Mechanical analysis and optimized performance of G-Code driven material extrusion components

Abstract

In this work an end-to-end optimization procedure to maximize the mechanical performance of Additive Manufacturing (AM) components is presented. Material Extrusion (ME) is the selected demonstrative AM technology, but the approach is applicable to other AM processes where the manufacturing toolpaths of the geometry of the printed component are described in G-Code format. The proposed methodology is integrated into the AM workflow and drives a two-step optimization process in order to select the optimal printing orientation for a user defined case. The G-Code file containing the manufacturing toolpaths is used as input. This approach allows to operate as close as possible to the geometry of the actual component, avoiding the use of the STereoLithography (STL) geometry. A voxelized mesh is built from the G-Code by solving a modified 2.5D Shortest Path Problem (SPP) and high-fidelity Finite Element (FE) simulations are performed with the resulting mesh. A printing pattern-based material model that distinguishes three different zones of the printed component is used. Due to the orthotropic nature of the ME process, the Tsai–Wu failure criterion is applied to obtain the indicators of the mechanical performance of the component. These computed metrics are used to drive an optimization process where a robust criterion based on the Machine Learning (ML) algorithm Anomaly Detection (AD) is applied in order to select the optimal build direction from a prespecified span of orientations. Two test cases and one case study illustrate the performance of the proposed methodology. The results validate the approach against experiments, indicate that the selected optimization criterion is robust against factors alien to the actual physical problem and show that the accuracy of the voxelized method greatly improves the “traditional” STL-based simulations.