Pool boiling of R-22, R-124 and R-134a on a plain tube

10  Download (0)

全文

(1)

~ ) Pergamon

Printed in Great Britain. All rights reserved 0017-9310/97 $17.00+0.00 P I I : S0017-9310(96)00239-6

Pool boiling of R-22, R-124 and R-134a on a plain

tube

C H U N G - B I A U C H I O U and D I N G - C H U N G LU

Departrc~ent of Mechanical Engineering, National Chiao Tung University, Hsinchu 300, Taiwan, Republic of China

and

C H I - C H U A N W A N G t

Energy & Resources Laboratories, Industrial Technology Research Institute, Hsinchu 310, Taiwan, Republic of China

(Received 20 March 1996)

Abstract--Pool boiling heat transfer on a plain tube having R-22, R-124 and R-134a as working fluids are reported in the present investigation. Experiments are conducted at saturation temperatures of 4.4°C and 26.7°C and ;tt reduced pressures of 0.1 and 0.2. It is shown that the Cooper correlation can predict the present pool boiling heat transfer data satisfactorily. In addition, a semi-analytical prediction method is proposed in this study ; this semi-analytical model can not only predict the present data with success, but also give rea:~onably good accuracy with the experimental data from other researchers. Copyright © 1996

Elsevier Science Ltd.

INTRODUCTION

Nucleate boiling heat transfer is a widespread indus- trial process. The importance of nucleate boiling arises from its ability to remove enormous quantities of heat per unit time and area from hot surface with a rela- tively low thermal temperature difference. This sug- gests a tremendous reduction in heat exchanger sur- face is likely by means of nucleate boiling. In recent years, significant progress has been made toward the understanding of nucleate boiling heat transfer, and correlations has been developed in order to precisely design flooded evaporators. However, it is still very difficult to predict boiling heat transfer coefficients with satisfactory accuracy. As illustrated by Stephan and Abdelsalam [1], many of the existing heat transfer data in nucleate boiling are inconsistent with each other. Cooper [2] argued that there is no agreement on the problem of relating heat flux to driving tem- perature potentia]. Recently, Kolev [3] examined experimental data for boiling of water on plain sur- faces from 14 separate investigators. These inves- tigators had reported about 5--6% error for tem- perature measurements and approximately 1-14% deviation for heat transfer measurements. However, Kolev [3] showed Lhat the spread o f the heat transfer data from different authors is over two orders of mag- nitude, which cannot be explained by measurement error. Therefore, it is essential to incorporate the cor- t Author to whom correspondence should be addressed.

rect physical mechanism of boiling heat transfer and the corresponding surface characteristics, to interpret the nucleate boiling heat transfer.

The ability to predict nucleate boiling heat transfer rates depends upon a knowledge of the mechanisms involved in the boiling heat transfer surface, and the growth and departure of the bubbles. There is no doubt that the growth and departure of the bubbles play an important role in the nucleation boiling heat transfer process. According to Mikic and Rohsenow [4], heat transfer by nucleate boiling is accomplished by the periodic removal of energy accumulated in the liquid that replaces the departing bubble, and hence, they concluded that the total heat flow rate Q at nucleate boiling can be separated into the heat flow Qmc by microconvection and Q,~ by natural convec- tion. This implies that the contribution of latent heat to total heat transfer is negligible. However, B16chl [5] showed that the contribution of latent heat cannot be omitted especially for higher heat flux and higher reduced pressure ; it is well known that the refrigerants used in refrigeration and air-conditioning industry generally fit into this category. Therefore, it is inter- esting to know the actual contribution of each heat transfer mechanism in the pool boiling process.

The main objective of the present study is to gain a better understanding of pool-boiling heat transfer through measuring the overall heat transfer coefficients of the R-22, R-134a, and R-124 refriger- ants. Experimental results are compared with the modified Bl6chl [5] semi-empirical model for pre- 1657

(2)

A heat transfer area [m 2] Cp specific heat [J kg -1]

Db bubble departure diameter [m] Do outside diameter [m]

Di inside diameter [m]

f bubble departure frequency [s-1] g gravitational acceleration [ms -z]

Gr Grashof number

h heat transfer coefficient [ W m -2 K -]] ifg latent heat [kJ kg -l]

Ja Jacob number

Jam modified Jacob number k wall thermal conductivity

[ W m ' K - ' ] / heating length [m]

