Development of a multiscale solver with sphere partitioning tracking

Abstract

A new method to compute the dispersed phase in a Lagrangian framework is shown in this contribution for computing incompressible bubbly flows. Each bubble is divided dynamically in equivolumetric elements and tracked into the Eulerian mesh for an appropriate assignment of the effect of the bubble in the cell. The coupling between phases is done considering in the momentum equation the interfacial forces along the bubble path during an Eulerian time step. The bouncing of the bubbles between themselves and the wall is modeled with a dynamic soft sphere model. The computational results obtained for different flow conditions are validated with the recently released experimental data on upward pipe flow. The test section used is a 52 mm pipe of 5500 mm of length maintained under adiabatic conditions with air and water circulating fluids. Time-averaged results of radial distribution for void fraction, chord length, number of bubbles detected, turbulence kinetic energy, dispersed and continuous velocity profiles show a good agreement.

Multiscale modelling of mass transfer in gas jets and bubble plumes

“…However, in some large‐scale systems, the requirement to track a large number of parcels can lead to very high computational demands. LPT models have been used to simulate many types of dispersed flows, including bubbly flows in columns, tanks and unconfined systems, three‐phase flow in columns, and bubbly pipe flow …”

“…In the LPT approach, continuous phase transitions can be included by coupling the method with an IT model. The LPT method has been commonly coupled to the VOF method to simulate bubbly flows in columns and plumes in addition to the motion of the gas‐liquid free surface . Less commonly, hybrid LPT‐VOF approaches have been developed to simulate the transitions between segregated jet flows and dispersed flows.…”