Numerical frameworks for small-scale bubble dynamics

Licentiate thesis, 2020

Designing efficient bubbly flow systems requires the prediction of the dynamics of the bubbles, the liquid and how the gas and liquid phases interact. Currently, the complex dynamics in bubbly flows are not yet fully explained, and we rely on incomplete models in our numerical methods. A major concern when investigating bubbly flows using numerical methods is the large range of length and time scales. The length scales may vary from nanometers, for the formation of vapor bubbles, to tens of meters, when system-size bubbly flow structures are induced. To predict the dynamics of the entire system, it is important to understand the phenomena at other scales, such as the evolution of vapor bubbles or the dynamics of individual bubbles.

In this PhD project, we aim at increasing our knowledge about the bubbly flow dynamics and developing numerical methods for investigation bubbly flows across all relevant length scales. So far into the project, we have focused on the small scale bubble dynamics where small scales refer to length scales of individual bubbles and below. We start by studying the evolution process of vapor bubbles by developing a multiphase DNS framework, and a less computationally expensive 1D framework, that resolve the conditions in both phases and takes into account phase change and thermal effects. These frameworks can be used to study both boiling and cavitation processes, and we use it to investigate the challenging case of laser-induced thermocavitation bubbles. These bubbles are studied as a promising tool to achieve good control in the process of crystallization. We simulate such bubbles and identify plausible mechanisms behind experimentally observed crystallization events and provide guidelines for appropriate setups to attain conditions favorable for crystallization.

Then, we shift the focus to investigate rising bubble dynamics at small scales. For this purpose, we develop an efficient multiphase DNS framework with a moving reference frame (MRF) technique that follows the bubbles. This method significantly reduces the size of the computational domain and eliminate the need for a priori estimations of sufficient domain sizes to capture the bubble dynamics. With the MRF method, we aim at obtaining the closures for bubble dynamics at small scales and use them to investigate bubbly flows up to industrial scales in the continuation of the project.