Fully coupled implicit finite-volume algorithm for viscoelastic interfacial flows

This paper proposes a robust, fully coupled implicit finite-volume algorithm combined with a front-tracking method to accurately simulate incompressible viscoelastic interfacial flows at high Weissenberg numbers without the need for a log-conformation approach.

The Gist

Imagine you are trying to simulate how a drop of thick, stretchy honey moves through a pool of water, or how a bubble rises through a gooey, elastic liquid like mucus.

In the world of computer science, this is a nightmare. It’s like trying to predict the movement of a crowd of people where everyone is connected by invisible, stretchy bungee cords, and some people are walking through water while others are walking through thick syrup.

This paper introduces a new "mathematical engine" (an algorithm) designed to solve these messy, stretchy, "interfacial" problems more accurately and reliably than ever before.

Here is the breakdown of how it works using everyday analogies.

1. The Problem: The "Stretchy Bungee Cord" Effect

Most computer simulations for fluids treat them like water: once you push it, it moves, and then it’s done. But viscoelastic fluids (like polymer solutions, blood, or even some foods) are different. They have "memory." When you deform them, they act like they are full of tiny, microscopic bungee cords. If you stretch them, they want to snap back.

In older computer models, scientists used a "segregated" approach. Imagine trying to solve a massive jigsaw puzzle by looking at only one piece at a time, then trying to fit it into the next. You might get the pieces to fit, but the whole picture often becomes unstable or "breaks" if the bungee cords get too tight (this is known as the "High Weissenberg Number Problem").

2. The Solution: The "All-at-Once" Approach

The authors proposed a "Fully Coupled Implicit" algorithm.

Instead of solving the puzzle piece by piece (the old way), they decided to solve the entire puzzle simultaneously.

The Analogy:
Imagine you are a conductor of an orchestra.

• The Old Way (Segregated): You tell the violinists to play, then you tell the drummers to play, then the flutists. You try to coordinate them, but it’s hard to keep the rhythm perfect because they aren't hearing each other in real-time.
• The New Way (Fully Coupled): You create a magical environment where every single musician hears every other musician instantly. When the drummer hits a beat, the violinist adjusts at the exact same microsecond. Because everyone is perfectly "coupled," the music (the simulation) is much smoother, more stable, and can handle much more intense, complex rhythms.

3. The "Front-Tracking" Secret Sauce

When you have two different liquids touching (like oil and water), there is a "border" or an interface. This paper uses a method called "Front-Tracking."

Think of this like a high-tech GPS tracker attached to the border of the two liquids. Instead of just guessing where the border is, the computer follows the "front" of the liquid very closely, like a specialized drone following the shoreline of a moving wave. This allows the simulation to handle very dramatic changes, like a droplet being stretched into a long thin thread without the computer "losing" the shape.

4. What did they prove?

To make sure their new "orchestra" actually worked, they put it through four "stress tests":

1. The Spinning Cup: They simulated a fluid spinning in a box. It worked perfectly, matching previous known results.
2. The Energy Test: They checked if their "bungee cords" were accidentally stealing energy from the system (a common math error). They proved their method was incredibly efficient and didn't "leak" energy.
3. The Stretching Droplet: They simulated a droplet being squeezed in a flow. Even when the "stretchy" forces were extremely high, the simulation didn't crash.
4. The Rising Bubble: This was the big one. They simulated a bubble rising through a thick, elastic liquid. They successfully captured a weird phenomenon called a "negative wake"—where the liquid actually flows backward behind the bubble. This is a very difficult thing to simulate, and they nailed it.

The Bottom Line

This paper provides a much more robust "brain" for computers to use when simulating complex, stretchy liquids. It allows engineers and scientists to predict how things like 3D-printed materials, biological fluids (like mucus), and industrial chemicals will behave, even when the forces involved are incredibly intense and chaotic.