M molecular weight [kg kmol-]] N bubble population

n bubble density [m -2] Pc critical pressure [kPa] Pr reduced pressure

Pr Prandtl number 0 heat flow rate [W]

q" heat flux [W m -2] NOMENCLATURE

Rp

rc

Twall

Tw,

Tw,

Tw2

Tw3

Tsatl

surface roughness [/am] critical radius [m]

wall surface temperature [K] average inner wall temperature [k] top wall temperature [K]

side wall temperature [K] bottom wall temperature [K] saturation temperature [K]. Greek symbols p density [kg m -3] (r surface tension [N m-l]. Subscripts e evaporation f fluid (liquid) g gas (vapor) mc microconvection nc natural convection t transient conduction tot total w wall.

dicting saturated nucleate boiling heat transfer rates, which included both the effects of the wall superheat and the effect of heat transfer surface characteristics.

EXPERIMENTALS

The schematic of the single-tube pool boiling appar- atus is shown in Fig. 1 (a). It consists of a cylindrical test vessel, a condenser and relevant connecting pipes made of stainless steel. Actually, the test setup is a natural circulation type apparatus. Heat is supplied to the tube by an internal cartridge heater, the evap- orated vapor refrigerant leaves the test section through the vapor pipe line, and condensed in a sep- arate condenser. The cylindrical boiling cell is made of stainless steel with a diameter of 88 mm and a length of 140 mm. The test vessel has a side- and a frontal-view window to observe the boiling phenom- ena.

The detailed geometries of the test tube is shown in Fig. l(b). The length of the test tube is 100 mm and has an outside diameter of 17.8 mm. Inside the test tube, a cartridge heater with diameter of 5.8 mm and a length of 95 mm was installed in the test tube. The cartridge heater, having a maximum power of 670 W, is coated by magnesium oxide and is covered by a stainless steel. The stainless steel tube has a copper sleeve with three grooves to locate the thermocouples. Three T-type thermocouples are installed in the cop- per sleeve located at 50 mm from the flange to measure

the temperature variations around the tube wall. Note that the location of the thermocouples are 90 ° apart as shown in Fig. 1 (b).

A pressure gauge calibrated with an accuracy of +0.01% is placed at the top of the test vessel to measure the system pressure. The liquid refrigerant temperature is recorded by two RTDs (resistance tem- perature device). All the thermocouples and RTDs were precalibrated by a quartz thermometer having a calibrated accuracy of 0.1 °C. A high resolution power supplier capable of measuring the current of 0.01 A is used to provide the power source of the heater.

EXPERIMENTAL PROCEDURES

The test section was initially cleaned and leak free before it was evacuated using a turbo-molecular vac- uum pump. The vacuum pump continued working for another 2 h after the vacuum gauge manometer

r e a c h e d 10 -4 torr, to ensure that it contained no non- condensable gases, then the refrigerants were charged into the system.

Pool boiling experiments were conducted for refrigerants R-22, R-124, and R-134a at saturation temperatures of 4.4 and 26.7°C and at reduced pres- sures of 0.1 and 0.2. The liquid refrigerants were gradually preheated to its corresponding saturated state before running each test. Power was then adjusted to a prior setting. The criterion of steady state condition was the variation of system pressure

(3)

Vapor line ~ - -

-

I ii

ii

II

1""-2 !V~[!2W ! Heater \ ~

Li~qu

/

id line

J

Condenser

C

CV: charging valve

DC: power supplier

P: pressure gauge

RV: relieve valve

TLI + TL2: liquid RTDs

TWl - TW3: wall TCs

WI - W2: view windows

(a)

Units: mm

F / L - ~ "

,o0

"1

,le

LC. ... J

- -

O-ring

0o~--- Thermocoul~

~

Copper

s oo,

_

~

~

~

MgO

~

Heater

180 °

(b)

Fig. 1. (a) Schematic diagram of experimental setup ; (b) detailed geometries of the test and test tube.

to be within _ 3 kPa and the temperature variations

of the wall surface were less than + 0.1 °C over 5 min.

All the data signals are collected and converted by a

data acquisition system (a hybrid recorder). The data

acquisition system then transmits the converted sig-

nals through GP]B interface to the host computer

for further operation. The experimental uncertainties

reported in the present investigation, following the

single-sample method proposed by Moffat [6], are

tabulated in Table 1. The maximum and minimum

uncertainties of the heat transfer coefficients were esti-

mated to be approximately 16.8% for Q = 2.6 W and

1.37% for Q = 668 W.

