Flow and bubble statistics of turbulent bubble-laden downflow channel

P. Cifani, J.G.M. Kuerten, B.J. Geurts

Research output: Contribution to journalArticleAcademicpeer-review

7 Citations (Scopus)
8 Downloads (Pure)


Direct numerical simulations of fully developed turbulent channel downflow at bulk Re equal to 6300, loaded with monodisperse bubbles at gas volume fractions α=0.5% , α=2.5% and α=10 have been carried out. Bubble deformability, surface tension, as well as discontinuity in the material properties across the bubble interfaces are explicitly accounted for. A full-scale channel of size 4πH × 2H × 4πH/3 in terms of the channel half-width H containing a number of bubbles up to O(103) is considered. The statistical structure of the bubbles, the probability density function describing the bubble velocity and the liquid kinetic energy spectra have been determined. A close range preferential clustering of the bubbles was found with a maximum density independent of the gas volume fraction at a separation distance of about 2.2R, with R the bubble radius. Preferential horizontal alignment and a general tendency to repulsion is shown for separation distances smaller than 3R. At larger separation distances a close to random distribution is observed for α=2.5% and α=10%, while tendency to vertical alignment is observed for α=0.5% . The pdf of the bubble velocity fluctuations was found to be well approximated by a Gaussian distribution. The liquid kinetic energy spectra in the channel core do not show a marked -3  scaling, which was previously reported for homogeneous isotropic turbulence and pseudo-turbulence.
Original languageEnglish
Article number103244
Number of pages17
JournalInternational journal of multiphase flow
Early online date12 Feb 2020
Publication statusPublished - May 2020


  • Bubbly flow
  • Turbulent channel flow
  • DNS
  • Bublle statistics
  • Kinetic energy spectra
  • UT-Hybrid-D
  • 22/2 OA procedure


Dive into the research topics of 'Flow and bubble statistics of turbulent bubble-laden downflow channel'. Together they form a unique fingerprint.

Cite this