Density-based topology optimization for 3D-printable building structures

This paper presents the study of a new penalty method for density-based topology optimization. The focus is on 3D-printable building structures with optimized stiffness and thermal insulation properties. The first part of the paper investigates the homogenized properties of 3D-printed infill patterns and in the second part a new penalty method is proposed and demonstrated. The method presents an alternative way to implement multi-material topology optimization without increasing computational cost. A single interpolation function is created, based on the homogenized properties of a triangular infill pattern. The design variables are linked to the different possible infill densities of the pattern. A high density represents a solid structure with high stiffness, but weak thermal properties, while an intermediate density provides the structure with good insulation qualities. On the other hand, when the air cavities become too large (i.e., low infill densities), the heat flow by convection and radiation again decreases the thermal performances of the material. The optimization study is performed using the GCMMA algorithm combined with a weighted-sum dual objective. One part of the equation aims to maximize stiffness, while the other attempts to minimize the thermal transmittance. Different case studies are presented to demonstrate the effectiveness of this multi-physics optimization strategy. Results show a series of optimized topologies with a perfect trade-off between structural and thermal efficiency.