Two-phase flow on the shell side of a shell and tube heat exchanger is complex. Several studies have produced flow pattern maps that show surprising differences in flow regime boundaries for data sets that contain relatively small variations in fluid and flow properties. Despite this, correlations for void fraction and pressure drop are sufficiently accurate to allow the thermal-fluid design of heat exchangers to be completed. However, these correlations are based on experimental data taken from tube bundles containing tubes with diameters less than 20 mm. This study examines their applicability to tube bundles containing tubes with a diameter of 38 mm. Results for void fraction and pressure drop are presented for air-water flows near atmospheric pressure. The results were obtained for flows through a thin-slice, in-line tube bundle containing 10 rows. The tube bundle contained a central column of tubes with half tubes placed on the shell wall to simulate the presence of other columns. The tubes were 38 mm in diameter and 50 mm long with a pitch to diameter ratio of 1.32. Previous studies have shown that the void fraction in a shell-side, gas-liquid flow becomes constant after only a few rows. Thus, the void fraction was only measured at one location. A single-beam, gamma-ray densitometer was used to measure void fractions near row 7 in the maximum gap between the rows. Corresponding pressure drops were obtained between rows 3 and 10. Data are presented for a mass flux range of 25–688 kg/m2s and a gas mass fraction range of 0.0005–0.6. The measurements are shown to compare reasonably well with predictions from correlations available in the open literature. This shows that these methods can be used for tube-bundles containing larger diameter tubes. Some elements of a heat-exchanger design require a more complex analysis. For example, tube vibration calculations require the distribution of void and phase velocity along the tube length. Such analysis can be provided by multiphase computational fluid dynamic (CFD) simulations. CFD approaches to modelling these flows require empirical inputs for the drag coefficient and the force on the fluid by the tubes. These are deduced from the measured data. The wall forces are shown to scale well with increased tube diameter, however, caution is required when selecting the drag coefficients.

This content is only available via PDF.
You do not currently have access to this content.