Multiscale Topology Optimization of Functionally Graded Multi-Morphology TPMS Structures

This research presents a progressive multiscale topology optimization framework for designing additively manufactured components with functionally graded multi-morphology TPMS lattices. To overcome the limitations of conventional methods including high computational cost, neglected local stresses, and unsafe morphology transitions, we propose a complete methodology. It integrates a density-based multi-material formulation, physics-admissible constitutive surrogates, Gaussian radial basis function blending for smooth transitions, stress-constrained optimization for local failure control, and physics-based machine learning to accelerate fine-scale analyses. All components are unified within a prototype software environment that ensures a traceable workflow from design to verified, AM-ready geometry. This work advances TPMS optimization from compliance-driven design toward a safer, more efficient, and industrially practical architected-material technology.