A stable partitioned FSI algorithm for incompressible flow and elastic solids

TL;DR

This paper introduces a stable, high-order partitioned fluid-structure interaction algorithm that remains stable regardless of density ratios, effectively handling added-mass effects without sub-iterations.

Contribution

The paper develops a novel AMP scheme with Robin boundary conditions for stable coupling of incompressible flow and elastic solids, extending previous methods and ensuring stability for all density ratios.

Findings

The AMP scheme is stable for any solid-to-fluid density ratio.

Traditional Dirichlet-Neumann coupling is unconditionally unstable with added-mass effects.

Numerical solutions match exact traveling wave solutions across various solid densities.

Abstract

A stable partitioned algorithm for coupling incompressible flows with compressible elastic solids is described. This added-mass partitioned (AMP) scheme requires no sub-iterations, can be made fully second- or higher-order accurate, and remains stable even in the presence of strong added-mass effects. The approach extends the scheme of Banks et al. [1,2] for compressible flow, and uses Robin (mixed) boundary conditions with the fluid and solid solvers at the interface. The AMP Robin conditions are derived from a local characteristic decomposition in the solid at the interface. Two forms of the Robin conditions are derived depending on whether the fluid equations are advanced with a fractional-step method or not. A normal mode analysis for a discretization of an FSI model problem is performed to show that the new AMP algorithm is stable for any ratio of the solid and fluid densities, including the case of very light solids when the added-mass effects are large. In contrast, it is shown that a traditional partitioned algorithm involving a Dirichlet-Neumann coupling for the same FSI problem is formally unconditionally unstable for any ratio of densities. Exact traveling wave solutions are derived for three FSI model problems of increasing complexity, and these solutions are used to verify the stability and accuracy of the corresponding numerical results obtained from the AMP algorithm for the cases of light, medium and heavy solids.