Steam bubble condensation in sub-cooled vertical pipe flow

Detailed experimental data on the evolution of a two-phase flow along a vertical pipe with an inner diameter of 195 mm and a length of 8 m were obtained for adiabatic air-water mixture as well as for steam-water mixture under adiabatic and non-adiabatic conditions at different pressure. The experiments were performed at the TOPFLOW facility of our institute. The measurements were carried out using the wire-mesh sensor technology. The data include local gas volume fraction values (lateral resolution: 3 mm) as well as bubble size distributions and local volume fraction distributions decomposed according to bubble size classes. A slight sub-cooling of the water (max. 4 K) was achieved by throttling the globe valve at the upper end of the vertical test section.

Since the wire-mesh sensor disturbs the flow in downflow direction (but still measures the undisturbed flow!) it doesn’t make sense to place a number of wire-mesh sensors behind each other in order to investigate the evolution of the flow along the pipe. Because of the large dimensions of the facility it is also not applicably to shift a single sensor to different height positions by dismantling the facility each time. For this reason the measuring plane was fixed at the upper end of the test section and a variable gas injection was used. The gas was injected by rings of holes at the pipe wall, which are placed at different height positions 1 mm and 4 mm holes can be used for the injection.

The previously developed Multi-Bubble-Size-Group Test Solver was extended to consider the phase transfer. The large number of bubble classes (50) in the simulation allows the investigation of the influence of the bubble size distribution. The interfacial area density is calculated according to the bubble size distributions and the assumption of ellipsoidal bubbles with a deformation according to the well-known Wellek-correlation. Simulations show, that there are clear differences in the condensation process along the pipe for poly-dispersed flow in comparison with the assumption of mono-dispersed flow is discussed.

The results of the simulations show a good agreement with the experimental data. The condensation process is clearly slower, if large bubbles are injected (4 mm holes). Also bubble break-up has a strong influence on the condensation process because of the change of the interfacial area. Some unsureness arises from the unknown interfacial area for large bubbles and possible uncertainties of the heat transfer coefficient.