Skip to main navigation Skip to search Skip to main content

A finite particle method based on a Riemann solver for modeling incompressible flows

  • Fan Zhang*
  • , Can Huang
  • , Huashan Zhang
  • , Tiegang Liu
  • , Moubin Liu
  • *Corresponding author for this work
  • University of Science and Technology Beijing
  • Peking University
  • North China University of Technology
  • Ocean University of China and Qingdao National Laboratory for Marine Science and Technology

Research output: Contribution to journal › Article › peer-review

Abstract

The weakly compressible SPH (WCSPH) method has been widely used for solving hydrodynamic problems, in which the fluid is assumed to be weakly compressible through limiting the density deviation to a small value. Despite its simplicity and robustness in implementation, the WCSPH method is generally hindered by its poor accuracy and spurious oscillations in the pressure field which may lead to numerical instability. In order to dampen the spurious oscillations without introducing excessive dissipation, a Roe-type approximate Riemann solver with a low-dissipation limiter was proposed and applied to the WCSPH method by Zhang et al. (2017). The present paper aims to extend this low-dissipation Riemann solver to the finite particle method (FPM), which can be referred to an enhanced or modified version of the conventional SPH method, for reducing spurious pressure oscillations and meanwhile achieving higher order of accuracy. Numerical results demonstrate that the Riemann solver based FPM is effective in modeling incompressible flows and it produces a more accurate flow field than the conventional WCSPH and Riemann solver based WCSPH methods.

Original languageEnglish
Pages (from-to)74-88
Number of pages15
JournalComputers and Mathematics with Applications
Volume124
DOIs
StatePublished - 15 Oct 2022

Keywords

  • Finite particle method
  • Riemann solver
  • Smoothed particle hydrodynamics
  • Weakly compressible flows

Fingerprint

Dive into the research topics of 'A finite particle method based on a Riemann solver for modeling incompressible flows'. Together they form a unique fingerprint.

Cite this