Improving the Performance of Partitioned Methods for Solid-Fluid Coupling
- Author(s): Akbay, Muzaffer
- Advisor(s): Shinar, Tamar
- et al.
Partitioned approaches, where the fluid and solid solvers are treated as black boxes with limited exposed interfaces, provide one convenient way of coupling solid and fluid solvers due their modularity and black-box design. However, they are often not preferred due to their stability and performance issues.
In this thesis, some of the crucial problems of partitioned approaches are addressed. A novel extended partitioned method for two-way solid-fluid coupling is presented, where the method achieves improved stability and extended range of applicability over standard partitioned approaches through three techniques. First, the black-box solvers are coupled through a small, reduced-order monolithic system, which is constructed on the fly from input/output pairs generated by the solid and fluid solvers. Second, a conservative, impulse-based interaction term is proposed to couple the solid and fluid rather than typical pressure-based forces. We show that both of these techniques significantly improve stability and reduce the number of iterations needed for convergence. Finally, a novel boundary pressure projection method is presented that allows for the partitioned simulation of a fully enclosed fluid coupled to a dynamic solid, a scenario that has been problematic for partitioned methods. The benefits of the methods are demonstrated by coupling Eulerian fluid solvers for smoke and water to Lagrangian solid solvers for volumetric and thin deformable and rigid objects in a variety of challenging scenarios. The application of the extended partitioned method to Smoothed Particle Hydrodynamics fluids requires some simple modifications to the system. These modifications are described and demonstrated on a publicly available solver. Finally, the boundary pressure projection method is revised further to handle solids with higher densities. The benefits of the revised boundary pressure projection method are demonstrated on various scenarios including multiple neighboring closed regions.