Abstract

A multiphase FSI framework using only open-source software has been developed, utilising components able to run on high-performance computing platforms. A partitioned approach is employed, ensuring a separation of concerns (fluid, structure and coupling), allowing design flexibility and robustness while reducing future maintenance efforts. Multiphase FSI test cases have been simulated and compared with published results and show good agreement. Simulation of a model representing an elastic aircraft wing with a fluid (fuel) sloshing inside is presented. This demonstrates the ability of this multiphase FSI framework in simulating complex and challenging cases involving a free liquid surface.

INTRODUCTION

The interaction between multiphase flow and elastic structures is an important phenomenon in a wide range of scientific and engineering disciplines, such as an aircraft wing with sloshing fuel tank (Saltari et al. 2021; Gambioli et al. 2019; Mastroddi et al. 2019; Titurus et al. 2019; Mastroddi et al. 2020; Gambioli et al. 2020) and the impact of ocean waves on elastic ocean structures (Gomes et al. 2020). Accurately simulating multiphase fluid-structure interaction (FSI) can help reveal the mechanism behind important and complex real-world phenomena, allowing for important design considerations such as how to protect the elastic structure from fatigue or failure, or to make use of the phenomenon to achieve active/passive control of the system. There is significant demand to develop an efficient and open-source numerical tool for the investigation of this phenomenon. Due to the nonlinear, timedependent and multi-physical nature of various multiphase FSI problems, a simulation tool that is both robust and highly scalable in parallel computing terms is challenging. There are notable commercial FSI solvers. However, few of them can achieve both numerical robustness and high scalability while also being able to tackle multiphase multi-physics FSI problems. Commercial software, such as ANSYS (Rao 2003) and COMSOL (Curtis et al. 2013) are able to provide fullycoupled FSI simulations, but it can be challenging to run large parallel simulations using them. Martínez-Ferrer et al. (2018) established an FSI simulation tool using OpenFOAM. Both fluid and structure domains were discretised with the Finite Volume method and solved by OpenFOAM using up to four CPU cores. Integrating both fluid and structural domains into a single library with a code-specific data mapping method is a good way to model multi-physical problems with small numbers of physical domains. However, maintaining these codes and incrementally adding more physical domains is then onerous, especially when different domain types cannot be tackled using the same discretisation strategy.

This content is only available via PDF.
You can access this article if you purchase or spend a download.