An intelligent algorithm for topology optimization in additive manufacturing

Additive manufacturing is a popular process due to advantages such as reduced tooling costs, short time to market, creative freedom for designers, reduced weight, and component consolidation, to name a few. However, there are issues such as component support structure that merit further investigation due to their effect on costs and quality of the fabricated components. In this paper, we propose a topology optimization (TO) algorithm based on solid isotropic material with penalization (SIMP) method. An intelligent combination of data clustering and neural networks enhances sensitivity analysis. The method is able to generate a support-free part design with desirable compliance. Two benchmark problems, short cantilever, and MBB beams are solved by the conventional and the proposed method. The results are validated through experimental work that includes fabricating the beams by fused deposition modeling (FDM) process. The experimental results display promising savings in material usage and manufacturing time.