DATA REDUCTION AND EXPERIMENTAL RESULTS

The heat transfer coefficient for each power input

was calculated as follows :

0

h - A(Twal I __ T~at )

(1)

where Q is the electric heating power. Tsar is the satu-

ration temperature based upon the measurement of

system pressure. The outside surface area, A, is evalu-

ated as

nDol.

Note that Do is the outside tube diameter,

and l is the length of the cartridge heater (l = 95 mm),

(4)

Table 1. Summary of estimated uncertainties

Primary Derived Q = 2.6 W Q = 668 W

measurements Uncertainty quantities (minimum) (maximum)

l 1 0 - 3 m AT 14.1% 0.94%

Do 10 5m A 0.83% 0.83%

T 0.1°C O 9.17% 0.57%

I 0.01 A q" 9.21% 1.00%

V I V h 16.8% 1.37%

TwaH is the mean average wall temperature at the outer surface, which can be calculated from the measure- ment of inside temperatures, and is given by

ln(Do/Di)

Twall = Twi - - ~ ~ (2) where Tw~ is the arithmetic mean of three inside wall temperatures :

Zwl + 2Zw2-+- Tw3

Tw, - 4 (3)

where Tw,, Tw2 and Tw3 are the inside wall temperature readings, respectively.

Figure 2 shows the comparison of heat transfer coefficients vs heat flux for R- 134a refrigerant between the present data and those of Webb and Pais [7]. The saturation temperatures shown in the figure are 4.4 and 26.7°C, respectively, which are identical to the test conditions of Webb and Pais [7]. As seen, the

10 5

10 4

~ ' 10 3

R-134a at 26.7°C (,,) Webb and Pais [7] (t~) Present

data

Cooper correlation _

[8], Rp = 0

_

4

~

rl , , , A ~ / / ' - . . .

Mostinski

/ / ~ correlation [9]

/ / / ~ Steph~. and Abdeisam

/ - correlation [ 11

R-134a at 4.40C (o) Webb and Pais [7] (o) Present

data

104 Cooper correlation [8],Rp=0.4 ~ ~ , ~

,03

co edikoin gl

y j , , ' ~ S t e p h a n and Abdelsam / / correlation [1]

I

t

I

10 2 10 3 10 4 10 5 q" l W m "2]

Fig. 2. Comparison of R-134a data with Webb and Pais [3] and correlations of Cooper [8], Stephan and Abdelsalam [1]

and Mostinski [9].

present data agree favorably with those of Webb and Pais [7]. Figure 2 also shows the heat transfer coefficients predicted by Cooper [8], Stephan and Abdelsalam [1], and Mostinski [9]. The Cooper [8] correlation is given as

h = 9 0 ( q " ) ° - 6 7 M - ° - S P r ~ ( - l o g , o P r ) -°55 (4) m = 0.12 - 0.2 log l 0 Rp. (5) As reported by Stephan and Abdelsalam [I], the commercial finish copper tubes generally have a sur- face roughness of 0.4 #m. Therefore, the surface roughness, Rp, is given as 0.4/~m in the present cal- culation. The choice o f surface roughness, Rp, has little effect on heat transfer rate as depicted by Cooper [8]. It is shown that the present data agree closely with the Cooper correlation. A slight over-prediction of the Cooper correlation is found for a heat flux over 30 kW m -2. One of the explanations is that the slope of the h vs q" curve for the Cooper correlation (equa- tion (4)) is independent of heat flux (0.67), and the present data indicate that the slope decreases a small amount for higher heat flux. The decrease of slope is much more noticeable for the enhanced tubes as illustrated by Webb and Pais [7]. F o r comparison purpose, the Stephan and Abdelsalam [1] correlation and the Mostinski correlation [9] are also drawn in the figure. As seen, the Stephan and Abdelsalam [1] correlation and the Mostinski [9] correlation con- siderably underpredict the experimental data. Gener- ally, about 20--30% underpredictions for the Stephan and Abdelsalam [1] correlation are reported, and approximately 40-100% underpredictions for the Mostinski [9] correlation are shown. Webb and Pais [7] also reported an underprediction of 20-25% of the Stephan and Abdelsalam [1] correlation.

