Error estimates of linear decoupled structure-preserving incremental viscosity splitting methods for the Cahn–Hilliard–Navier–Stokes system Journal: Arxiv

Baolin Kuang, Hongfei Fu ∗, Xiaoli Li · 2025

Read Error estimates of linear decoupled structure-preserving incremental viscosity splitting methods for the Cahn–Hilliard–Navier–Stokes system Journal: Arxiv by Baolin Kuang, Hongfei Fu ∗, Xiaoli Li (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-01141.txt.

Introduction

Baolin Kuang, Hongfei Fu, and Xiaoli Li design and analyze schemes for a coupled model of phase separation and fluid flow. The Cahn-Hilliard-Navier-Stokes system describes two fluids mixing or unmixing while they move, a model used for alloys, polymers, and oil-water flows. Simulating it is hard because the physics couples tightly and energy must be tracked exactly. Their schemes split the work so each time step solves only constant-coefficient equations plus one linear equation.

The designs come in first- and second-order time accuracy. The authors prove the schemes have a unique answer, conserve mass, and never create energy. The hard part is the error analysis: using induction and regularity results, they prove the first-order scheme converges at the best possible rate in several dimensions. Numerical experiments back the theory. The paper serves simulation specialists who need speed and guarantees.

Catalog record

CREATORBaolin Kuang, Hongfei Fu ∗, Xiaoli Li
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE102 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-01141

