Large-scale Simulations in Complex Flows
Title: Large-scale Simulations in Complex Flows
SNIC Project: Berzelius-2021-6
Project Type: LiU Berzelius
Principal Investigator: Luca Brandt <luca@mech.kth.se>
Affiliation: Kungliga Tekniska högskolan
Duration: 2021-09-01 – 2022-03-01
Classification: 20306
Keywords:

Abstract

The physical process of a rising bubble (or swarm of bubbles) in ambient turbulence is a fundamental problem of fluid mechanics. The flow agitation induced by the raising bubble motion, i.e. Bubble Induced Turbulence (BIT), plays a fundamental role in industrial processes such as heat-exchanger, chemical reactors and cooling towers. In nature, the presence of air bubbles in the ocean controls the air-sea gas transfer, important for both CO2 absorption and oxygen concentration in the sea. In most of the mentioned applications, BIT usually combines with pre-existing turbulence, which in its simplest model can be assumed as Homogeneous and Isotropic Turbulence (HIT). The interaction of BIT and HIT originates a complex scenario which is difficult to study with experimental techniques and that is better suited for numerical studies. In particular, the rising motion of the bubble significantly alters the canonical turbulence behavior, exemplified by the -5/3 power-law scaling of the turbulence spectra, and creates different scaling laws, such as the -3 power-law at smaller scales. To these days, it is unclear how BIT and HIT interactions affect the general flow behavior. Most notably, the energy transfer mechanisms remain to be fully unveiled and the main unknown is given by how bubble’s size alters the onset of -3 and -5/3 scaling in the energy spectra. This has significant implications in all applications, and its understanding requires access to flow features with details that only Direct Numerical Simulation (DNS) can provide. In this project, we aim at understanding the effect of the bubble size on the turbulent kinetic energy spectra, by performing DNS of rising bubbles (i.e. BIT) in ambient HIT. In order to fully address this phenomena, large scale separation (between energy generation scales and energy dissipation scales) is needed, which can only be addressed with large simulations. We will therefore study this phenomena on a cube with 10243 grid points, with sustained HIT. In particular, we wish to perform 3 simulations, with bubbles of different sizes, by adding them on top of the developed turbulent velocity field. By maintaining the other simulation parameters constant for all simulations, we will be able to study the differences between each case and assess what determines the modification of the turbulent kinetic energy spectra.