| Abstract Scope |
Metamaterials (MMs), known for their outstanding properties, have garnered attention from diverse fields such as mechanical, thermal, electromagnetic and acoustic. MMs’ properties are realised by their underlying architectures, some of which have been effectively derived from various designing strategies. Our recent work (Wang and Panesar, 2022) proposed a two-step designing strategy for realising multi-scale architecture by combining topology optimisation (TO) and machine learning (ML). 2D mechanical metamaterials with various unit cell geometries were explored. This study advances on the two-step strategy by extending the geometry to 3D to design truss and plate-based metamaterials. Multiple objective functions incorporating thermal, buckling, vibration etc. are explored to realise multi-functional metamaterials. We also benchmark the two-step strategy against other TO-based strategies, inverse homogenisation and dehomogenisation. The evaluation highlights the structural performance, computational efficiency and flexibility to include manufacturing aspects. Recommendations for engineering applications will be discussed. |