Investigation on the effect of density ratio on the convergence behavior of partitioned method for fluid–structure interaction simulation. (July 2020)

• Record Type: Journal Article
• Authors: Ha, Sang Truong; Choi, Hyoung Gwon
• Is Part Of: Journal of fluids and structures. Volume 96(2020)
• Publication Date: 2020-07
• Subjects: Fluid–structure interaction -- Strong-coupling partitioned algorithm -- Finite element method -- Density ratio -- Mooney–Rivlin model; Fluid-structure interaction -- Periodicals; Fluid mechanics -- Periodicals; Structural dynamics -- Periodicals; Structural analysis (Engineering) -- Periodicals; 620.106
• DOI: 10.1016/j.jfluidstructs.2020.103050
• Languages: English
• ISSNs: 0889-9746

Abstract

Abstract: In order to investigate the effect of density ratio of fluid and solid on the convergence behavior of partitioned FSI algorithm, three strong-coupling partitioned algorithms (fixed-point method with a constant under-relaxation parameter, Aitken's method and Quasi-Newton inverse least squares (QN-ILS) method) have been considered in the context of finite element method. We have employed the incompressible Navier–Stokes equations for a Newtonian fluid domain and the total Lagrangian formulation for a non-linear motion of solid domain. Linear-elastic (hyper-elastic) model has been employed for solid material with small (large) deformation. A pulsatile inlet-flow interacting with a 2D circular channel of linear-elastic material and a pressure wave propagation in a 3D flexible vessel have been simulated. Both linear-elastic and hyper-elastic (Mooney–Rivlin) models have been adopted for the 3D flexible vessel. From the present numerical experiments, we have found that QN-ILS outperforms the others leading to a robust convergence regardless of the density ratio for both linear-elastic and hyper-elastic models. On the other hand, the performances of the fixed-point method with a constant under-relaxation parameter and the Aitken's method depend strongly on the density ratio, relaxation parameter selected for coupling iteration, and degree of deformation. Although the QN-ILS of this work is still slower than a monolithic method for serial computation, it has an advantage of easier parallelization due to the modularity of the partitioned FSI algorithm. Highlights: Performance comparison of three strong-coupling partitioned algorithms. Effect of density ratio on the convergence behaviors of partitioned methods. Strong-coupling partitioned algorithms for hyperelastic(Mooney–Rivlin) model. Comparison of a monolithic method and QN_ILS method for a light solid.