🔢 mathematics

A volume penalization method for solving conjugate scalar transport with interfacial jump conditions

This paper presents a novel volume penalization method that effectively handles interfacial jump conditions in conjugate scalar transport problems on complex geometries, achieving high accuracy with less than 3% deviation compared to body-fitted mesh simulations.

Original authors: Ming Liu, Yosuke Hasegawa

Published 2026-08-11

📖 3 min read🧠 Deep dive

CC BY 4.0

Original authors: Ming Liu, Yosuke Hasegawa

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a world where heat and chemicals don't just flow smoothly through a pipe or a room, but have to jump over invisible hurdles. In the real world, when a hot gas touches a cold solid wall, or when a chemical reaction happens right at the boundary between a liquid and a gas, things get messy. The temperature might suddenly spike, or the concentration of a chemical might change instantly. Scientists call this "conjugate scalar transport with interfacial jump conditions." It's a fancy way of saying: "How do we track things like heat or pollution when they behave differently on either side of a wall, and sometimes even jump across it?"

To solve this, researchers usually use computer simulations. The old-school way is to build a digital map (a mesh) that fits perfectly around every curve of the object, like molding clay around a statue. It's accurate, but it's a nightmare to make if the object has a weird shape, like a crumpled piece of paper or a rough rock. A newer, cooler method is called the "Volume Penalization Method" (VPM). Think of VPM as a "ghost wall." Instead of molding the map to the object, you just paint the object onto a simple grid. The computer treats the solid object as a super-thick, sticky sponge that stops things from moving through it. It's fast and easy, even for crazy shapes. But, until now, this "ghost wall" trick had a glitch: it couldn't handle those sudden jumps in heat or chemicals at the boundary without creating fake, nonsensical results inside the solid object.

This paper introduces a clever fix for that glitch. The authors, Ming Liu and Yosuke Hasegawa, developed a new way to use the Volume Penalization Method that allows for these "jump conditions" without breaking the simulation. Instead of just forcing the wall to stop things, they added a special "source term"—think of it as a tiny, invisible pump located exactly at the boundary—that pushes the right amount of heat or chemical across the line, even if the amount changes suddenly. They tested this new method on several scenarios, including a simple one-dimensional diffusion problem and more complex fluid-solid interactions. The results showed that their new "ghost wall" is highly accurate, matching the results of the old, difficult "molding clay" method with an average error of less than 3.0%. Crucially, unlike the old version of this method, their new approach doesn't create fake data inside the solid object, making it a reliable tool for simulating complex thermal and chemical processes in everything from micro-channels to industrial reactors.

Technical Summary: A Volume Penalization Method for Solving Conjugate Scalar Transport with Interfacial Jump Conditions

Problem Statement
Conjugate scalar transport (e.g., heat or mass transfer) involving interfacial jump conditions is prevalent in thermal and chemical processes, such as microchannel heat transfer, chemical reactions at interfaces, and phase change systems. In these scenarios, discontinuities often exist in the scalar value itself (e.g., temperature or concentration jumps) and/or its flux across the fluid-solid interface. While body-fitted mesh (BFM) simulations can handle these conditions by explicitly coupling separate phase domains, they require cumbersome grid generation for complex geometries. Immersed Boundary Methods (IBMs), particularly the Volume Penalization Method (VPM), offer a way to embed arbitrary geometries in Cartesian grids without complex meshing. However, existing VPM schemes struggle with Neumann boundary conditions and general interfacial jump conditions. Conventional approaches often generate non-physical scalar fields within the solid region and fail to accurately enforce flux jumps or scalar discontinuities without additional, complex coupling procedures.

