A conservative diffuse interface method for two-phase flows with provable boundedness properties

TL;DR

This paper introduces a conservative, bounded diffuse interface method for two-phase flows using central differences, ensuring stability and accuracy without reinitialization, and demonstrating practical advantages over existing methods.

Contribution

It develops a novel central difference scheme that guarantees phase field boundedness in two-phase flow simulations without reinitialization, combining low computational cost with theoretical stability.

Findings

Proves boundedness of phase field for specific parameters.

Achieves competitive accuracy with reduced computational cost.

Demonstrates ease of implementation and parallelizability.

Abstract

Central finite difference schemes have long been avoided in the context of two-phase flows for the advection of the phase indicator function due to numerical overshoots and undershoots associated with their dispersion errors. We will show however, for an incompressible flow, in the context of a specific diffuse interface model, one can maintain the boundedness of the phase field while also taking advantage of the low cost and ease of implementation of central differences to construct a non-dissipative discretization scheme for the advective terms. This is made possible by combining the advection and reinitialization steps of a conservative level set scheme introduced by Olsson and Kreiss [J. Comput. Phys., 210, 225 (2005)] to form a phase field equation similar to that of Chiu and Lin [J. Comput. Phys., 230, 185 (2011)]. Instead of resorting to specialized upwind methods as in these articles, we prove that the boundedness of the phase field is guaranteed for certain choices of the free parameters ($\epsilon$ and $\gamma$) for a specific central difference scheme that we propose. The proposed discretely conservative and bounded phase field equation, which is free of any reinitialization or mass redistribution, possesses desirable properties that can be leveraged in the coupled finite difference discretization of the two-phase momentum equation. Additionally, as compared to the state-of-the-art conservative and bounded two-phase flow methods, the proposed method boasts competitive accuracy-vs-cost trade-off, small memory requirements, ease of implementation and excellent parallelizability, providing a viable alternative for realistic two-phase flow calculations.