Figure 3 shows the heat transfer coefficients versus heat flux at reduced pressures o f 0.1 and 0.2 for R- 22, R-134a, and R-124. As seen in this figure, the deviations between the heat transfer coefficients are quite small for a given reduced pressure. This result reveals that the reduced pressure plays a significant role in correlating pool boiling heat transfer data. As seen in the Cooper correlation (equation (4)), the primary correlation parameter is the reduced pressure. Eventually, the Cooper [8] correlation can predict the present experimental data better than other corre- lations.

(5)

10 5 10 4 /z ' ~ 10 3 10 2 10 5 (a) P:= 0.1 (=) R-22 (0) R-124 (~) R-134a A A A i = 6 = i I r 10 3 10 4 10 5 q" [W m "2] 10 4 (b) 103 - P r = 0.2 (•) R-22 ~ (o) R-124 (t,) R- 134a A

• ~ | "4~"

J A

|

I I I 10 2 10 3 10 4 105 q" [W m "2]

Fig. 3. Pool boiling data of R-22, R-124 and R-134a (a) at

P r = 0.1 ; (b) a t P~ = 0.2. 105~ 10 4 '7, E e., (a) 10 3 -- 10 2 10 s T ~ t = 4 . 4 ° C ( • ) R - 2 2 ( e ) R - 1 2 4 (A) R - 1 3 4 a gnu

.::..

n l n ~ • ° at • I I t 10 3 10 4 10 5 q" [W m "2] 10 4 E -~ 10 3 (b) Tsat = 2 6 . 7 ° C (e) R-124 (a) R- 134a ~i

~"..

~ e A I I I 10 2 10 3 10 4 10 5 q " [ W m "2]

Fig. 4. Pool boiling data of R-22, R-124 and R-134a (a) at 4.4°C; (b) at 26.7°C.

Figure 4(a) and (b) presents the heat transfer coefficients versus heat flux at saturation temperatures at 4.4°C and 26.7°C, respectively. As seen, the heat transfer coefficients for R-124 are considerably lower than those of R-22 and R-134a. The main reason can be seen from tlhe physical properties of the tested refrigerants illustrated in Table 2. For a prescribed saturation temperature, the corresponding saturation pressure for R-22 and R-134a are higher than that of R-124. As a result, higher heat transfer coefficients for R-22 and R-134a are seen.

Figure 5 show the photographs of nucleate boiling for refrigerants of R-22, R- 134a and R- 124 at reduced pressure of 0.1 and 0.2 at a heat flux of 49.3 kW m -z. The increase of reduced pressure will decrease the bubble size and consequently the regime of isolated bubble will sustain at a higher heat flux. These figures show no significant distinctions between bubble size and frequency at the same reduced pressure.

Figure 6 shows the photographs of nucleate boiling at saturation temperature of 26.7°C and at heat fluxes

of 12.2, 49.3 and 110.7 kW m -2 for R-22 and R-124. Examination of the pictures indicates that the flow pattern is in isolated bubble regime for a heat flux of 12.2 kW m -2, and then moves to the regime of slugs and column at a heat flux of 110.7 kW m -2. The number of active nucleation sites increase sharply with increasing heat flux. For a given heat flux, the size of the bubble diameter for R-22 is smaller than that of R-124. This is because, as indicated in Table 2, the reduced pressure for R-124 is 0.118 compared to 0.218 for R-22. Consequently, it is expected that the bubble size of R-124 is much larger than R-22.

MODELING OF HEAT TRANSFER

In the modeling of saturated pool boiling from an active bubble site, the heat transfer rate can be con- sidered to be the sum of four principal components [10]. Namely, the contribution of microconvection due to bubble growth and departure, latent heat trans- port in the vapor bubble, natural convection and the

(6)

Table 2. Properties of tested refrigerants Reduced

M Pc ifg at 20°C Pr pressure Ts~t°C

(kg kmol- ~) (kPa) (kW kg- ') Tsa~ = 4.4°C T~, = 26.7°C Pr = 0.1 Pr = 0.2

R-22 86.48 4990 187.28 0.114 0.218 0.23 23.54

R-124 136.47 3634 147.96 0.053 0.111 23.36 47.53 R- 134a 102.03 4056 182.44 0.084 0.172 9.35 3 h82

