A new hybrid recursive regularised Bhatnagar–Gross–Krook collision model for Lattice Boltzmann method-based large eddy simulation

ABSTRACT

A new Lattice Boltzmann collision model for large eddy simulation (LES) of weakly compressible flows is proposed. This model, referred to as the Hybrid Recursive Regularised Bhatnagar–Gross–Krook (HRR-BGK) model, is based on a modification of previously existing regularised collision models defined with the BGK Lattice Boltzmann method (LBM) framework. By hybridising the computation of the velocity gradient with an adequate Finite Difference scheme when reconstructing the non-equilibrium parts of the distribution function, a hyperviscosity term is introduced in the momentum equation, whose amplitude can be explicitly tuned via a weighting parameter. A dynamic version of the HRR-BGK is also proposed, in which the control parameter is tuned at each grid point and each time step in order to recover an arbitrarily fixed total dissipation. This new collision model is assessed for both explicit and implicit LES considering the flow around a circular cylinder at $\mathrm{Re}=3900$. The dynamic HRR-BGK is observed to yield very accurate results when equipped with Vreman's subgrid model to compute the target dissipation.

KEYWORDS:

• Large eddy simulation
• Lattice Boltzmann method
• circular cylinder Reynolds 3900

Acknowledgments

This work was carried out using the ProLB solver. Professor Eric Lamballais is warmly acknowledged for providing experimental data.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Jérôme Jacob  http://orcid.org/0000-0001-9287-4167

Orestis Malaspinas http://orcid.org/0000-0001-9427-6849

Pierre Sagaut http://orcid.org/0000-0002-3785-120X

Additional information

Funding

This work was supported by the French project CLIMB, with the financial support of BPIFrance (Project No. P3543-24000), in the framework of the program ‘Investissement d'Avenir: Calcul Intensif et Simulation Numérique’. This work was performed using HPC resources from GENCI-TGCC/CINES (Grant 2017-A0012A07679).