Unified GPU-based simulation of porous flow, absorption, and diffusion in cloth-liquid coupling.

Area of Science:

• Computational physics
• Fluid dynamics simulation
• Computer graphics

Background:

• Modeling cloth-liquid interactions in particle-based simulations is computationally challenging due to coupled phenomena.
• Existing methods struggle with stability and efficiency for processes like porous flow, absorption, and diffusion.

Purpose of the Study:

• To develop a GPU-based framework for stable and efficient cloth-liquid coupling in particle simulations.
• To integrate porous flow, absorption, and diffusion into a unified Smoothed Particle Hydrodynamics (SPH) pipeline.

Main Methods:

• Utilized a virtual-pressure formulation inspired by Darcy's law for robust porous flow near boundaries.
• Implemented saturation-based mass exchange for absorption/emission and sequential redistribution for diffusion.
• Employed a Bitonic sort-based GPU hashing structure for efficient neighbor searching.

Main Results:

• The proposed hashing scheme significantly reduced sorting overhead compared to baseline GPU implementations.
• The integrated pipeline demonstrated stable wetting behavior and bounded mass variation in simulations.
• Quantitative analyses confirmed improved efficiency and scalability with increasing particle counts.

Conclusions:

• The developed framework offers an efficient and numerically stable solution for GPU-based porous cloth-liquid simulations.
• This approach advances the simulation of complex fluid-fabric interactions in computer graphics and physics.