sensible heat t r a n s f e r due to M a r a n g o n i flow. The fluid flow i n d u c e d by the t e m p e r a t u r e - r e l a t e d surface tension g r a d i e n t o n a l i q u i d / v a p o r interface is k n o w n as the t h e r m o c a p i l l a r y or M a r a n g o n i flow. T h e o r - etical calculations o f the t h e r m a l c o n t r i b u t i o n o f M a r - a n g o n i flow to the total h e a t flux is negligible, except in very h i g h s u b c o o l i n g [10]. Therefore, it is n o t

included in m o s t theoretical analysis. Mikic a n d R o h s e n o w [4] suggested t h a t the l a t e n t h e a t m a y n o t be a p r e d o m i n a n t f a c t o r in practical cases. C o n - sequently, m o s t p r e v i o u s investigations h a d divided pool boiling h e a t t r a n s f e r into two m e c h a n i s m , n a m e l y the c o n t r i b u t i o n o f m i c r o c o n v e c t i o n a n d o f n a t u r a l c o n v e c t i o n :

R-22, P r = 0.1 R-22, Pr = 0.2

R-124, Pr = 0.1 R - 1 2 4 , Pr = 0.2

R - 1 3 4 a , Pr = 0.1 R - 1 3 4 a , Pr = 0.2

Fig. 5. Photographs of boiling phenomena at Pr = 0.1 and Pr = 0.2 and at q" = 49.3 kW m -2 for R-22, R- 124 and R-134a.

(7)

R-22, q" = 12.2 kW m -2 R-124, q " = 12.2 kW m 2

R-22, q" = 49.3 kW m -2 R-124, q" = 12.2 kW -2

R-22, q"= 110.7 kW m 2 R-124, q " = 110.7 kW m -2 Fig. 6. Photographs of boiling phenomena at T s a t = 26.7°C for R-22 and R-124 at q" = 12.2, 49.3 and

110.7 kW m 2

An¢

-- =

hA

T = A 0mc + A-- q,c (6) A

where qmc is the average of the microconvection heat flux over the time interval, and qnc is due to natural convection heat flux. Using the conjugate error func- tion solution to evaluate the heat flux from the surface to this region yields :

2 t~mc- x ~ ~ x ~ A T =

htAT

Amc =

nND~

therefore, Anc= A--Amc

(7)

(8)

(9)

(10)

The contribution of natural convection, q,o is given by

q.c = hnc(Tw~lJ- Tsat) (11) where h~c is the heat transfer coefficient of natural convection from horizontal cylinder proposed by Bay- ley [11], i.e.

hnc = ~O.l(GrPr) (I/3).

(12) Equation (7) suggests that the heat transfer in mic- roconvection is only due to the transient conduction into the sublayer. This assumption was confirmed for boiling of water at low pressure. However, B16chl [5] showed that the contribution of latent heat transfer to the total heat transfer rate cannot be neglected. He adds a second term to equation (7) for hmc to take into account the latent heat transport effect resulting from

(8)

et al.

the superheated liquid evaporation at the surface of the bubbles :

_ ~ n ~ ~ f

1

q:

hmcAT

= A T +

gpgifgfOb

= q't'+

(13)

w h e r e D b and f a r e the bubble departure diameter and bubble departure frequency, respectively. To apply the above-mentioned model, information for the bub- ble departure diameter and bubble departure fre- quency must be available. The bubble diameter can be calculated using the Cole and Rohsenow cor- relation [12], and is given by

Db = C(Jam) 5/4 ~g(pf__a

Pg) (14)

Tsat Cpfpf

Jam

(15)

pgif,

where C = 1.5 × 10 -4 for water ) ? 4.65 x 10 -4 for other fluidsJ' The correlation for the prediction of bubble frequency used in the present investigation is taken from Zuber [131:

fDb = 0.59 [ag(Pf~ Pg) ]

TM.

(16)

L

Pf

J

The active nucleation site density can be derived from equations (6), (7), (8) and (1 l) :

N 1 h - h,c

- n - - - ( 1 7 )

A 7rD 2 hmc - hnc" b

A unique feature of B16chl's [5] model is the way the active nucleate site density is determined from the size distribution of active nucleation cavities on the heat transfer surface. It is generally agreed [14-16] that the pool boiling active nucleation site density can be determined as a function of the cavity radius in the form

n = C ~ 1 (18)

where rc is the minimum cavity radius at a specified condition, and is given by

2aTsat

rc -- (19)

pgif~AT"

The constants C and m 1 in equation (18) characterize the boiling surface, and can be obtained from the experimental data.

