Density-based topology optimization using a deformable mesh

Abstract

Conventional density-based topology optimization suffers from blurred or stair-stepped boundaries of the optimized structure, difficulty in converting outputs to usable designs, and a significant increase in computational cost when fine meshes are used. To address these challenges, we propose a density-based topology optimization method that allows the deformation of the reference mesh by incorporating nodal displacements as additional design variables, which offers higher design flexibility even with coarse meshes. This approach improves resolution without the burden of dense discretization and eliminates post-processing by directly producing a binarized density distribution with smooth boundaries. Demonstrated on benchmark problems, our method is expected to streamline the design-to-fabrication process and holds promise for more efficient and accurate structural optimization.