ECCOMAS 2024

Stable, Efficient, and Higher-order Mesh Movement in Large-displacement Fluid-structure Interaction

  • Schwentner, Teresa (Graz University of Technology)
  • Bošnjak, Domagoj (Graz University of Technology)
  • Fries, Thomas-Peter (Graz University of Technology)

Please login to view abstract download link

Mesh movement is a crucial aspect of simulations where the domain is time-dependent, e.g., in fluid-structure interactions (FSI). The fluid domain changes over time and hence the mesh moves. The accuracy of the simulation highly depends on the quality of the mesh [1]. Various different methods for mesh updates have been proposed [2]. However, these methods are often not robust enough in case of higher-order meshes. Consequently, instead of updating the mesh, in this work a new mesh is generated in every time step [3], based on a block-structure consisting of valid coarse quads. Transfinite maps are used to generate sub-meshes within the block-structure, allowing any order and number of elements. The geometry information is assigned to the block edges by user-input. Whereas, on the FSI-interface the geometry data is a result of the simulation itself. The remaining nodes of the block-structure are moved to generate valid meshes throughout the simulation. For moderately large movements, these manipulations of the block-structure may be implemented based on case-dependent, empirical rules. However, larger displacements and more blocks in the block-structure require stable and efficient update schemes for the blocks (instead of the resulting mesh in the simulation). Hence, different methods for this movement are compared. First, a pseudo-solid approach applied to the block-structure nodes is used, where the displacement on the FSI-interface is prescribed as a boundary condition. Valid blocks and hence valid elements are not guaranteed. Thus, as a second approach, an optimization scheme is applied to the blockstructure, using gradient descent. The optimization scheme prevents the generation of invalid blocks and requires minimal user-input. This task is fast, as only the block nodes are optimized. Herein, the quality and applicability of these methods are compared. REFERENCES [1] Y. Bazilevs, K. Takizawa and T.E. Tezduyar, Computational Fluid–Structure Interaction: Methods and Applications. John Wiley & Sons, Ltd, 2013. [2] A. Shamanskiy and B. Simeon, Mesh moving techniques in fluid-structure interaction: robustness, accumulated distortion and computational efficiency, Computational Mechanics, Vol. 67, pp. 583–600, 2021. [3] T. Schwentner and T.P. Fries, Fluid-structure interaction with fully coupled mesh generation, Proc. Appl. Math. Mech., Vol. 23, e202300067, 2023.