Partitioning a topology-optimized structure into additively manufacturable parts using a feature-mapping approach: a novel decomposition optimization method

Structures re-designed with topology optimization (TO) for additive manufacturing (AM) are restricted to the maximum build size of an AM machine. This work proposes a novel mathematical method that sequentially solves two subproblems, namely the topology and part decomposition problem, to obtain the optimal material layout and partitioning of large-scale structures into additively manufacturable components. Presented as decomposition optimization, the latter subproblem adopts a feature-mapping approach to introduce and optimize the layout of partitioning profiles, referred to as partitioning rectangles, in the design domain. The size of components, thickness, and material properties of component interfaces are directly introduced in the problem formulation through the geometric description of a partitioning rectangle. Further, to ensure each component in the structure is additively manufacturable, an incomplete decomposition fraction constraint is introduced to limit the amount of structural material that exists outside of a partitioning rectangle. The proposed methodology is tested with two numerical examples to demonstrate its capabilities.