Self-generating multiscale configurations, their CAD features in support of 3D printing and their CAE efficiencies

Abstract

Additive manufacturing enables the production of a number of multiscale configurations such as lattice structures. However, due to their multiscale complexities, the acquisition of explicit, high-fidelity, resource-saving digital description of lattice configurations is still technically challenging. This article is aimed to introduce a general algorithm for digital presentation of lattice structure, whose constituting cells can be spatially-varying. The natural suitability of the present algorithm with additive manufacturing can be summarised as follows. Firstly, the designed lattice here can be represented fully in line with Computer-Aided Design (CAD) conventions. Secondly, the designed lattice can functionally approximate any multiscale configurations in the sense of resulting in similar responding fields under given loading conditions. Thirdly, the designed lattice can be digitally memorised in a highly compact manner, and an unzipping scheme in parallel with its additive manufacturing process is thus proposed to maintain the number of CAD control points in memory at a low level. Fourthly, mechanical properties of the designed lattice, both its overall compliance and its localised properties, such as its strength, can be evaluated instantly, with the use of a machine-learning-based asymptotic homogenisation and localisation method.