High-Fidelity TPMS Shell Structures: A New Algorithm for Enhanced Structural Performance
ORAL
Abstract
Shell structures based on thickened Triply Periodic Minimal Surfaces (TPMS) have emerged as a promising class of architected materials due to their unique geometric features, which impart exceptional mechanical properties, including high specific strength and excellent manufacturability via additive manufacturing.
However, despite this growing interest, most TPMS implementations in physical applications rely on approximations using Periodic Nodal Surfaces (PNS). While this approach facilitates computational modeling and fabrication, it introduces deviations from true TPMS geometries. These discrepancies can significantly impact structural behavior under particular loading conditions, where minor geometric distortions can lead to substantial variations in internal stress distribution.
To address this issue, we proposed a new algorithm capable of generating (1) higher fidelity TPMS approximations—closer to zero mean curvature—and (2) a broader variety of TPMS—hundreds of new geometries. This capability allows for the design of shell structures that are more precisely tailored to specific mechanical performance criteria.
This work presents a comparative analysis between shell structures generated using conventional PNS-based TPMS and those produced via this novel algorithm. Through simulation and experimental validation, we show that the proposed algorithm produces mechanical metamaterials with superior performance across multiple metrics compared to their conventional counterparts. Notably, under hydrostatic compression, the proposed structures exhibited a more uniform Von Mises stress distribution, indicating enhanced resistance to failure. Additionally, the evaluated bulk modulus approached the theoretical Hashin-Shtrikman upper bound, underscoring the efficiency of the proposed geometries.
To validate these findings, a subset of the designed structures was fabricated using additive manufacturing and tested in a high-pressure chamber. These experiments were designed to replicate the hydrostatic conditions typical of deep-sea environments, providing empirical support for the computational results.
However, despite this growing interest, most TPMS implementations in physical applications rely on approximations using Periodic Nodal Surfaces (PNS). While this approach facilitates computational modeling and fabrication, it introduces deviations from true TPMS geometries. These discrepancies can significantly impact structural behavior under particular loading conditions, where minor geometric distortions can lead to substantial variations in internal stress distribution.
To address this issue, we proposed a new algorithm capable of generating (1) higher fidelity TPMS approximations—closer to zero mean curvature—and (2) a broader variety of TPMS—hundreds of new geometries. This capability allows for the design of shell structures that are more precisely tailored to specific mechanical performance criteria.
This work presents a comparative analysis between shell structures generated using conventional PNS-based TPMS and those produced via this novel algorithm. Through simulation and experimental validation, we show that the proposed algorithm produces mechanical metamaterials with superior performance across multiple metrics compared to their conventional counterparts. Notably, under hydrostatic compression, the proposed structures exhibited a more uniform Von Mises stress distribution, indicating enhanced resistance to failure. Additionally, the evaluated bulk modulus approached the theoretical Hashin-Shtrikman upper bound, underscoring the efficiency of the proposed geometries.
To validate these findings, a subset of the designed structures was fabricated using additive manufacturing and tested in a high-pressure chamber. These experiments were designed to replicate the hydrostatic conditions typical of deep-sea environments, providing empirical support for the computational results.
–
Presenters
-
Victor Riera Naranjo
- Georgia Institute of Technology