A geometry-driven optimization framework for tuning cyclic stiffness and dissipative response in additively manufactured implant-like lattices.

Abstract

This study presents a geometry-driven optimization framework for additively manufactured implant-like lattice architectures, focusing on the control of architecture-dependent cyclic mechanical response. The proposed Manufacturability-Aware Surrogate-Assisted Topology Optimization framework integrates a genetic algorithm, a Gaussian-process surrogate model, and gradient-based refinement within a unified pipeline that enforces additive manufacturing constraints. The framework uses a deformation-derived mechanical response proxy to guide the ranking of manufacturable lattice architectures according to their expected stiffness-related and dissipation-related behavior. Optimized M-88 and M-2525 architectures were fabricated from PLA and evaluated under quasi-static and cyclic compression. The finer M-2525 architecture exhibited higher quasi-static stiffness and a more storage-dominated cyclic response, whereas the coarser M-88 architecture showed a stronger loss-modulus contribution and more dissipative cyclic behavior. These cyclic trends were consistent with the quasi-static compression results, indicating that geometry-driven design can shift the response of lattice architectures between stiff elastic behavior and more viscous or dissipative behavior. Overall, the proposed framework provides a computational-experimental route for tuning geometry-dependent cyclic response in manufacturable additively manufactured lattices.