Physical Formulation and Numerical Algorithm for Simulating N Immiscible Incompressible Fluids Involving General Order Parameters

TL;DR

This paper introduces a novel physical formulation and an efficient numerical algorithm for simulating multiple immiscible incompressible fluids using general order parameters, enabling accurate and computationally feasible multi-phase flow modeling.

Contribution

It develops a new phase field formulation with general order parameters and a decoupling numerical scheme, improving simulation accuracy and efficiency for N-phase fluid systems.

Findings

Successfully simulates multiple fluid phases with large property contrasts.

Produces physically accurate results consistent with established theories.

Maintains computational complexity comparable to simpler models.

Abstract

We present a physical formulation, and a numerical algorithm, based on a class of general order parameters for simulating the motion of a mixture of $N$ ($N\geqslant 2$) immiscible incompressible fluids with given densities, dynamic viscosities, and pairwise surface tensions. The introduction of general order parameters leads to a more strongly coupled system of phase field equations, in contrast to that with certain special choice of the order parameters. However, the general form enables one to compute the N-phase mixing energy density coefficients in an explicit fashion in terms of the pairwise surface tensions. From the simulation perspective, the increased complexity in the form of the phase field equations with general order parameters in actuality does not cause essential computational difficulties. Our numerical algorithm reformulates the ($N-1$) strongly-coupled phase field equations for general order parameters into $2(N-1)$ Helmholtz-type equations that are completely de-coupled from one another, leading to a computational complexity essentially the same as that of the simpler phase field equations associated with special choice of order parameters. We demonstrate the capabilities of the method developed herein using several test problems involving multiple fluid phases and large contrasts in densities and viscosities among the multitude of fluids. In particular, by comparing simulation results with the Langmuir-de Gennes theory of floating liquid lenses we show that the method produces physically accurate results for multiple fluid phases.