TY - GEN
T1 - ALL-HEX MESHING STRATEGIES FOR DENSELY PACKED SPHERES
AU - Lan, Yu Hsiang
AU - Fischer, Paul
AU - Merzari, Elia
AU - Min, Misun
N1 - Publisher Copyright:
© 2021 held by the authors of the individual papers.
PY - 2021
Y1 - 2021
N2 - We develop an all-hex meshing strategy for the interstitial space in beds of densely packed spheres that is tailored to turbulent flow simulations based on the spectral element method (SEM). The SEM achieves resolution through elevated polynomial order N and requires two to three orders of magnitude fewer elements than standard finite element approaches do. These reduced element counts place stringent requirements on mesh quality and conformity. Our meshing algorithm is based on a Voronoi decomposition of the sphere centers. Facets of the Voronoi cells are tessellated into quads that are swept to the sphere surface to generate a high-quality base mesh. Refinements to the algorithm include edge collapse to remove slivers, node insertion to balance resolution, localized refinement in the radial direction about each sphere, and mesh optimization. We demonstrate geometries with 102–105 spheres using ≈ 300 elements per sphere (for three radial layers), along with mesh quality metrics, timings, flow simulations, and solver performance.
AB - We develop an all-hex meshing strategy for the interstitial space in beds of densely packed spheres that is tailored to turbulent flow simulations based on the spectral element method (SEM). The SEM achieves resolution through elevated polynomial order N and requires two to three orders of magnitude fewer elements than standard finite element approaches do. These reduced element counts place stringent requirements on mesh quality and conformity. Our meshing algorithm is based on a Voronoi decomposition of the sphere centers. Facets of the Voronoi cells are tessellated into quads that are swept to the sphere surface to generate a high-quality base mesh. Refinements to the algorithm include edge collapse to remove slivers, node insertion to balance resolution, localized refinement in the radial direction about each sphere, and mesh optimization. We demonstrate geometries with 102–105 spheres using ≈ 300 elements per sphere (for three radial layers), along with mesh quality metrics, timings, flow simulations, and solver performance.
UR - http://www.scopus.com/inward/record.url?scp=85139514598&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85139514598&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85139514598
T3 - Proceedings of the 29th International Meshing Roundtable, IMR 2021
SP - 293
EP - 305
BT - Proceedings of the 29th International Meshing Roundtable, IMR 2021
A2 - Tomov, Vladimir Z.
A2 - Gu, Xianfeng
A2 - Verdicchio, John
PB - Zenodo
T2 - 29th International Meshing Roundtable, IMR 2021
Y2 - 21 June 2021 through 25 June 2021
ER -