PSI - Issue 54

Martin Matušů et al. / Procedia Structural Integrity 54 (2024) 135 – 142 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

136

2

powder. Two different manufacturers (GE and ECKART) provided the powder of AlSi10Mg aluminium alloy. The printing parameters remained consistent throughout the testing campaign. The primary goal was to investigate if the fatigue behaviour changes with multiple reuses of the powder left after previous builds. The printing procedure is complex, but a brief explanation is necessary. CL 31AL powder from GE or from ECKART is used, while both powder manufacturers claim the same distribution of grain size of 15-53 µ m. The chamber is filled with powder, and the laser melts the objects [1]. Firstly, the contour of the specimen is melted (depth of 0.3mm), followed by the skin (depth of 2.5mm) which is melted using zig-zag pattern in each layer. The core (the rest of the print in depths) is melted by laser beam of a higher energy input using a random criss-cross pattern every second layer. The unused powder is sieved through a sieve with the mesh of 60  m. However, the question at hand is whether the powder is affected by the printing process, despite being sifted. In this paper, we will assess the fatigue parameters of three different platforms of samples, each subjected to three different heat treatments. The aim is to determine whether the effects of recyclability can be mitigated through heat treatment. All samples have the same geometry, and the testing ratio was kept consistent to ensure no additional influences are introduced into the equation. Some of the results were already published in another paper of a different focus [2]. Nomenclature Θ Stabilized temperature increase R o Initial temperature rate R γ Closing temperature rate. N Number of cycles N f Number of cycles till failure k logarithmic slope of S-N curve This study focuses on examining the fatigue behaviour of the AlSi10Mg aluminium alloy produced through the Laser Power Bed Fusion (LPBF) technology. Hourglass-shaped fatigue specimens with a critical cross-section diameter of 9 mm and a transition fillet radius of 60 mm were selected for the fatigue experiments (refer to Figure 1b). The fatigue specimens featured a machined head with an M18x1 thread, allowing them to be securely fastened to the Amsler HFP 422 resonant pulsator equipped with a 100 kN load cell. The loading was applied exclusively in tension by the pulsator, maintaining the stress ratio R of 0.1. All specimens were printed vertically within a single build cycle using the Concept Laser M2 printer. The printing parameters for both the skin and the core of the specimens were kept consistent. The skin, which is an integral part of the outer structure and is printed in every layer, can be considered as a contour. This approach significantly enhances the surface quality of the printed specimens. The roughness measurements were performed using the MarSurf LD 120 drive unit for fatigue specimens, yielding Ra = 3.09 μm and Rz = 24.28 μm. The core of the specimen was printed using parameters that differed from those used for the skin, as indicated in Table 1. Four separate batches of samples were printed over a span of 2 years, each batch focusing on investigating the fatigue behaviour of different heat treatments and different surface treatments. The main topic discussed in this paper is the comparison of these four different platforms. Each platform was fully utilized to print samples for fatigue tests or tensile tests, depending on the specific topic being examined. As an example, Figure 1 a) illustrates a printing platform after completion, showcasing 58 printed samples, including 14 tensile specimens. This paper also addresses three distinct heat treatments, as indicated in Table 2. However, the discussion concerning the differences in fatigue strength or tensile properties among these heat treatments will be presented in a separate paper prepared for the Fatigue Design 2023 conference. 2. Fatigue strength analysis 2.1. Description of specimens and experimental setup