Efficient fluid structure interaction simulation of vocal fold oscillations using a homogenized Lattice Boltzmann Method

A. Kummerländer, Bogac Tur, Maik Haase, Fedor Bukreev, Michael Döllinger, Mathias J. Krause, S. Kniesburges

TL;DR: Providing both high computational efficiency and physiological results, the proposed HLBM-6MM solver enables systematic parameter studies of phonation previously infeasible with traditional FSI approaches.

Computational modeling of human phonation is a uniquely challenging fluid–structure interaction (FSI) problem. It is constrained both by the large computational cost and numerical restrictions of existing methods, particularly in handling vocal fold (VF) contact. To combine physiological fidelity with high computational efficiency, this study introduces a novel FSI coupling strategy in a highly optimized computational workflow. The core methodological advance lies in mapping the structural dynamics of a Six-Mass-Model (6MM) into a time-dependent porosity field, efficiently integrating it with a homogenized Lattice Boltzmann Method (HLBM) for the fluid dynamics. The fluid is modeled using the filtered Brinkman– Navier–Stokes equations, platform-transparently discretized by an HLBM variant in the OpenLB framework. By representing the moving VFs as a time-dependent porosity field, the solver implicitly handles complex geometry changes and contact without remeshing, while the two-way coupled 6MM captures both elastic motion and collision. Simulations in a 3D laryngeal channel produced stable, self-sustained oscillations at physiologically plausible values: fundamental frequency ( 𝑓 0 ) ≈ 248 Hz, mean subglottal pressure ≈ 1.01 kPa, and an open quotient ≈ 0.64. The model robustly achieved complete glottal closure in each cycle, yielding credible peak contact forces ( ≈ 27 mN). Providing both high computational efficiency (up to 60 oscillation cycles/hour on a NVIDIA RTX 2000 Ada laptop GPU) and physiological results, the proposed HLBM-6MM solver enables systematic parameter studies of phonation previously infeasible with traditional FSI approaches.