucleate Boiling Single Phase Regime

Saturated Flow Boiling Regime

0 20 40 60 0 2000 4000 6000 hr, sa t (W/m 2 oC ) 0 20 40 60 q (kW/m2₎ 0 2000 4000 6000 hr,sa t (W/m 2 oC )

R-410A Saturated Heat Transfer Coefficients at G= 150 kg/m2_s

Tsat= 5 C

Tsat= 10 C

Tsat= 15 C

R-410A Saturated Heat Transfer Coefficients at G= 200 kg/m2_s

Tsat= 5 C

Tsat= 10 C

Tsat= 15 C

Fig. 9. Variations of saturated flow boiling heat transfer coefficient with imposed heat flux for various saturated temperatures (Tsat= 5, 10 and

15°C) at (a) G = 150 kg/m2

(10)

flow. Beyond the ONB, the bubble departure frequency increases significantly with the imposed heat flux, which causes a substantial increase in the heat transfer rate from the wall to the refrigerant. We further noted that at a slightly higher wall superheat for DTsat> 5°C the heat

transfer rate from the heating surface to the refrigerant increased with the refrigerant mass ﬂux for Tsat= 5°C

(Fig. 6a). Besides, the bubbles in the liquid refrigerant

moved vigorously and turbulently due to the higher refrig-erant mass ﬂux. At an even higher wall superheat for DTsat> 8°C the heat transfer in the R-410A saturated

flow boiling is noticeably better at higher refrigerant mass flux. The increase is more prominent at a higher wall superheat. This increase in the heat transfer rate with the mass flux is primarily due to the increasing single-phase convection in the boiling flow with the increasing mass flux. The results from Fig. 6 also indicate that the wall superheat at the ONB is only slightly influenced by the change in the mass flux. The wall superheat at the ONB is about 3–4°C which is much lower than that in smooth tube [18]. However, the imposed heat flux needed for the ONB is somewhat higher at the higher refrigerant mass flux.

Finally, we note fromFig. 7that even at the high mass flux of 200 kg/m2s the refrigerant saturated temperature only has relatively slight effects on the boiling curves. At the lower G of 150 kg/m2s the effects of Tsatcan be ignored

(Fig. 7a). It is further noted that the wall superheat and

imposed heat flux at the ONB are somewhat affected by the refrigerant saturated temperature. But the effects of Tsat

on DTsatand qONBdo not exhibit a monotonic trend.

4.2. Saturated ﬂow boiling heat transfer coeﬃcient

The variations of the saturated flow boiling heat transfer coefficient with the imposed heat flux are shown inFigs. 8

and 9 for various refrigerant mass ﬂuxes and saturated

temperatures.

The results inFig. 8indicate that at given G and Tsatboth

the single-phase forced convection and saturated flow boil-ing heat transfer coefficients increase with the imposed heat flux. Beyond the ONB, the increase in the boiling heat trans-fer coefficient is rather significant at increasing imposed heat flux. For example, at G = 200 kg/m2s and Tsat= 5°C

the average saturated boiling heat transfer coeﬃcient for q = 50 kW/m2 is about 85% higher than that for

Fig. 10. Photos of bubbles in the saturated ﬂow boiling of R-410A around z = 80 mm for Tsat= 15°C and q = 10 kW/m 2

for (a) G = 100 kg/m2s, (b) G = 150 kg/m2s and (c) G = 250 kg/m2s.

(11)

q = 10 kW/m2(Fig. 8a). This large increase in the saturated flow boiling heat transfer coefficient is ascribed to the higher nucleation density on the heating surface, higher bubble departure frequency and faster bubble growth for a higher imposed heat flux. It is further noted that at a high imposed heat flux for q > 40 kW/m2 the increase of hr,sat with the

heat flux becomes much slower. Furthermore, a higher refrigerant mass flux always results in a somewhat higher saturated flow boiling heat transfer coefficient. For instance, the average saturated flow boiling heat transfer coefficient over the entire heat flux range at Tsat= 15°C

for G = 250 kg/m2s is respectively about 25% and 8.5% higher than those for G = 150 and 200 kg/m2s (Fig. 8b).

Finally, it is noted fromFig. 9that the variations of the saturation flow boiling heat transfer coefficient with the saturated temperature of the refrigerant is within the exper-imental uncertainty. Hence the effects of Tsaton hr,satcan

be neglected.

4.3. Bubble characteristics

To illustrate the characteristics of bubbles associated with the R-410A saturated ﬂow boiling in the horizontal

annular finned duct, the photos taken for the cases at dif-ferent refrigerant mass fluxes and imposed heat fluxes from a small region around the middle axial location at z = 80 mm are shown inFigs. 10 and 11.