Bier

et al.

[17] suggested an alternative form to represent the active nucleation distribution in the form of a Rosin-Rammler-Sperliling distribution :

In(N/A) = ln(N/A)max[1--(re/rmax)

mE]

(20)

where rm,x is the total number of the nucleation sites available on the heat transfer surface.

Based on their experimental data, Jamialahmadi

et

al.

[ 18] proposed a simple exponential form:

ln(N/A) = A, + A2rc + A3r~.

(21)

Despite the fact that these previous works have achieved significant progress in theories, it seems that these forms cannot accurately reflect the dependence of the nucleation site density on the boiling surface conditions (Yang and Kim [19]). Kocamustafagullari and Ishii [20] indicated that equation (19) is difficult to use in practice. Generally, the empirical constants used in these above-mentioned equations are only applicable to their own data. It is noted that the active nucleation sites should depend upon other refrigerant properties such as surface tension. Therefore, an empirical form of active nucleation sites is developed based on the present R-22, R-134a, and R-124 to yield :

n = (e")/Ja"2

(22) where Y(1 + M °'2) z = (23) (1 + Pr)(1.0-- 10a) Y = 6 4 0 . 9 3 - 4 5 . 8 7 X + 1.117X 2 - 2 9 9 6 / X (24) X = ln(1/r~) (25)

CpfpfAT

Ja

= (26) pgifg

We then try to apply the correlation to predict experimental data from other researchers.

DISCUSSION OF THE MODEL

Figure 7(a) shows the heat transfer rate for the contributions of transient conduction, latent heat transport, and natural convection at a reduced pres- sure of 0.2 and of 0.01 for R-123. As seen, the latent heat transfer is the dominant heat transfer mechanism for a given reduced pressure of 0.2 and at a wall superheat less than 2°C. The contribution of transient conduction to total heat transfer increases sharply with the increase of wall superheat, and is comparable to latent heat transfer for a wall superheat of 5°C. The natural heat transfer does not show a noticeable increase vs wall superheat. Despite the increase of wall superheat results, an increase of the natural con- vection component, the natural convection heat flux, reveals maximum characteristics vs wall superheat. The reason for this phenomenon can be explained from the change of 'influence area' of departing bubbles. As shown in Fig. 6, the active bubble sites increase significantly with heat flux (wall superheat). Therefore, the 'influence area', Amc, for a departing bubble is increasing with the wall superheat, and eventually reduces the surface area for natural con-

(9)

10 ~ 10 4 103

E

'~ 1021- 1011 l0 o R-123 / / / / q e Pr = 0.2 / / ~ - - p r = 0.01 / /

/

AT [K]

l

101 10 6 10 5 10 4 10 3 10 2 1 0 1 - - ~

(b,

J A

- Water / ; / q ~ Pr = 0.2 / / Pr = 0.004 //t

g

/

/ I 100 10 ! AT [K]

Fig. 7. The contributions of transient conduction, latent heat, and natural convection heat flux for (a) R-123 at P~ = 0.2

and P~ = 0.01 ; (b) water at Pr = 0.2 and P~ = 0.01.

E

O "3 O t~ i . " O O ~3 .o 104 10 3

(a)

,,v~o~

~

( ' ) R-22 x ' r ~ k , 4 r -

(.) R-124

.Z]~Ai~#

Z / I I 103 104

Experimental heat transfer coefficient [W m "2 K l l

A ~ 10 4

{

/ / ~ ' ~ (GO) R-] 1, Webb and Pals

~o | ~ . ~ 1 ~ ' = (O) R-12, Webb and Pais

,.. ~')~],,,,I (O) R-22, Webb and Pals 103_ ~ - ¢ / / (D) R-123, WebbandPais

/ ~ / (A) R-134a, Webb and Pais

l o 3 lO 4

Experimental heat transfer coefficient [W m "2 K "l] Fig. 8. Comparisons between the proposed prediction method and (a) the present experimental data ; (b) the exper-

imental data from refs. [7] and [22].

vection. The net result is a maximum phenomenon for the natural co~:vection component. In the present investigation, the 'influence area' ratio is taken to be 4 as suggested by Hsu and Graham [21].

