Optimized printing orientation determination for improving the stiffness of as-printed handles

Purpose This study aims to improve the stiffness of as-printed handles by finding appropriate printing orientations. Design/methodology/approach First, a series of benchmark handles is designed using Taguchi method. Then, for each uniformly sampled printing orientation, every benchmark handle is sliced and undergoes stiffness evaluation (i.e. displacement and mean stress) by using finite element analysis (FEA). This generates a substantial batch of handle-orientation-stiffness samples. With the data, an effective stiffness-prediction network is developed based on the artificial neural network. Finally, using the developed network, the particle swarm optimization is adapted to determine the optimized printing orientation for each input handle, aiming to improve its stiffness. Findings Compared with the common slicing software, the printing orientations proposed in this study, based on FEA, result in varying degrees of improvement in stiffness for four handles. Specifically, the displacement and mean stress are reduced by 16.86% and 18.14% on average. The experiments show that the approach has the potential to effectively improve the stiffness of a handle. Originality/value Although the anisotropic property in mechanics is unavoidable and difficult to formally describe in 3D printing, the proposed approach can effectively characterize the relationship between the stiffness and the printing orientation for each handle. And, it also can determine an optimized printing orientation for each handle to enhance its stiffness after printing.