Minimum size control for binary topology optimization

Abstract Topology optimization methods employing binary (also known as discrete) design variables currently lack mathematical formulations to ensure length scale control in their solutions. This paper proposes and applies a morphology-mimicking filtering scheme to provide a minimum size control (often also referred to as minimum length scale control) in this class of binary designs. The Topology Optimization of Binary Structures (TOBS) method was chosen as the foundational framework for this length scale control study. Thermal and structural compliance scenarios were explored under this approach. Numerical results show that the proposed filter efficiently imposes the desired minimum length scale. The optimized designs were also less dependent on the filtering parameters when compared to designs optimized using standard techniques that employ continuous design variables.