A closed inspection of the photos in Fig. 10 indicates that at a higher mass flux the bubble population is less on the heating surface. The results suggest that at a higher mass flux the liquid refrigerant moves at a higher speed, which in turn tends to sweep the bubbles more quickly away from the heating surface and the vapor condensation inside the bubbles is quicker. Thus, on the heating surface the bubbles are smaller and isolated. Besides, at a higher G the bubble departure frequency is higher and the bub-bles are in violent agitating motion. Obviously, a higher refrigerant mass flux results in a smaller bubble departing diameter. Note that at the low mass flux the bubble coales-cence is important and a number of big bubbles form in the duct.

How the imposed heat ﬂux aﬀects the bubble character-istics is presented in Fig. 11 for three typical cases with q = 10, 15 and 20 kg/m2 at G = 150 kg/m2s and Tsat=

5°C. The photos show that the effects of the imposed heat flux on the bubble characteristics in the saturated flow

Fig. 11. Photos of bubbles in the saturated ﬂow boiling of R-410A around z = 80 mm for Tsat= 5°C and G = 150 kg/m 2

s for (a) q = 10 kW/m2, (b) q = 15 kW/m2and (c) q = 20 kW/m2.

(12)

boiling is rather significant. A higher wall superheat results from the higher imposed heat flux, which in turn results in a larger number of active nucleation sites and a higher bubble generation frequency. Thus, the corresponding boiling heat transfer from the surface is better. Moreover, at a high imposed heat flux a large number of bubbles generated from cavities in the heating surface tend to coa-lesce to form large bubbles. At even higher imposed heat flux for q > 20 kW/m2, the bubble departure frequency is very high and the bubble coalescence occurs rather fre-quently so that it is difficult to distinguish the individual bubbles. It should be mentioned that at a high heat flux for q > 20 kW/m2 a large number of big bubbles form and they may obstruct the flow space between the fins. This will slow the small bubbles to escape from the heating face and the liquid refrigerant to rush to the heating sur-face. Hence, the boiling heat transfer can be retarded by increasing the imposed heat flux for a sufficiently high q (>50 kW/m2).

4.4. Correlation equations

An empirical equation to correlate the present data of the heat transfer coefficient for the saturated flow boiling of R-410A in the horizontal annular finned duct is pro-posed here. The propro-posed correlation is modified from that of Kandlikar[14]and can be expressed as

hr;sat¼ hr;f ½1:25 Co0:5 Fr0:35þ 0:95 Bo0:625 ð5Þ

here hr,fis the all-liquid non-boiling heat transfer coeﬃcient

and is determined from the empirical correlation for the previous single-phase heat transfer tests for liquid R-410A. It can be expressed as[18]

hr;f ¼ 0:2092 kf Dh Re0:78_Pr1=3 lave l_wall 0:14 ð6Þ In the above equation Co, Fr and Bo are respectively the convection, Froude and boiling numbers, deﬁned as Co¼ qg qf 0:5 ð1 xÞ x 0:8 ð7Þ Fr¼ G 2 q2 f g Dh ð8Þ and Bo¼ q G ifg ð9Þ The empirical correlation proposed above can correlate the present data for hr,sat with an average deviation of 10.6%

(Fig. 12). Besides, most data fall within ±20% of the

corre-lation. Then, the mean bubble departure diameter in the saturated flow boiling of R-410A in the horizontal annular finned duct estimated from the present flow visualization can be correlated as

R-410A Saturated Flow Boiling Present data

Proposed correlation, eq. (5)

0 1000 2000 3000 4000 5000

hr,sat (Present data)

0 1000 2000 3000 4000 5000 hr,sat (Proposed Correlation) +20% -20%

Fig. 12. Comparison of the measured data for heat transfer coeﬃcient in the saturated ﬂow boiling of R-410A with the proposed correlation.

R-410A Subcooled Flow Boiling Present Data

Proposed Correlation, eq. (10)

-20% +20% (Experimental data) (Proposed correlation) 0.00 0.05 0.10 0.15 0.20 0.25 0.00 0.05 0.10 0.15 0.20 0.25

_

dp dp

Fig. 13. Comparison of the measured data for mean bubble departure diameter in the saturated ﬂow boiling of R-410A with the proposed correlation.

(13)

dp dp ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi r=ðgDqÞ p ¼ 0:065 qf qg 0:667 Re0:75_Bo0:5 ð10Þ

Fig. 13 also shows that most of the present experimental

data fall within ±20% of the above correlation with an average deviation of 16.1%.

