A Variational Multiscale method with immersed boundary conditions for incompressible flows

Soonpil Kang, Arif Masud

Research output: Contribution to journalArticlepeer-review

Abstract

This paper presents a new stabilized form of incompressible Navier–Stokes equations for weak enforcement of Dirichlet boundary conditions at immersed boundaries. The boundary terms are derived via the Variational Multiscale (VMS) method which involves solving the fine-scale variational problem locally within a narrow band along the boundary. The fine-scale model is then variationally embedded into the coarse-scale form that yields a stabilized method which is free of user defined parameters. The derived boundary terms weakly enforce the Dirichlet boundary conditions along the immersed boundaries that may not align with the inter-element edges in the mesh. A unique feature of this rigorous derivation is that the structure of the stabilization tensor which emerges is naturally endowed with the mathematical attributes of area-averaging and stress-averaging. The method is implemented using 4-node quadrilateral and 8-node hexahedral elements. A set of 2D and 3D benchmark problems is presented that investigate the mathematical attributes of the method. These test cases show that the proposed method is mathematically robust as well as computationally stable and accurate for modeling boundary layers around immersed objects in the fluid domain.

Original languageEnglish (US)
Pages (from-to)1397-1422
Number of pages26
JournalMeccanica
Volume56
Issue number6
DOIs
StatePublished - Jun 2021
Externally publishedYes

Keywords

  • Immersed boundary
  • Incompressible fluids
  • Interfacial stabilization
  • Variational Multiscale method
  • Weakly imposed essential boundary conditions

ASJC Scopus subject areas

  • Condensed Matter Physics
  • Mechanics of Materials
  • Mechanical Engineering

Fingerprint

Dive into the research topics of 'A Variational Multiscale method with immersed boundary conditions for incompressible flows'. Together they form a unique fingerprint.

Cite this