A discrete unified gas-kinetic scheme for immiscible two-phase flows

TL;DR

This paper extends the discrete unified gas-kinetic scheme (DUGKS) to simulate immiscible two-phase flows, offering improved stability, accuracy, and flexibility over previous models, validated through various flow simulations.

Contribution

The paper introduces a DUGKS-based model for two-phase flows that is more stable, adaptable to non-uniform meshes, and capable of handling wider viscosity and density ratios.

Findings

Accurately tracks interfaces in two-phase flows.

Demonstrates stability over a range of flow conditions.

Handles wider viscosity and density ratios than previous models.

Abstract

In this work, we extend the discrete unified gas-kinetic scheme (DUGKS) [Guo et al., Phys. Rev. E 88, 033305 (2013)] to continue two-phase flows. In the framework of DUGKS, two kinetic model equations are used to solve the quasi-incompressible phase-field governing equations [Yang et al., Phys. Rev.E 93, 043303 (2016)]. One is for the Chan-Hilliard (CH) equation and the other is for the Navier-Stokes equations. The DUGKS can correctly recover the quasi-incompressible phase-field governing equations through the Chapman-Enskog analysis. Unlike previous phase-field-based LB models, the Courant-Friedricks-Lewy condition in DUGKS is ajustable which can increase numerical stability. Furthermore, with the finite-volume formulation the model can be easily implemented on non-uniform meshes which can improve numerical precision. The proposed model is validated by simulating a stationary drop, layered Poiseuille flow, rising bubble and Rayleigh-Taylor instability and comparing with the quasi-incompressible lattice Boltzmann method (LBM). Numerical results show that the method can track the interface with high accuracy and stability. The model is also capable of dealing with a wider range of viscosity and density ratios than the quasi-incompressible lattice Boltzmann model. The present model is a promising tool for numerical simulation of two-phase flows.