TPMS-based hyperboloidal primitive architectured structures with a novel hybridization method

Abstract

Herein, by drawing inspiration from quantum mechanics of atomic and molecular orbitals, similar architected meta-structures (SAMs) and hybridization of SAMs (HSAMs) involving double-hybridized SAMs (DSAMs) and triple-hybridized SAMs (TSAMs) are introduced. HSAMs involves evaluating a specific meta-structure from SAMs, reinforcing it to compensate for elastic stiffness deficiencies, enhance structural integrity, mechanical properties, and isotropy. Through linear and nonlinear finite element modelling (FEM) and compression tests, anisotropy, mechanical properties and energy absorption (EA) characteristics of these architected meta-structures are quantified and validated against additively manufactured samples. Further, machine learning approaches involving U-Net, eXtreeme gradient boost (XGBoost), Multilayer Perceptron (MLP), and random forest (RF) are employed to predict their deformation behaviors and mechanical responses. Results revealed that DSAMs and TSAMs are effective means for improving isotropy, mechanical properties, and EA characteristics. Remarkable improvements of 29% and 42% in isotropy, 63% and 130% in Young's modulus, and 38% and 60% in EA are achieved for DSAMs and TSAMs, respectively. Lastly, XGBoost outperforms MLP and RF in predicting mechanical responses. © 2026 Elsevier Ltd