Human-Informed Topology Optimization: interactive application of feature size controls

Abstract This paper presents a new topology optimization framework in which the design decisions are made by humans and machines collaboration. The Human-Informed Topology Optimization approach eases accessibility of tools enables improved identification for so-called ‘everyday’ ‘in-the-field’ situations. is based on standard density-based compliance minimization. However, engineer enabled to actively use their experience expertise locally alter minimum feature size requirements. done conducting short initial solution prompting evaluate quality. user can identify potential areas concern material distribution. In these areas, requirement be altered as deemed necessary user. algorithm rigorously resolves problem using updated filtering map, resulting solutions that eliminate, merge, or thicken topological members concern. demonstrated 2D benchmark examples extension 3D shown. Its ability achieve performance improvement with few computational resources buckling stress concentration examples.