Simulation method for bubble flows and fluid–structure interaction

RWTH Publications (RWTH Aachen)

Abstract

In this thesis, a numerical method for the simulation of bubble flows and fluid–structure interaction (FSI) is presented. It is implemented in the open-source solver framework m–AIA. The method combines a lattice Boltzmann solver with a cumulant-based collision step for fluid dynamics, an Eulerian–Lagrangian multiphase model for bubble transport, and the finite element method (FEM) library MFEM for solid mechanics. An interpolated bounce-back boundary condition enables direct coupling between the fluid and solid domains. The implementation leverages MPI-based parallelization strategies and supports both CPU-based systems with a hybrid OpenMP–MPI parallelization and GPU-accelerated high-performance computing systems. The developed method is validated against established benchmark cases. The multiphase model is validated using Deen's bubble column, which features a highly unsteady turbulent bubble flow. The simulation results show good agreement with experimental and numerical data for averaged quantities and turbulence statistics. The FSI method is validated through simulations of the flow around an oscillating cylinder, vortex-induced vibrations of an elastically mounted cylinder, and the Turek–Hron flexible flag benchmark. In all cases, accurate predictions of delicate FSI phenomena are achieved with good agreement to reference data. The method is then applied to simulations of the electrochemical machining (ECM) process, where gas bubbles generated during material removal affect electrical conductivity and thus machining accuracy and surface quality. Large eddy simulations of gas transport for the machining of a generic compressor blade geometry are performed on grids of up to 1.2 billion cells. The influence of the inlet geometry on electrolyte flow and gas distribution is investigated, and the effect of blade curvature on the local gas concentration and electrolyte conductivity is discussed. Additionally, FSI simulations of the precise electrochemical machining (PECM) process with oscillating cathodes are conducted on a GPU-accelerated system using 2.5 billion fluid cells. Here, the effect of the tool oscillation frequency on the workpiece deformation is analyzed. The results reveal that the higher oscillation frequency significantly increases blade displacement, which adversely affects machining precision and increases the risk of short-circuit events. These simulations provide valuable insights into the underlying physical phenomena of the ECM process and demonstrate the capability of the developed method to analyze complex multiphysical processes on high-performance computing systems.

Authors 1

1. Daniel Lauwers corresponding Aachen

   RWTH Aachen University

   Affiliation as printed

   RWTH Aachen