Topology optimization for additive manufacturing: a front propagation-based approach

Abstract

This thesis a method is described to include manufacturing constraints of the additive manufacturing process into topology optimization. Topology optimization often results in complex structures, which can only be produced by additive manufacturing. Although additive manufacturing has much fewer design restrictions compared to conventional manufacturing methods, it has its own limitations. The most prominent limitation is the overhang constraint, which causes the need for support structures. In this thesis front propagation is utilized to crudely mimic the printing process. Making use of existing numerical methods for front propagation, a numerically efficient algorithm is produced that can identify parts of the structure that do not adhere to the overhang constraint, during the topology optimization process. Since the sensitivities of the algorithm are available, it can be used in the gradient based topology optimization to ensure printability of the final part. Due to the continuous description of the constraint, it can be applied to unstructured meshes and for variable overhang angles. Furthermore, it is shown that the constraint can be parallelized which is demonstrated on large scale 3D problems.