Technical Summary

Technical Summary: Fully Coupled Implicit Finite-Volume Algorithm for Viscoelastic Interfacial Flows

1. Problem Statement

The numerical simulation of viscoelastic interfacial flows—where two or more immiscible fluids with viscoelastic properties interact—presents significant computational challenges. Traditional numerical approaches often rely on segregated algorithms, which solve the continuity, momentum, and constitutive equations sequentially. These methods suffer from weak coupling between velocity, pressure, and polymer stress, leading to stability issues, slow convergence, and the requirement for heavy under-relaxation.

Furthermore, simulating flows with high elasticity (high Weissenberg numbers, $Wi$) often leads to numerical instabilities, traditionally necessitating complex mathematical workarounds like the log-conformation approach. There is a critical need for a robust, efficient, and tightly coupled framework capable of handling strong interfacial forces (surface tension) and high elasticity without these additional complexities.

2. Methodology

The authors propose a fully coupled implicit finite-volume algorithm implemented in their in-house solver, MultiFlow. The core components of the methodology include:

• Governing Equations: The framework solves the incompressible Navier-Stokes equations coupled with a general differential constitutive model (specifically the Giesekus, linear Phan-Thien-Tanner (LPTT), and exponential Phan-Thien-Tanner (EPTT) models). These models account for shear-thinning and limited extensibility.
• Fully Coupled Implicit Discretization: Unlike segregated solvers, this method discretizes the continuity, momentum, and constitutive equations such that pressure, velocity, and all six unique components of the polymer stress tensor are solved simultaneously in a single linear system of equations ($A \cdot \zeta = b$).
• Stress-Velocity Coupling: To ensure stability in a collocated variable arrangement, the authors employ a "both-sides diffusion" (BSD) technique. This introduces two mathematically equivalent diffusion terms with opposite signs on different computational stencils, which provides necessary numerical stabilization without sacrificing second-order accuracy.
• Pressure-Velocity Coupling: A momentum-weighted interpolation (MWI) is used to prevent pressure-velocity decoupling, treating all pressure terms implicitly.
• Interface Tracking: The interface between phases is modeled using a state-of-the-art front-tracking method. This involves a Lagrangian surface mesh that is advected using a "normal-only advection" (NOA) scheme to maintain mesh quality, with surface tension treated as a volumetric source term.

3. Key Contributions

• Unified Implicit Framework: The development of the first fully coupled implicit finite-volume algorithm specifically designed for viscoelastic interfacial flows.
• Elimination of Log-Conformation: A major contribution is the demonstration that the tightly coupled implicit approach can predict high-$Wi$ flows accurately without requiring a log-conformation transformation, simplifying the implementation and reducing computational overhead.
• Robust Interfacial Handling: The integration of the front-tracking method with a fully implicit solver allows for the stable simulation of flows with large density ratios and strong surface tension effects.

4. Results and Validation

The framework was validated through four representative test cases:

1. Lid-Driven Cavity (Single-Phase): Confirmed the algorithm's accuracy against existing literature for LPTT fluids, showing excellent agreement in velocity profiles and second-order mesh convergence for stress components.
2. Taylor Vortices: Demonstrated that the stress-velocity coupling maintains second-order accuracy and does not introduce significant artificial numerical diffusion.
3. Droplet in Shear Flow (Interfacial): Successfully simulated a Newtonian droplet in a Giesekus fluid. The model remained robust and physically consistent at very high Weissenberg numbers (up to $Wi = 10^4$), matching previous benchmark results.
4. Rising Bubble (Interfacial/Complex): Validated against experimental data for a bubble rising in an EPTT liquid. The algorithm successfully captured the jump discontinuity in terminal rise velocity (the transition from subcritical to supercritical regimes) and the formation of a "negative wake" (reverse flow) behind the bubble.

5. Significance

This research provides a powerful and robust numerical tool for the rheology and manufacturing communities. By enabling the reliable simulation of strongly elastic interfacial flows at high Weissenberg numbers without the need for log-conformation, the algorithm opens new doors for modeling complex processes such as 3D printing of soft materials, the dispersion of respiratory droplets (e.g., viruses in mucus), and lubrication dynamics in bearings. It represents a significant step forward in the computational modeling of non-Newtonian multiphase flows.