Stress-Driven Anisotropic Voronoi Structures
ACM Transactions on Graphics · 16.11.2026
DOI: 10.1145/3864020 ↗Kurz gefasst
Fachartikel in der Zeitschrift ACM Transactions on Graphics, online erschienen am 16.11.2026. Verfasst von Dennis Bukenberger und 2 weiteren Autor:innen. Die Arbeit nennt 59 Literaturangaben.
Abstract (Originalsprache)
This paper presents a method for constructing anisotropic 3D Voronoi infill structures whose morphology adapts to principal stress directions induced by external loads. Existing stress-aligned infill methods predominantly generate truss-like architectures, whereas wall-like structures are known to achieve higher stiffness at comparable material volume. Our method bridges this gap by enabling the construction of stress-aligned wall-like infills. This is achieved through stress-driven sampling and a Lloyd relaxation constrained by the underlying stress field. A novel adaptive metric, the L q norm, is introduced to compute anisotropic Centroidal Voronoi Tessellations (CVTs) with cells aligned to locally varying stress directions. The final structure is generated by carving shape-preserving harmonic distance fields within the Voronoi cells, producing a continuous wall-like microstructure. The resulting infills outperform existing wall- and truss-based approaches in stiffness at comparable material volume while maintaining substantially lower construction times. We demonstrate these properties through finite-element analyses and physical experiments on 3D printed specimens.
Zitieren
APA
Bukenberger, D., Wu, J., & Westermann, R. (2026). Stress-Driven Anisotropic Voronoi Structures. ACM Transactions on Graphics. https://doi.org/10.1145/3864020
Vancouver
Bukenberger D, Wu J, Westermann R. Stress-Driven Anisotropic Voronoi Structures. ACM Transactions on Graphics. 2026. doi:10.1145/3864020
BibTeX
@article{Bukenberger2026,
author = {Bukenberger, Dennis and Wu, Jun and Westermann, Rüdiger},
title = {Stress-Driven Anisotropic Voronoi Structures},
journal = {ACM Transactions on Graphics},
year = {2026},
doi = {10.1145/3864020}
}
EUVIMED