Cut-element based immersed boundary method for moving geometries in compressible liquid flows with cavitation

Felix Örley, Vito Pasquariello, Stefan Hickel, Nikolaus A. Adams

Research output: Contribution to journalArticlepeer-review

40 Scopus citations

Abstract

The conservative immersed interface method for representing complex immersed solid boundaries or phase interfaces on Cartesian grids is improved and extended to allow for the simulation of weakly compressible fluid flows through moving geometries. We demonstrate that an approximation of moving interfaces by a level-set field results in unphysical oscillations in the vicinity of sharp corners when dealing with weakly compressible fluids such as water. By introducing an exact reconstruction of the cut-cell properties directly based on a surface triangulation of the immersed boundary, we are able to recover the correct flow evolution free of numerical artifacts. The new method is based on cut-elements. It provides sub-cell resolution of the geometry and handles flows through narrow closing or opening gaps in a straightforward manner. We validate our method with canonical flows around oscillating cylinders. We demonstrate that the method allows for an accurate prediction of flows around moving obstacles in weakly compressible liquid flows with cavitation effects. In particular, we show that the cavitating flow through a closing fuel injector control valve, which is an example for a complex application with interaction of stationary and moving parts, can be predicted by the method.

Original languageEnglish
Pages (from-to)1-22
Number of pages22
JournalJournal of Computational Physics
Volume283
DOIs
StatePublished - 5 Feb 2015

Keywords

  • Cavitation
  • Cut-cell method
  • Fluid-structure interaction
  • Immersed boundary method
  • Large eddy simulation
  • Moving boundary

Fingerprint

Dive into the research topics of 'Cut-element based immersed boundary method for moving geometries in compressible liquid flows with cavitation'. Together they form a unique fingerprint.

Cite this