Efficiency and Stability Trade-offs in FSI Simulations of Coronary Artery Blood Flow

E. Guleryuz[1], A. Taha[1], S. Koric[1], Q. Lu[1], M. Vellakal[1]
[1]National Center for Supercomputing Applications (NCSA), University of IL at Urbana-Champaign., USA
Published in 2019

Some of the most common fluid-structure interaction (FSI) applications in biomedical research include blood flow in the veins and arteries. COMSOL Multiphysics® simulation software offers two main multiphysics schemes for FSI problems: partitioned (segregated) and monolithic (fully-coupled). The monolithic approach treats the fluid and solid domains as a single continuum. In the partitioned approach, governing equations of fluid and solid domains are solved separately in the course of a simulation. For both partitioned and monolithic approaches, several direct and iterative linear solver options are available. In general, fully-coupled schemes and direct solvers are known to be more robust. This robustness comes with the expense of computationally more demanding linear systems to be solved. The particular choices of multiphysics scheme and linear solver will determine the simulation time and memory footprint of the simulation. Analysts need to achieve a balance between numerical stability and computational efficiency. To construct a high-level guidance map, we perform numerical experiments with several scheme and solver combinations. We model the fluid structure interaction (FSI) of the coronary artery blood flow under pulsatile flow conditions and quantify the trade-offs between computational efficiency and numerical stability in choosing among alternative simulation setups with the objective of reaching the best simulation strategy for such FSI application.