Programmable Mechanical Metamaterials based on Spinodal Architectures

The thesis explores the characteristics and properties of anisotropic metastructures produced using building blocks of spinodoids, a metamaterial known for its spatially-variant architecture and potential for diverse mechanical functionalities. The current inverse design of spinodoids employs an artificial neural networks-based framework for characterizing and reconstructing spinodoids (Kumar, Tan et al. (2020)). Nevertheless, the complex relationships between properties and structures in spinodoid elements pose challenges and inefficiencies in their simulation and analysis through computational design methods. Consequently, a homogenisation method, which aims to represent the spinodoid metamaterial properties with a homogenised block finite element (FE) model, is proposed to address these challenges. The target mechanical properties are evaluated from the static and dynamic FE analysis of the spinodoid structures, including stiffness evaluation, coupled deformation analysis, modal analysis and wave propagation analysis. In the static analysis, this project focuses on the stiffness evaluation, which was carried out for individual spinodoid blocks (Chapter 3) and more complex assembly models (Chapter 4). A more advanced static analysis for deformation coupling was also conducted on spinodoid assembly models. The outcomes of these analyses contribute to a deeper understanding of how spinodoid metamaterials with variant architecture can behave under various loading conditions. These static cases also investigate how spinodoid block models can be properly represented by homogenised block models in FE simulations. Furthermore, the dynamic analyses, including modal and wave propagation analyses, were conducted for spinodoid assembly models in Chapter 4, shedding light on the dynamic characteristics of these structures assigned with different combinations of geometric parameters. These dynamic cases also prove that homogenised block models can present dynamic behaviours similar to those of the spinodoid models if they are both assigned identical material properties. The optimized spinodoid FE models were also fabricated using Stereolithography (SLA) technique for subsequent mechanical testing. The experimental results of the spinodoid structures were then compared with corresponding FE simulation data to assess the spinodoid property evaluation framework. This work demonstrates the feasibility and significance of employing spinodoid in architectural design through computational design, fabricating spinodoid structures with advanced manufacturing techniques, and finally experimental testing of spinodoid structures. It also characterises the mechanical performance of the spinodoid specimens and comparative evaluation with the computational models.