Sub-Hinze scale breakup models for high-fidelity simulation of bubbly flows
Abstract
Executive Summary The proposed project aims to develop computational capabilities for prediction of microbubbles in large-scale simulation of breaking waves in regimes relevant to naval surface ships. To make such calculations possible, our plan utilizes key recent advancements in high-performance computing of two-phase flows, as well as advancements in physical understanding of microbubble generation mechanisms at small scales. High-fidelity CFD methods, such as Large-Eddy Simulation (LES), are becoming widely available in most engineering research settings and are gradually being adopted for applied problems. Recent advancements in development of geometric volume of fluid (VoF) method, has enabled robust, fully conservative, and non-dissipative discretizations of two-phase flow equations. These properties meet the strict numerical requirements for LES, and thus provide a timely opportunity for demonstration of efficient and accurate LES of turbulent two-phase flows. However, direct simulations that resolve key processes leading to formation of microbubbles in turbulent bubbly flows are still prohibitively expensive by orders of magnitude. With reasonably affordable LES, only large bubbles and broad range of turbulence scales can be resolved. Our recent investigations suggest that a dominant mechanism contributing to the formation of microbubbles is the breakup of thin air films that are trapped during the liquid-liquid impact events. This phenomenon, called the Mesler entrainment phenomenon, can generate hundreds of microbubbles per single impact event. We have recently performed detailed simulations of this phenomenon by analyzing single impact events, and have developed a qualitative understanding of various effects contributing to the breakup process and size of bubbles generated in Mesler impacts. Our proposal is based on realization of the fact that the onset of impact events leading to the Mesler mechanism can be detected with reasonable LES resolutions (but the subsequent microbubbles are difficult to resolve). Our preliminary studies suggest that nonlinear breaking waves generate interface curvatures and velocity fluctuations hospitable for Mesler entrainment. A proper characterization of the Mesler mechanism from analysis of individual impacts, enables quantitative prediction of generated microbubbles in LES settings. Our plan is to develop such a characterization and by proper tabulation of the corresponding data we allow real-time injection of subgrid bubbles in LES simulation. These bubbles can subsequently be tracked by employing Lagrangian methods. Our plan also includes development of fast impact detection algorithms suitable for parallel LES simulations. To validate the developed subgrid scale models, LES results will be compared with the existing and ongoing experimental measurements.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 12, 2016
- Source ID
- N000141512726
Entities
People
- Parviz Moin
Organizations
- Office of Naval Research
- Stanford University
- United States Navy