Machine Learning-Guided Design of Gradient Soft Materials via Digital Light Processing
발표자
김미소 (한국과학기술원)
연구책임자
김미소 (한국과학기술원)
초록
내용
Grayscale digital light processing (g-DLP) enables control of light intensity to fabricate soft materials with spatially programmable mechanical properties. In this talk, we introduce a synergistic design platform that combines g-DLP with machine learning-guided multi-objective optimization and a custom-designed viscoelastic polyurethane acrylate resin system. The resin exhibits a record-wide tunable modulus range (8.3 MPa to 1.2 GPa) with excellent damping performance, while the AI framework uses Bézier-based optimization to generate optimal gradient distributions. Our structures achieve up to 83% reduction in strain concentration and significant improvements in fracture delay and fatigue durability in applications from artificial cartilage to automotive bumpers. This work pioneers a data-driven pathway for engineering next-generation soft materials through the integration of chemistry, photoprinting, and machine learning.