Validation of a G-code generator for Additive Manufacturing using Image-based structural mechanics simulations

1. Introduction

Additive Manufacturing (or AM, 3D printing) is a highly interesting technology with many benefits: complex and lightweight geometries, design freedom and customization, innovative material options, rapid prototyping, no-tooling production, reduced material waste, supply chain flexibility, cost-effective low- volume production, on-demand manufacturing, etc.

Ideally, a printed part should be designed completely in the computer, by simulating and optimizing all its relevant properties before manufacturing it. However, the inhomogeneity of the printed parts is a key issue for many AM processes. This leads to varying material properties and makes it difficult to simulate properly the parts’ behavior from the CAD data directly.

We propose that simulating the printing process to model the resulting porosity overcomes this issue.

This work involves CAD, 3D printing, 3D imaging, mechanical testing, simulation of the printing process and simulation of the mechanical properties.

We use a micro-CT scan to understand the porous microstructure of a printed part. Structure-mechanical properties simulated on the scan agree with physical measurements. Based on knowledge of this microstructure, a G-code generator is presented. It converts the CAD data into a 3D model of the part looking just like the scan, i.e., made up of the same type of porous microstructure instead of being a solid piece.

Structural-mechanical properties simulated on this micro-structured model also agree with the physical experiment. The agreement of the experimental and simulated stress-strain curves for the micro-CT scan validates the mechanical simulation.

The agreement of the experimental and simulated stress-strain curves for the model created from CAD and G-Code validates the G-code generator code.

Micro-CT can reveal the porous microstructure resulting from AM processes. With this knowledge, one can introduce the porous microstructure into the geometric model of a printed part and correctly predict its stress-strain curve. 

Part modelled with GeoDict [1] and printed using fused filament fabrication by an Ultimaker 3 printer in polycarbonate. Mechanical tests done by J. Krummenacker, (Leibniz Institute for Composite Materials, IVW). 3D µCT imaging done by F. Schreiber (Fraunhofer Institute for Industrial Mathematics, ITWM). Printing process simulated using the new G-code generator code. Mechanical properties  computed using ITWM’s FeelMath solver [2], part of M2M’s GeoDict [1] software. C. Bauer (ZF Group, formerly M2M) contributed technical help and F. Arnold (M2M) created graphics and improved the slides.

[1] GeoDict simulation software Release 2024, by Math2Market GmbH, Germany, doi.org/10.30423/release.geodict2024
[2] M. Kabel, T. Böhlke, M. Schneider, Efficient fixed point and Newton–Krylov solvers for FFT-based homogenization of elasticity at large déformations,Computational Mechanics 54 (6), 1497-1514