PSI - Issue 34

Sanne van den Boom et al. / Procedia Structural Integrity 34 (2021) 87–92 Sanne van den Boom et al. / Structural Integrity Procedia 00 (2021) 000–000

88

2

Area with prescribed downward displacement

Symmetry plane

Simply supported at the holes

Supports at the slanted surface

Fig. 1: Original design, making use of symmetry, and boundary conditions for the ladder step.

We predict the mechanical behavior for the following three designs: 1. Design A - uniform infill: The original FFF design with a uniform infill distribution, as it would be sliced in common slicer software; 2. Design B - optimized infill: A design with an optimized infill distribution which has roughly the same total weight as design A - uniform infill ; 3. Design C - optimized design: A solid topology optimized redesign of the component, which again has roughly the same weight as previous designs. Experimental testing is done on a total of 9 specimens, 3 of each type. Each specimen is loaded until failure in a 25 ton bench using a rubber patch. Digital Image Correlation (DIC) measurements are done to measure the deformation pattern. The goals of these experiments are fourfold: (i) to gain insight in the mechanical behavior of mechanically loaded fused filament fabricated PET-G parts; (ii) to gain preliminary insight in the uncertainties of the mechanical properties associated to the 3-D printing process of complex components; (iii) to validate the numerical simulations that were developed for the printed material and homogenized material properties; (iv) to verify the improved perfor mance of the component with an optimized infill distribution, and of the topology optimized design. We show that FFF of PET-G is suitable for the production of highly loaded components. We also demonstrate that our models predict the mechanical behavior in su ffi cient detail, and that optimization techniques can be used to improve the performance. This paper is structured as follows: in Section 2, the basic design and boundary conditions are described. Section 3 describes the numerical method, optimization procedures, and optimized designs. Section 4 describes the manufac turing of the specimens and the experimental setup. In Section 5, both the numerical and experimental results are discussed, and the di ff erent designs are compared. Finally, Section 6 draws conclusions. The original design of the ladder step, as made by the ECAM, is illustrated in Figure 1. Here, only half of the ladder step is shown, where the blue arrows indicate the symmetry plane. The ladder step is designed to hang from ropes and rest against a slanted surface. In Figure 1, two holes are visible where the ladder steps are suspended from the ropes. The ladder step is simply supported along the bottom edge of the holes, and the parts that are supported by the slanted surface are fixed orthogonal to the surface on which the ladder step rests and in the sideways direction. In the downward direction, the part is free to slide frictionless down the slanted surface. The surface area on the top is constrained to a reference point by using kinematic coupling with uniform weighing and a downward displacement is prescribed on the reference point. Symmetry conditions are prescribed on the symmetry plane. 2. Designs and boundary conditions All numerical simulations are performed in Abaqus 2019 HF2, using the Micromechanics plugin v1.15 (Simulia, 2017) for homogenization of material and infill properties, and Tosca for the optimization. Material and infill properties Following Somireddy et al. (2018), e ff ective orthotropic material properties are obtained for the 3-D printed material in solid sections such as the wall layers. The geometry of the Representative Volume Element (RVE) (Figure 2b) that is used to this end is based on micrographs (Creusen, 2019) and represents the voids that appear in between roads in 3-D printing. The constitutive material in this RVE is PET-G with material properties as summarized in Figure 2a. For the homogenization of the infill properties a similar approach is chosen; 3. Numerical methods and design optimization