Flow boiling of R134a in vertical mini-diameter tubes

Abstract

The current study is a part of a long term experimental project devoted to investigate flow boiling heat transfer, pressure drop and flow visualization of R134a in small to mini/micro-diameter tubes. The experimental facility was first designed and constructed by X. Huo (2005) with the contribution of L. Chen (2006). In the present study, the experimental facility was upgraded by changing the heating system from AC to DC heating and also upgrading the logging system through using a faster data logger and developing a new Labview program. The objectives of the current study include (i) contribute in identifying the reasons behind the wide scatter in the published flow boiling heat transfer results, (ii) contribute in understanding the fundamentals of flow boiling heat transfer in mini/micro-diameter tubes and (iii) evaluation of the existing heat transfer and pressure drop prediction methods. Two sizes of stainless steel tubes were investigated in the current study; 0.52 mm and 1.1 mm diameter. In the current study, the 0.52 mm tube was roughly called a “micro-tube” whilst the 1.1 mm tubes were called “mini-tubes”.

The present study proposes two possible reasons for the scatter in the published heat transfer results. The first reason is the variations in the heated length from one study to another–there is no criterion for choosing the heated length. The second reason is the variations in the inner surface characteristics of the channels from one study to another. These two important parameters were not taken into consideration by researchers in the past studies. Accordingly, the effect of the heated length was investigated in the current study using a seamless cold drawn tube with diameter of 1.1 mm and heated length ranging from 150 to 450 mm. The effect of the tube inner surface was also tested here by conducting the test in two stainless steel tubes with diameter of 1.1 mm and manufactured by two different processes. The first tube was manufactured by welding technique whilst the second tube was a seamless cold drawn tube. Both tubes were identical in design and dimensions. The inner surface of each tube was examined first using SEM analysis and demonstrated that, the surface morphology is completely different. The local heat transfer coefficient was determined through measuring the local wall temperature using 14 K-type thermocouples attached to the wall using thermally conducting but electrically insulating epoxy supplied by Omega. Pressure drop was measured directly across the heated section and a high speed camera was used for the flow visualization at 1000 frames/s. All measurements were recorded after the system attained steady state. The experimental conditions include mass flux range of 100 – 500 kg/m2 s, system pressure range of 6 – 10 bar, inlet sub-cooling of about 5K and exit quality up to about 0.9.

The most frequently observed flow regimes in the 0.52 mm tube were found to be slug (elongated bubble), transition to annular and annular flow regimes. In the 1.1 mm tube, the observed regimes were found to be slug, churn and annular. The transition from slug flow to annular flow in the 0.52 mm tube occurred smoothly with little disturbances at the liquid vapour interface compared to the 1.1 mm tube. Additionally, increasing the heated length of the 1.1 mm tube was found to shift the transition to annular flow to occur at lower vapour quality.

The heat transfer results demonstrated that the behaviour of the local heat transfer coefficient in the 0.52 mm diameter tube is different compared to that in the 1.1 mm tubes. Also, the tube inner surface characteristics and the heated length were found to strongly influence the local behaviour of the heat transfer coefficient. Flow boiling hysteresis was investigated and the results indicated that hysteresis exists only at very low heat fluxes near the boiling incipience.

Existing heat transfer and pressure drop correlations were examined using the results of the 0.52 and 1.1 mm seamless cold drawn tubes. The pressure drop data were predicted very well using the Muller-Stienhagen and Heck (1986) correlation, the homogeneous flow model and the correlation of Mishima and Hibiki (1996). On the contrary, all macro and microscale heat correlations failed to predict the current experimental data. The mechanistic models failed to predict the data of all tubes with the same accuracy. Accordingly, two heat transfer correlations were proposed in the current study. The first correlation is based on dimensionless groups whilst the second is based on the superposition model of Chen (1966). Both correlations predicted the current experimental data and the data of Huo (2005) and Shiferaw (2008) very well.