For a lower reduced pressure (Pr = 0.01), the dis- tributions of the transient conduction, latent heat, and natural convection are quite different from those at higher reduced pressure. The effect of natural con- vection generally cannot be omitted, and both of the transient conduction and latent heat increases sig- nificantly with the wall superheat. However, the ratio,

q'~'/q'(,

decreases with the wall superheat, and the latent heat contribution may be discarded for a heat flux over 10 000 W m -2. Figure 7(b) shows the heat trans- fer rate for the irdividual contribution of transient conduction, latent heat transfer, and natural con- vection at a reduced pressure of 0.2 and of 0.004 for water. As seen, the distributions of these heat transfer components for water is analogous to those of R-123.

Figure 8(a) and (b) shows the predictions for the proposed method with the present experimental data, and the experimental data of R-1 I, R-12, R-22, R-123 and R- 134a from Webb and Pais [7] and of R-22 from Gorenflo and Fath [22]. As shown in Fig. 8(a), the present method can predict 95% of the experimental data within 20%. In addition, based on the present experimental data bank, the proposed method can also predict the experimental data from Webb and Pais [7] and Gorenflo and Fath [22] with reasonably good accuracy as depicted in Fig. 8(b). Actually, 75% of the experimental data can be predicted within 20%.

C O N C L U S I O N S

(1) Pool boiling data for R-22, R-124 and R-134a on a plain tube are reported at saturation tem- peratures of 4.4°C and 26.7°C and at reduced pres- sures of 0.1 and 0.2. For a given saturation tempera-

(10)

ture, the heat transfer coefficient for R-124 is the smallest due to its low reduced pressure. Also, it is shown that the C o o p e r [8] correlation can predict the present pool boiling heat transfer quite satisfactorily. (2) A modified BlSchl [5] m e t h o d is proposed in this study; this semi-analytical model not only can predict the present data with success but also predict the experimental data from W e b b and Pais [7] and of R-22 from Gorenflo and F a t h [22] with reasonably good accuracies.

(3) The modified B16chl [5] m e t h o d suggests that the governing heat transfer mechanism depends on the reduced pressure and wall superheat. F o r a reduced pressure greater than 0.2, the natural convection is almost negligible, and the d o m i n a n t heat transfer mechanism is latent heat transport instead of transient conduction. However, for a reduced pressure of 0.01, the contribution o f natural convection generally can- not be omitted, and the contribution o f transient con- duction is higher than that of the latent heat. In addition, the transient conduction becomes the con- trolling heat transfer mechanism at higher wall super- heat.

REFERENCES

1. Stephan, K. and Abdelsalam, M., Heat transfer cor- relations for natural convection boiling. International

Journal of Heat and Mass Transfer, 1980, 23, 73-87.

2. Cooper, M. G., Heat flow rates in saturated nucleate pool boiling--a wide-ranging examination using reduced properties. Advances in Heat Transfer, 1984, 16, 157- 239.

3. Kolev, N. I., How accurate can we predict nucleate boil- ing. Experimental Thermal and Fluid Science, 1995, 10, 370-378.

4. Mikic, B. B. and Rohsenow, W. M., A new correlation of pool-boiling data including the effect of heating surface characteristics. Journal of Heat Transfer, 1969, 91, 245- 250.

5. Bl6chl, R., Zum Einflul3 der Oberfl/ichenstruktur unter- shiedlich bearbeiteter Heizfl~ichen auf die W/ir- mefibertragung beim Blasensiden. Ph.D. dissertation, University of Karlsruhe, 1986.

6. Moffat, R. J., Describing the uncertainties in exper- imental results. Experimental Thermal and Fluid Science, 1988, 1, 3-17.

7. Webb, R. L. and Pais, C., Nucleate pool boiling data for

five refrigerants on plain, integral-fin and enhanced tube geonertries. International Journal of Heat and Mass

Transfer, 1992, 35(8), 1893-1904.

8. Cooper, M. G., Saturation nucleate pool boiling-- a sim- ple correlation. International Chemical Engineering Sym-

posium Series, 1984, 86, 785-792.

9. Mostinski, I. L., Calculation of heat transfer and critical heat flux in boiling liquids based on the law of cor- responding states. Teploenergetika, 1963, 10, 66-71. 10. Tong, W., Bar-Cohen, A. and Simon, T. W., Thermal

transport mechanisms in nucleate pool boiling of highly- wetting liquids. In Ninth International Heat Transfer

Conference, Vol. 2, 1990, pp. 27-32.

11. Bayley, F. J., An analysis of turbulent free convection heat transfer. Proceedings of the Institute of Mechanical

Engineering, 1955, 169, 361.