Methodology
The authors propose a novel treatment within the VPM framework to solve conjugate scalar transport with general interfacial jump conditions. The methodology proceeds in three main stages:

Level-Set Representation: An arbitrary fluid-solid interface is represented by a level-set function ( $\phi_0$ ), converted into a phase indicator ( $\phi$ ) that transitions smoothly from 0 (fluid) to 1 (solid) over a thin interfacial region of thickness $2\delta$ .
Neumann Boundary Condition Treatment: For a single-phase advection-diffusion problem with a Neumann condition, the authors derive a source term in divergence form, $\nabla \cdot (q_w \phi \mathbf{n})$ , to be added to the governing equation. Unlike previous schemes that force derivatives via a penalty term throughout the solid volume, this new term is non-zero only within the interfacial region. This ensures that the prescribed scalar flux $q_w$ is applied precisely while keeping the molecular diffusivity zero inside the solid, thereby preventing non-physical diffusion into the solid phase.
Generalization to Conjugate Transport with Jumps: To handle both scalar jumps ( $c_f = \alpha c_s$ ) and flux jumps, the authors introduce an "equivalent scalar" $h = \alpha c$ . This transformation removes the scalar discontinuity, allowing the use of a unified governing equation for both phases. The flux jump condition is reformulated into an equivalent flux jump $q^*_w$ , which incorporates both the prescribed flux jump and the effect of the scalar jump. The unified governing equation for the equivalent scalar $h$ is:
$\frac{\partial h}{\partial t} + \mathbf{u} \cdot \nabla h = \nabla \cdot (D \nabla h) + \nabla \cdot (q^*_w \phi \mathbf{n})$
where $D$ is the phase-dependent diffusivity.

The scheme is implemented in the open-source finite volume code OpenFOAM (version 8) using second-order linear discretization for spatial terms and the Euler implicit scheme for transient terms.

Key Contributions

Novel Source Term Formulation: The paper introduces a divergence-form source term for VPM that precisely enforces interfacial flux jumps without generating non-physical solutions in the solid region.
Unified Governing Equation: By defining an equivalent scalar and flux, the method unifies the governing equations for fluid and solid phases, eliminating the need for explicit interface coupling or separate domain solutions.
Extension to General Jump Conditions: The method successfully extends VPM to handle simultaneous jumps in both the scalar value and its flux, a capability previously lacking in finite-volume-based immersed boundary schemes.
Error Analysis: The authors provide an analytical error analysis demonstrating that the scheme achieves first-order accuracy with respect to the grid size, consistent with conventional VPMs, even in the presence of interfacial curvature.

Results
The proposed method was validated through three distinct test cases:

One-Dimensional Diffusion: A transient diffusion problem with a Neumann boundary condition was solved. The results showed good agreement with analytical solutions, with an average relative error of approximately 1.25%. Crucially, unlike existing VPM schemes, the present method maintained a zero scalar field inside the solid region, avoiding non-physical diffusion. The error remained stable across varying diffusivity values, whereas existing schemes showed increasing errors with higher diffusivity.
Fluid-Solid Coupled Diffusion: Simulations of a cylinder in a square domain with both scalar and flux jumps were performed. The VPM results were compared against body-fitted mesh (BFM) simulations using chtMultiRegionFoam. The relative deviations between the VPM and BFM results were within 3.0% (average 1.46%) across four test cases with varying diffusivity ratios and jump magnitudes.
Fluid-Solid Coupled Advection-Diffusion: A steady laminar flow ($Re=40$) past a cylinder with interfacial jumps was simulated. The VPM results again showed strong agreement with BFM benchmarks, with a relative deviation of 1.59%. The method accurately reproduced the scalar distribution and flux jumps near the interface.

Significance and Claims
The paper claims that the proposed VPM offers a robust and efficient alternative to body-fitted meshes for conjugate scalar transport problems involving complex interfacial geometries and jump conditions. The primary significance lies in the ability to solve coupled fluid-solid problems using a single unified governing equation while accurately enforcing local jump conditions. The authors note that this approach avoids the non-physical artifacts common in traditional VPM implementations for Neumann conditions.

The authors modestly position this work as a foundational step, suggesting that the method's ability to handle complex geometries makes it suitable for future applications in rough walls and complex heat transfer surfaces. They also highlight the potential for combining this VPM with adjoint methods for shape and topology optimization in radiative, turbulent, and conjugate heat transfer problems, though these specific applications are identified as directions for future work rather than immediate results of the current study.

More like this