Analysis and reduction of spurious noise generated at grid refinement interfaces with the lattice Boltzmann method

The present study focuses on the unphysical effects induced by the use of non-uniform grids in the lattice Boltzmann method. In particular, the convection of vortical structures across a grid refinement interface is likely to generate spurious noise that may impact the whole computation domain. This issue becomes critical in the case of aeroacoustic simulations, where accurate pressure estimations are of paramount importance. The purpose of this article is to identify the issues occurring at the interface and to propose possible solutions yielding significant improvements for aeroacoustic simulations. More specifically, this study highlights the critical involvement of non-physical modes in the generation of spurious vorticity and acoustics. The identification of these modes is made possible thanks to linear stability analyses performed in the fluid core, and non-hydrodynamic sensors specifically developed to systematically emphasize them during a simulation. Investigations seeking pure acoustic waves and sheared flows allow for isolating the contribution of each mode. An important result is that spurious wave generation is intrinsically due to the change in the grid resolution (i.e. aliasing) independently of the details of the grid transition algorithm. Finally, the solution proposed to minimize spurious wave amplitude consists of choosing an appropriate collision model in the fluid core so as to cancel the non-hydrodynamic mode contribution regardless the grid coupling algorithm. Results are validated on a convected vortex and on a turbulent flow around a cylinder where a huge reduction of both spurious noise and vorticity are obtained.

Thomas Astoul, Gauthier Wissocq, Jean-François Boussuge, Alois Sengissen, Pierre Sagaut. Analysis and reduction of spurious noise generated at grid refinement interfaces with the lattice Boltzmann method. Journal of Computational Physics, Elsevier, 2020, 418, pp.109645. ⟨10.1016/j.jcp.2020.109645⟩. ⟨hal-02960150⟩

Journal: Journal of Computational Physics

Date de publication: 01-10-2020

Auteurs:
  • Thomas Astoul
  • Gauthier Wissocq
  • Jean-François Boussuge
  • Alois Sengissen
  • Pierre Sagaut

Digital object identifier (doi): http://dx.doi.org/10.1016/j.jcp.2020.109645

x >