5. Concluding remarks

Experimental measurement has been conducted to investigate the heat transfer characteristics of the saturated flow boiling of R-410A in a horizontal annular finned duct. Results for the heat transfer coefficient and associated bub-ble characteristics were examined in detail. In addition, empirical equations to correlate the measured heat transfer coefficient and mean bubble departure diameter were proposed. The major results can be summarized in the following.

(1) The boiling hysteresis and the temperature under-shoot at ONB are insignificant for the saturated flow boiling in the horizontal annular finned duct which is rather different from that in the smooth duct. (2) The effects of the imposed heat flux and refrigerant

mass flux on the boiling curves are stronger than that of refrigerant saturated temperature. Raising the imposed heat flux and refrigerant mass flux can cause a substantial increase in the boiling heat transfer coefficient.

Acknowledgments

The ﬁnancial support of this study by the engineering division of National Science Council of Taiwan, ROC through the contract NSC 93-2212-E-009-005 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] S.M. Sami, B. Poirier, Comparative study of heat transfer character-istics of new alternatives to R-22, ASHRAE Trans. 120 (1997) 824– 829.

[3] S.M. Sami, B. Poirier, Two phase ﬂow heat transfer of binary mixtures inside enhanced surface tubing, Int. Common. Heat Mass Transfer 25 (1998) 763–773.

[4] C.C. Wang, W.Y. Shieh, Y.J. Chang, Nucleate boiling performance of R-22, R-123, R-134a, R-410A and R407C on smooth enhanced tubes, ASHRAE Trans. 121 (1998) 1314–1321.

[5] C.C. Wang, J.G. Yu, P.L. Shieh, D.C. Lu, An experimental study of convective boiling of refrigerants R-22 and R-410A, ASHRAE Trans. 121 (1998) 1144–1150.

[6] T. Ebisu, K. Torikoshi, Heat transfer characteristics and correlations for R-410A ﬂowing inside a horizontal smooth tube, ASHRAE Trans. 121 (1998) 556–561.

[7] H. Wijaya, M.W. Spatz, Two-phase ﬂow heat transfer and pressure drop characteristics of R-22 and R-32/R125, ASHRAE Trans. 118 (1995) 1020–1027.

[8] J. Shen, K. Spindler, E. Hahne, Pool boiling heat transfer of refrigerant mixtures R-32/125, Int. Comm. Heat Mass Transfer 26 (1999) 1091–1102.

[9] Y.J. Chang, S.K. Chiang, T.W. Chung, C.C. Wang, Two-phase frictional characteristics of R-410A and air–water in a 5 mm smooth tube, ASHRAE Trans. 123 (2000) 792–797.

[10] M.S. Chitti, N.K. Anand, Condensation heat transfer inside smooth horizontal tube for R-22 and R-32/R-125 mixture, HVAC&R Res. 2 (1996) 79–103.

[11] Y.Y. Hsieh, T.F. Lin, Saturated ﬂow boiling heat transfer and pressure drop of refrigerant R-410A in a vertical plate heat exchanger, Int. J. Heat Mass Transfer 45 (2002) 1033–1044.

[12] J.C. Chen, A correlation for boiling heat transfer to saturated ﬂuids in convective ﬂow, Ind. Eng. Chem. Proc. Des. Dev. 5 (1966) 322–329.

[13] K.E. Gungor, R.H.S. Winterton, A general correlation for ﬂow boiling in tubes and annuli, Int. J. Heat Mass Transfer 29 (1986) 351– 358.

[14] S.G. Kandlikar, A general correlation for saturated two-phase ﬂow boiling heat transfer inside horizontal and vertical tubes, ASME J. Heat Transfer 112 (1990) 219–228.

[15] Z. Liu, R.H.S. Winterton, A general correlation for saturated and subcooled ﬂow boiling in tubes and annuli, based on a nucleate pool boiling equation, Int. J. Heat Mass Transfer 34 (1991) 2759–2766. [16] C.P. Yin, Y.Y. Yan, T.F. Lin, B.C. Yang, Subcooled ﬂow boiling

heat transfer of R-134a and bubble characteristics in a horizontal annular duct, Int. J. Heat Mass Transfer 43 (2000) 1885–1896. [17] S.J. Kline, F.A. McClintock, Describing uncertainties in

single-sample experiments, Mech. Eng. 75 (1) (1953) 3–12.

[18] C.H. Chen, Experimental of ﬂow boiling and evaporation heat transfer of R-410A and bubble characteristics in horizontal annular duct, MS. thesis, National Chaio Tung university, Hsinchu, Taiwan, ROC, 2001.