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Error estimates of linear decoupled structure-preserving incremental viscosity splitting methods for the Cahn–Hilliard–Navier–Stokes system Journal: Arxiv Baolin Kuang, Hongfei Fu ∗, Xiaoli Li Abstract We propose first- and second-order time discretization schemes for the coupled Cahn–Hilliard–Navier–Stokes model, leveraging the incremental viscosity splitting (IVS) method. The schemes combine the scalar auxiliary variable method and the zero-energy-contribution approach, resulting in a linear, decoupled numerical framework. At each time step, they only require to solve a sequence of constant-coefficient equations, along with a linear equation with one unknown, making the algorithms computationally efficient and easy to implement. In addition, the proposed schemes are proven to be uniquely solvable, mass-conserving, and unconditional energy dissipation. Most importantly, leveraging the mathematical induction method and the regularity properties of the Stokes equation, we perform a rigorous error analysis for the first-order scheme in multiple space dimensions, establishing an unconditional and optimal convergence rate for all relevant variables under different norms. A user-defined, time-dependent parameter plays an important role in the error analysis of the proposed structure-preserving IVS methods. Ample numerical examples are carried out to verify the theoretical findings and to demonstrate the accuracy, effectiveness and efficiency of the proposed schemes. ∎ 1 Introduction The phase field method, also known as the diffusive interface method, has become a prominent mathematical framework for modeling interfacial dynamics in multi-component systems anderson1998diffuse ; christlieb2014high . The governing equations, namely the phase field model, are typically derived through a variational approach within an appropriate functional setting aimed at minimizing the total free energy. Notably, the Cahn–Hilliard–Navier–Stokes (CHNS) system anderson1998diffuse ; gurtin1996two ; kay2007efficient represents the fundamental building blocks of hydrodynamic phase-field models for multiphase fluid flow dynamics. Despite being extensively studied over the past two decades chen2020novel ; zhao2021second ; liu2024simple ; qian2023fully ; han2015second ; kay2007efficient , numerical research on the CHNS system remains challenging due to its inherent complexities, including: (i) the presence of the stiffness terms linked to interfacial width parameters in the phase equation, the nonlinear convection term in the momentum equation, and the nonlinear couplings between velocity and phase-field variables through convective transport and stress effects; (ii) the intrinsic velocity-pressure coupling imposed by the incompressibility constraint; (iii) the requirement for numerical methods to preserve essential physical laws, such as energy dissipation law and mass conservation law; and (iv) the theoretical difficulty in deriving rigorous optimal-order error estimates for fully decoupled linear schemes. Due to the inherent complexity of the CHNS system, considerable efforts have been devoted to establishing error estimates for efficient numerical schemes. For example, Diegel et al. diegel2017convergence proposed a second-order convex splitting scheme using mixed finite elements, achieving optimal convergence rates for the velocity in the \bm{L}^{2} -norm and for the phase-field variable and the chemical potential in the H^{1} -norm. However, no theoretical estimate was provided for the pressure. Cai and Shen cai2018error conducted a rigorous error analysis for a linear, weakly coupled energy-stable scheme with mixed element discretization. Apart from error estimates for the phase-field variable, the chemical potential and the velocity, they also obtained the optimal-order error estimate for the pressure in the \ell^{2}(0,T;L^{2}) -norm. Xu et al. xu2020error proposed a first-order in time semi-discrete scheme based on projection methods goda1979multistep and stabilization technique shen2010numerical , which is fully decoupled and energy-stable. They proved the optimal-order accuracy for the phase-field variable in the \ell^{\infty}(0,T;H^{1}) -norm, for the chemical potential in the \ell^{2}(0,T;H^{1}) -norm, and for the velocity in the \ell^{\infty}(0,T;\bm{L}^{2})\cap\ell^{2}(0,T;\bm{H}^{1}) -norm, and for the pressure also in the \ell^{2}(0,T;L^{2}) -norm. Recently, Li et al. li2022fully proposed first- and second-order fully decoupled time discretization schemes based on the multiple scalar auxiliary variable (MSAV) approach for gradient systems, combined with (rotational) pressure-correction methods goda1979multistep ; guermond2006overview for the Navier–Stokes equations. They carried out a rigorous error analysis for the first-order scheme, and optimal convergence rates for the phase-field variable and chemical potential in the H^{1} -norm, and the velocity in the \bm{L}^{2} -norm were presented. Notably, they also provided a half-order error estimate for the pressure in the stronger \ell^{\infty}(0,T;L^{2}) -norm. However, extension of the error analysis to three-dimensional model remains unresolved. Recent efforts by Cai et al. cai2023optimal derived optimal L^{2} -norm error estimates for convex-splitting finite element methods using novel quasi-projection techniques, however, again no error estimate for the pressure was presented. Despite these great advancements, the construction of fully decoupled linear schemes with constant coefficients and rigorous error analysis in a stronger norm (especially for the pressure) remains limited. Moreover, most existing works rely on projection methods or mixed finite element discretization to handle the velocity-pressure coupling, which may cause artificial boundary conditions for the velocity or require special finite element space pairs for the velocity and pressure. Apart from conventional projection methods, the viscosity-splitting (VS) method blasco1998fractional provides an efficient alternative for solving the Navier–Stokes equations. This approach consists of two main steps: the first step involves solving a linear elliptic problem without incompressibility constraint and pressure term, and the second step preserves the viscosity and addresses a generalized Stokes problem. In contrast to the projection methods, a key advantage is that it preserves full original velocity boundary conditions throughout the time-stepping process, thereby avoiding artificial boundary conditions. Furthermore, El-Amrani et al. extended its application to incompressible non-Newtonian fluids el2024error and to thermally coupled Navier–Stokes equations involving convection–conduction dynamics el2024time , where the velocity was shown to achieve first-order accuracy in the \ell^{\infty}(0,T;\bm{L}^{2})\cap\ell^{2}(0,T;\bm{H}^{1}) -norm, while only 1/2 -order accuracy in the \ell^{2}(0,T;L^{2}) -norm was proved for the pressure. We point out that el2024error also asserted that the pressure approximation can attain full first-order accuracy by employing the same arguments as in guillen2011new ; yakoubi2023enhancing . Inspired by the incremental pressure-projection scheme goda1979multistep , Yakoubi yakoubi2023enhancing enhanced the VS method by proposing an incremental viscosity-splitting (IVS) scheme for the Navier–Stokes equations, which further improves the accuracy and robustness of the VS method. They demonstrated that the scheme achieves first-order convergence for the velocity in the \ell^{\infty}(0,T;\bm{L}^{2})\cap\ell^{2}(0,T;\bm{H}^{1}) -norm, and for the pressure also in the \ell^{2}(0,T;L^{2}) -norm. More recently, Obbadi et al. obbadi2025stable proposed a second-order IVS method for the incompressible Navier–Stokes equations and provided the corresponding stability analysis. However, the existing VS and IVS methods typically employ a semi-implicit treatment for the nonlinear convection terms, which results in variable-coefficient equations that must be updated at each time step. Recently, by introducing a time-dependent auxiliary variable, Sun et al. Yang2024error proposed an efficient first-order auxiliary variable viscosity-splitting method for the incompressible fluid flows, which yields constant-coefficient Poisson-type equations at each time step. They also established an energy dissipation law based on kinetic energy rather than a modified energy involving velocity and additional pressure gradient, while, no error analysis has been conducted for this scheme. To the best of our knowledge, research on the IVS method remains limited. In particular, no existing work has considered its application to complex two-phase flow models such as the CHNS system, nor has any rigorous error analysis been carried out for such coupled systems. To address this significant gap, we propose a novel linear, fully decoupled, and structure-preserving numerical framework for the CHNS system by integrating the incremental viscosity-splitting (IVS) method yakoubi2023enhancing with the SAV-ZEC approach zgdsisc2024 . The SAV-ZEC approach that integrates the SAV

Read the complete text