Flow boiling heat transfer in small tubes

W , H max(

) W , H

min( , H and

W are the height and width of the channel, respectively.

Recently in an experiment for R-134a liquid flow in mini-channels, Agostini et al.

[6] compared their heat transfer data with some existing correlations. They concluded that the correlation of Shah and London [7] was good for Re <500 and the Gnielinsky correlation [8] was appropriate for 2,300< Re <10⁶ and 0.6< Pr <10⁵. Moreover, the Dittus-Boelter correlation [9] was suitable for Re >10⁵ and 0.7< Pr <16,700.

1.2.2 Flow boiling heat transfer in small tubes

Heat transfer associated with flow boiling in a conventional channel is normally considered to result from two mechanisms: nucleate boiling and convective boiling. In the situation with a dominant nucleate boiling mechanism, the boiling heat transfer coefficient is mainly dependent on the imposed heat flux and the refrigerant saturated temperature or pressure. While for the cases with a predominated convective boiling mechanism, heat transfer coefficient is mainly dependent on the refrigerant mass flux and vapor quality. Therefore in the study of flow boiling heat transfer in mini- and micro- channels, which mechanism is prevalent is the main issue of many investigations.

Yu et al. [10] recently examined flow boiling heat transfer for water in a 2.98 mm diameter channel. They found that the boiling heat transfer of water in the small channel (vapor quality above 0.5) was dependent on the heat flux but was independent of the mass flux. They concluded that the nucleate boiling was dominant over the

convective boiling in small channels. The results are significantly different from these for the conventional channels, where the mass flux effect can be substantial. Similarly, Sumith et al. [11] measured the saturated flow boiling heat transfer and pressure drop of water in the test section which was made of a stainless steel tube with an inner diameter of 1.45 mm. They indicated that the dominant flow pattern in the tube was a slug-annular or an annular flow, and liquid film evaporation dominated the heat transfer.

An experimental measurement carried out by Yan and Lin [12, 13] to study the evaporation heat transfer and pressure drop of R-134a in a tube bank forming by 28 small pipes (Di=2.0 mm) revealed that both the nucleate and convective boiling mechanisms were important and the evaporation heat transfer in the small pipes was significantly higher than that in large tubes. A visualization investigation was presented by Nino et al. [14] to examine R-134a refrigerant in a multiport microchannel tube with Dh=1.5 mm. They proposed a method to describe the fraction of time or the probability that a flow pattern existed in a particular flow condition. A recent study from Fujita et al. [15] for R-123 boiling in a horizontal small tube with an inside diameter of 1.12 mm suggested that heat transfer in the flow was dominated by the nucleate boiling and the effects of the refrigerant mass flux and vapor quality to the boiling heat transfer were very weak. A similar study for R-113 boiling conducted by Lazarek and Black [16] noted the negligible variation of the boiling heat transfer coefficient with the local vapor quality, which implied that the wall heat transfer process was again controlled by nucleate boiling. In a vertical small tube with an inside diameter of 1 mm with R-141b refrigerant flowing in it, Lin et al. [17] found that at low quality, nucleate boiling dominated. But at higher quality, convective boiling dominated. In a further study [18], they examined the same refrigerant in four

mm². Their results indicate that the mean heat transfer coefficient in a tube or channel is independent of the mass flux and tube diameter but is a function of the imposed heat flux. Cornwell and Kew [19] observed boiling in refrigerant R-113 flow and measured the heat transfer for two geometries: one had 75 channels with Dh=1.03 mm and other had 36 channels with Dh=1.64 mm. Their experimental work suggested the presence of three two-phase flow patterns in the channels: isolated bubble, confined bubble and annular-slug flow. In a continuing study [20], they investigated refrigerant R-141b boiling in a horizontal tube with its inner diameter ranging from 1.39 to 3.69 mm and proposed that flow boiling in a narrow channel might be through one of four mechanisms: nucleate boiling, confined bubble boiling, convective boiling and partial dry-out. They further indicated that, except at very low heat flux, the boiling showed a strong dependence on the heat flux, a weak dependence on the mass flux, and independence of the quality. Besides, they introduced a new dimensionless group

named as the confinement number,

1/ 2 conf

h

[ /(g( ))]

N D

l g

σ ρ ρ−

≡ , which represented the

importance of the restriction of the flow by the small size of the channel. The dimensionless number Nconf can be used to find the transition from the isolated to confined bubble regimes. To a first approximation, they showed that the confined boiling occurred when Nconf >0.5. The effects of the tube confinement were found to be significant for micro- and mini-channels.

Examining the boiling of refrigerants in a small circular tube (D=2.46 mm) and a rectangular duct (Dh=2.40 mm) with nearly the same hydraulic diameters, Tran et al.

[21] showed that there was no significant geometry effect for the two channels tested.

Furthermore, their results implied that the nucleation mechanism dominated over the convection mechanism in small-channel evaporators over the full range of quality (0.2

~ 0.8), which was contrary to the situations in larger channels where the convection

mechanism dominates at qualities typically above 0.2. Bao et al. [22] also found that the boiling heat transfer coefficient was a strong function of the heat flux and system pressure in a 1.95 mm diameter tube with R-11 and R-123, while the effects of the mass flux and vapor quality were very small, suggesting that the heat transfer was mainly through the nucleate boiling. Wambsganss et al. [23] studied boiling heat transfer of refrigerant R-113 in a small diameter (2.92 mm) tube and evaluated 10 different heat transfer correlations. They found that the high boiling number and slug flow pattern led to the domination by the nucleation mechanism and the two-phase correlations based on this dominance also predicted the data best. Moreover, Warrier et al. [24] used FC-84 in five parallel channels with each channel having a hydraulic diameter of 0.75 mm and compared their results with five widely used correlations.

They then proposed two new correlations, one for subcooled flow boiling heat transfer and the other for saturated flow boiling heat transfer. Oh et al. [25] examined R-134a in capillary tubes of 500 mm long and 2, 1 and 0.75 mm inside diameters. They concluded that the heat transfer in the forced convective boiling was more influenced by the refrigerant mass flux than by the boiling number and the heat transfer coefficient was controlled by the Reynolds number. Besides, their results also showed that the dry-out point moved to the lower quality with decreasing size of the tubes.

Vaporization of CO2 in 25 flowchannels of 0.8 mm ID was recently examined by Pettersen [26]. He also observed the two-phase flow pattern with another separate test rig with 0.98-mm heated glass tube. The results showed that nucleate boiling dominated at low/moderate vapor fractions, where the boiling heat transfer coefficient increased with the heat flux and refrigerant temperature but was less affected by the mass flux and vapour fraction. Moreover, the dryout effects became very important at higher mass flux and temperature, where the boiling heat transfer coefficient dropped

regimes, the latter becoming more important at high mass flux.

More complete information on the two phase flow boiling heat transfer in compact evaporators or microchannels is available from the recent critical review conducted by Ghiaasiaan and Abdel-khalik [27], Thome [28], Sobhan and Garimella [29], Kandlikar [30, 31] and Watel [32].