12. Cole, R. and Rohsenow, W. M., Correlation of bubble departure diameters for boiling of saturated liquids.

Chemical Engineering Progress Symposium Series, 1966,

65, 6-16.

13. Zuber, N., Hydrodynamic aspects of boiling heat trans- fer. U.S. AEC Report, AECU 4439, 1959.

14. Griffith, P. and Walis, G. D., The role of surface con- ditions in nucleate boiling. Chemical Engineering Pro-

gress Symposium Series, 1960, 56, 49 63.

15. Judd, R. L. and Shouki, M., Nucleate boiling on an oxide coated glass surfaces. A S M E Journal of Heat

Transfer, 1975, 97, 494--495.

16. Fujita, Y., Ohta, H., Hidaka, S. and Nishikawa, K., Nucleate boiling heat transfer on horizontal tubes in bundles. In Eighth International Heat Transfer Confer-

ence, 1986, Vol. 5, pp. 2131-2136.

17. Bier, K., Gorenflo, M. and Salem, M., Size distribution of surface cavities and of stable nuclei for pool boiling heat transfer to refrigerants. In Fifteenth International

Congress of Refrigeration, Venice, Italy, 1979, paper B1-

91.

18. Jamialahmadi, M., B15chl, R. and Mfiller-steinhagen, H., Pool boiling heat transfer to saturated water and refrigerant 113. The Canadian Journal of Chemical Engin-

eering, 1991, 69, 746-754.

19. Yang, S. R. and Kim, R. H., A mathematical model of the pool boiling nucleation site density in terms of the surface characteristics. International Journal of Heat and

Mass Transfer, 1988, 31, 1127-1135.

20. Kocamustafagullari, G. and Ishii, M., Interracial area and nucleation site density in boiling system. Inter-

national Journal of Heat and Mass Transfer, 1983, 26,

1377-1387.

21. Hsu, Y. Y. and Graham, R. W., NASA TN D-594, 1961. 22. Gorenfio, D. and Fath, W., Heat transfer at pool boiling outside of finned tubes at high saturation pressures. In

Proceedings of the Seventeenth International Congress of Refrigeration, Wien, Vol. B, 1987, pp. 955-968.

數據

Fig.  1. (a) Schematic diagram of experimental  setup ; (b) detailed geometries of the test and test tube
Fig. 1. (a) Schematic diagram of experimental setup ; (b) detailed geometries of the test and test tube p.3
Figure  2  shows  the  comparison  of  heat  transfer  coefficients vs heat flux for R- 134a refrigerant between  the present data and  those of Webb and  Pais  [7]

Figure 2

shows the comparison of heat transfer coefficients vs heat flux for R- 134a refrigerant between the present data and those of Webb and Pais [7] p.4
Table 1. Summary  of estimated  uncertainties

Table 1.

Summary of estimated uncertainties p.4
Fig. 3. Pool boiling data of R-22, R-124 and R-134a (a) at
Fig. 3. Pool boiling data of R-22, R-124 and R-134a (a) at p.5
Fig.  5. Photographs of boiling phenomena at Pr =  0.1 and Pr =  0.2 and at  q"  =  49.3 kW m -2 for R-22,  R-  124 and R-134a
Fig. 5. Photographs of boiling phenomena at Pr = 0.1 and Pr = 0.2 and at q" = 49.3 kW m -2 for R-22, R- 124 and R-134a p.6
Table 2. Properties of tested refrigerants  Reduced

Table 2.

Properties of tested refrigerants Reduced p.6
Figure  8(a)  and  (b)  shows  the  predictions for  the  proposed method with the present experimental data,  and the experimental data of R-1 I, R-12, R-22, R-123  and R- 134a from Webb and Pais [7] and of R-22 from  Gorenflo  and  Fath  [22]

Figure 8(a)

and (b) shows the predictions for the proposed method with the present experimental data, and the experimental data of R-1 I, R-12, R-22, R-123 and R- 134a from Webb and Pais [7] and of R-22 from Gorenflo and Fath [22] p.9
Fig. 7. The contributions of transient conduction, latent heat,  and natural convection heat flux for  (a)  R-123 at  P~ =  0.2
Fig. 7. The contributions of transient conduction, latent heat, and natural convection heat flux for (a) R-123 at P~ = 0.2 p.9

參考文獻

相關主題 :