Abstract
The mass transfer between a rising bubble and the surrounding liquid is mainly determined by an extremely thin layer of dissolved gas near the bubble interface. Resolving this concentration boundary layer in numerical simulations is computationally expensive and limited to low Péclet numbers. Subgrid-scale models mitigate the resolution requirements by approximating the mass transfer near the interface. In this contribution, we validate an improved subgrid-scale model with a single-phase simulation approach, which solves only the liquid phase at a highly-resolved mesh. The mass transfer during the initial transient rise of moderately deformed bubbles in the range Re = 72–569 and Sc = 102–104 is carefully validated. The single-phase approach is able to mirror the two-phase flow field. The time-dependent local and global mass transfer of both approaches agree well. The difference in the global Sherwood number is below than 2.5%. The improved subgrid-scale model predicts the mass transfer accurately and shows marginal mesh dependency.
Original language | English |
---|---|
Article number | e17641 |
Number of pages | 17 |
Journal | AIChE Journal |
Volume | 68 |
Issue number | 7 |
DOIs | |
Publication status | Published - Jul 2022 |
Keywords
- high-Schmidt number problem
- machine learning
- mass transfer
- multiphase flows
- subgrid-scale modeling