A Parallel Partitioned Approach On Fluid-Structure Interaction Simulations Using the Multiscale Universal Interface Coupling Library

Abstract

Fluid-Structure Interaction (FSI) is a phenomenon that appears in a wide range of scientific research and engineering applications at different spatial and temporal scales. There are many in-house/commercial solvers capable of modelling FSI, but high numerical robustness and high scalability codes are still in demand. In this study, a numerical framework for FSI simulations has been developed using a partitioned approach aimed at both high numerical robustness and good computational scalability. Open-source software is used for each component of the coupled solution, with OpenFOAM and FEniCS adopted to simulate the computational fluid dynamics and computational structural mechanics, respectively. A coupling interface between the fluid and structural computational domains is realised using the open-source Multiscale Universal Interface (MUI) scientific code coupling library. To achieve a tight and stable coupling, various FSI coupling algorithms have been implemented in the MUI. The behaviour of this framework has been assessed for simulations of a blunt trailing edge hydrofoil at different working conditions with vortex-shedding induced vibration.

A High-performance Open-source Solution for Multiphase Fluid-Structure Interaction

“…1 shows the results of a 300-million-cell to 300-million-cell coupled case with a uniform 50/50 split of resources; this shows good parallel performance on up to 12,000 MPI tasks. The performance of the framework has also been demonstrated in our previous work (Liu et al, 2021b), which has shown acceptable scalability, with a parallel efficiency (i.e., speedup over number of MPI tasks) of 68% for 1,000 MPI tasks for FSI simulations using this framework.…”

“…Systematic verifications and validations have been carried out for the coupling between OpenFOAM and FEniCS using the MUI library with a single-phase fluid solver in previous work (Liu et al, 2021b). It covers both two-dimensional (2-D) and three-dimensional (3-D) aero/hydroelastic cases within both laminar and turbulence regimes.…”

preCICE v2: A sustainable and user-friendly coupling library

“…Often, the categories of use cases are not strict. MUI, for example, has recently also been used for fluid-structure interaction 30 . At the same time, current work on preCICE aims towards certain multi-scale coupling patterns (cf.…”

“…The example is meant to serve as a first impression of how the preCICE API looks and how to use it along with deal.II. 30 code_aster adapter on GitHub: https://github.com/precice/code_aster-adapter, GPLv2…”

preCICE v2: A sustainable and user-friendly coupling library

“…The code is available on GitHub 30 under the GPLv2 license and can be easily extended by adding more coupling fields in adapter.py and providing their names as arguments to the method adapter.writeCouplingData().…”