3D-Printed Twisting Tubular Metamaterials with Tunable Mechanical and Torsional Characteristics

Abstract

Innovative tubular compression-torsion mechanical metamaterials with potential applications in sensors, actuators, and energy harvesting are introduced. The metamaterials are parametrically designed and fabricated through stereolithography appearance 3D printing method, and their behaviour is analyzed using experiments, finite element method, and machine learning (ML). The effects of unit cell geometrical parameters, such as the unit cell size, the struts length, the angle between the struts, and the in-plane and out-of-plane thicknesses, on the performance of the structure, are investigated. Additionally, the influence of section geometry, longitudinal and circumferential cell numbers, and the unit cell assembly layouts on the metamaterial's twisting capability and mechanical properties is studied. The results demonstrate the tunability of the metamaterial by setting the unit cell design parameters, compression-torsion conversion property, volume fraction, and effective Young's modulus. ML is employed to predict a geometric design suitable to fit known working specifications. ML facilitates tuning the proposed metamaterial based on the desired responses which are expected from the structure. The problem is addressed through a multi-target supervised regression using the Random Forest regressor algorithm with optimized hyperparameters.