Experimental study of Gas-Water Two-Phase flow in a bubble Column: Insights into the Transition and structure of flow in pool scrubbing tests

• Gas-liquid two-phase flow development in a bubble column was measured using a Wire-Mesh Sensor. • 3D flow field data were reconstructed from stacked axial profiles of void fraction and gas velocity. • Quantitative data on bubble diameter change from the nozzle exit to the water surface were obtained. • The dataset is intended for CFD model benchmarking, and part of the results is publicly available. Pool scrubbing is a key mechanism for reducing aerosol and iodine releases in severe accidents, yet the decontamination factor (DF) depends strongly on the gas‑liquid two‑phase flow. To improve mechanistic understanding and support CFD‑based analysis, we performed air‑water experiments in a large vertical bubble column (diameter 500 mm , height 8 , 000 mm ) at ambient conditions. Gas flow rates ranged from 100 - 2 , 000 L mi n -1 (superficial velocity up to ≈ 0.17 m s -1 ), five water depths ( 500 - 2 , 500 mm ) were tested, and a 40 mm nozzle was used for injection. Time–resolved wire–mesh sensors (WMS) yielded time–averaged 3D fields of void fraction and gas–phase velocity and enabled bubble‑size distributions at each height; in this work we summarize the axial evolution using the percentiles D B 16 , D B 50 , and D B 84 and present representative distributions at selected elevations. At high gas flows, bubble diameters decrease rapidly from the nozzle outlet, with the main breakup occurring over a 200 - 900 mm region and a stabilized distribution emerging by approximately 900 mm , whereas at low gas flows this process is slower, requiring on the order of 1 , 500 - 2 , 000 mm for the distribution to approach a uniform small‑bubble regime. Water depth strongly influenced the flow: high levels yielded rapid stabilization and cross‑sectional spreading, while shallow pools kept gas concentrated near the axis, allowing larger bubbles to reach the surface before a stabilized size distribution formed. The dataset is provided as time‑averaged quantities to serve as practical benchmarks for steady‑state CFD with RANS turbulence modeling; selected files are supplied as supplementary material. The bubble‑size profiles also aid interpretation of DF trends measured in companion tests with the same facility and can inform the verification and refinement of coalescence‑breakup models in severe‑accident/system codes (e.g., MAAP, MELCOR).