PSI - Issue 68

Shahriar Afkhami et al. / Procedia Structural Integrity 68 (2025) 929–935 S. Afkhami et al. / Structural Integrity Procedia 00 (2025) 000–000

931

3

were carried out to indicate the primary culprit behind the failure of each design and its corresponding raw metal used in the fabrication. 2. 2. Materials and experimental procedure Topology optimization of the original bicycle crank arm and its validation (Fig. 1) was carried out using Altair Inspire. As indicated in Fig. 1, material stiffness was considered the optimization domain (constraint), and the optimized component was manufactured along its major axis to minimize the required support structures. Three raw powders were considered in this research to be used for L-PBF as the selected AM technique to fabricate the components: stainless steel 316L, titanium alloy Ti64, and aluminum alloy Al5X1. All these powders were processed via L-PBF using the AM parameters recommended (optimum) by their vendor. The powders’ specifications and parameters are available in [6]. All the tested components were designed and manufactured by the industrial contributor of this research (Etteplan), while the subsequent tests and analyses were done at LUT University. Optical microscopy was utilized using an Olympus microscope to evaluate the AM metals’ microstructural features (low magnification), defect content, and defect distribution; quantitative defect content analysis was done based on the optical micrographs using ImageJ software. Also, a SU3500 (Hi-Tech Instrument) scanning electron microscope (SEM) was used for high-magnification images. In addition, the surface roughness of the manufactured components was measured using a KEYENCE VR-3200 3D measuring microscope.

Fig. 1. Schematics of the original design (left), and the 3D view of the optimized design of the component (right) (linear dimensions in mm).

Next, the manufactured components were subjected to cyclic loads until failure for the fatigue tests ( R = 0.01); the direction of the applied loads and fixed spot of the components in the test rig setup are indicated by red arrows in Fig. 1. It should be noted that components made of 316L and Al5X1 were shot blasted after AM, while Ti64 components were tested in their as-built condition. Finally, using the SEM, a fractography analysis was carried out on selected samples that broke at the end of the fatigue tests. It is worth mentioning that the maximum stress concentration factor from each optimized design (for each metal) was calculated using the finite element analysis (FEA) and FEMAP software (Nastaran solver) to normalize the fatigue test results and their comparisons between various raw metals (316L, Ti64, and Al5X1). Due to the different strength levels of the raw metals [ σ y(Al5X1) < σ y(316L) < σ y(Ti64) ], the wall thickness and specific corner angles were different in the optimized designs for each metal powder. Also, the FEA was used to turn the fatigue test data from force values into corrected stress values to make a stress-based comparison possible between various designs and materials. 3. Results and discussion The optical micrographs of the samples and their SEM images, as shown in Figs. 1 (left) and (right), respectively, exhibited microstructural features similar to the noted one for these additively manufactured metals in the literature: for 316L, an austenitic hierarchical microstructure with epitaxial growth between deposited layers and cellular subgrain structures [7]; for Al5X1: a fine-grained and difficult to etch microstructure comprising of an alpha aluminum matrix (large pores in the etched sample (right image) are caused by the etching process and are not a part of the microstructure) [8]; and for Ti64: arrays of α and β titanium intertwined lamellas forming a basket-weave microstructure [9,10]. Defect analyses based on the defect distribution maps, Fig. 2 (middle), show that Al5X1 had a