Design, additive manufacturing, and machine learning prediction of multi-material diamond TPMS structure for improved mechanical performance

This study addresses the challenge of designing mechanical metamaterials with customizable properties to meet the growing demand for lightweight, high-strength, and energy-absorbing materials in advanced engineering applications. Traditional manufacturing techniques and single-material designs often limit the tunability and performance of these structures. To overcome these constraints, this research explores the design and fabrication of multi-material diamond lattice structures using fused deposition modeling (FDM) process. Triply periodic minimal surface (TPMS) diamond lattice structures were fabricated with polylactic acid (PLA) for structural rigidity and thermoplastic polyurethane (TPU) for flexibility and energy absorption. Compression testing revealed the influence of material distribution and geometry on the mechanical performance. The energy loss/shape recovery of the proposed lattice structures was also studied. The results showed that an increase in the relative density and PLA% resulted in a reduction in energy loss% and an increase in the compression modulus, energy absorption, and peak load. Artificial neural network (ANN) machine learning model was used to predict the design parameters based on the desired mechanical properties. The predicted ANN lattice structures with varying objectives were also tested and showed excellent agreement between the experimental and desired values. This study highlights the integration of AM and machine learning as transformative tools in the development of next-generation multi-material metamaterials, offering significant potential for aerospace, biomedical, and other high-performance applications.