An exactly curl-free staggered semi-implicit finite volume scheme for a first order hyperbolic model of viscous flow with surface tension

TL;DR

This paper introduces a semi-implicit finite volume scheme for a hyperbolic two-phase flow model with surface tension and viscosity, ensuring curl-free properties and efficient handling of stiff source terms in low-Mach regimes.

Contribution

The paper develops a novel curl-free staggered semi-implicit finite volume method that preserves differential identities and efficiently solves stiff nonlinear source terms in two-phase flow models.

Findings

Exact preservation of curl constraints via compatible operators

Efficient semi-analytical integration of stiff source terms

Improved accuracy and efficiency in low-Mach number regimes

Abstract

In this paper, we present a semi-implicit numerical solver for a first order hyperbolic formulation of two-phase flow with surface tension and viscosity. The numerical method addresses several complexities presented by the PDE system in consideration: (i) The presence of involution constraints of curl type in the governing equations requires explicit enforcement of the zero-curl property of certain vector fields (an interface field and a distortion field); the problem is eliminated entirely by adopting a set of compatible curl and gradient discrete differential operators on a staggered grid, allowing to preserve the Schwartz identity of cross-derivatives exactly. (ii) The evolution equations feature highly nonlinear stiff algebraic source terms which are used for the description of viscous interactions as emergent behaviour of an elasto-plastic solid in the stiff strain relaxation limit; such source terms are reliably integrated with an efficient semi-analytical technique. (iii) In the low-Mach number regime, standard explicit Godunov-type schemes lose efficiency and accuracy; the issue is addressed by means of a simple semi-implicit, pressure-based, split treatment of acoustic and non-acoustic waves, again using staggered grids that recover the implicit solution for a single scalar field (the pressure) through a sequence of symmetric-positive definite linear systems that can be efficiently solved via the conjugate gradient method.