PSI - Issue 79

Andrea Avanzini et al. / Procedia Structural Integrity 79 (2026) 88–96

89

1. Introduction The aim of this work is to evaluate the applicability of the infrared thermographic (IRT) methods for rapid estimation of the fatigue limit of additively manufactured (AM) metal alloys. These materials are increasingly used by industry due to their attractive mechanical properties, often comparable with those achieved with traditional processing routes, combined with the design advantage associated with additive manufacturing. However, the fatigue properties of AM materials are difficult to determine, due to the challenges inherently associated with their production methods. Hardly controllable factors, such as porosity, layer adhesion or surface finish make estimating the fatigue limit both complex and time-consuming. Given the number of parameters that govern the behavior of AM materials, the availability of methods for rapid estimation of fatigue life would be extremely relevant. Over the last decades, infrared thermographic (IRT) methods have emerged as viable alternative to traditional fatigue tests (Luong, 1998) (La Rosa, 2000) potentially allowing fast determination of the fatigue limit, based on investigations on a limited number of specimens. The validity of IRT approaches has been demonstrated for various metal alloys and polymers or composites produced with conventional techniques (Munier et al., 2014), (Liu et al., 2023)(Montesano et al., 2013)(Meneghetti & Quaresimin, 2011)(Li et al., 2021), but limited, and sometime contrasting, findings are available when AM is concerned (Douellou et al., 2019)(Balit et al., 2020)(Matuš ů et al., 2024). Moreover, different thermal indexes, that can be extracted from thermographic imaging, have been proposed for fatigue life estimation, including temperature based (initial rise slope and steady temperature) and energy-based (Q-parameter from cooling curve and Φ parameter) parameters (Wei et al., 2024) and various fitting strategies for the interpretation of thermographic data have been developed (Cura et al., 2005) but no systematic investigation on their applicability to AM alloys has been conducted yet. In the present work, starting from the analysis of methodologies reported in literature for performing this type of investigation by IR thermal camera on metal alloys, we aimed to compare the various methods, verifying their applicability to AM materials on a case study. To this aim, specimens of a Ni-based (IN625) alloy, produced via Laser Powder Bed Fusion (L-PBF) and subjected to different heat treatments (HT), underwent cyclic stepped tests, under thermographic monitoring. In the post-processing phase, different methods were implemented, and the results were first compared as a function of the approach and then assessed against conventional fatigue data from an ongoing conventional study. 2. Methods 2.1. Materials and specimens The material under investigation is a Ni-based alloy, IN625, produced via L-PBF using a Renishaw printing system operating under an argon atmosphere. Details of printing parameters and powder characteristics are available in previous investigations on the same material (Ferraresi et al., 2022). The dumbbell specimens used in the present work were printed vertically, their length was 80 mm and the cross section in their gauge region was 5 × 3 mm. These specimens underwent different heat treatments (HT) in a vacuum furnace, namely Stress Relieving (SR), Direct Aging (DA), Solution+Aging (SA) and Annealing (AN). A detailed description of heating and cooling thermal curves for each treatment can be found in (Ferraresi et al., 2022) together with material characterization. For the purpose of the present work, it suffices to underline that these treatments mainly differed for the maximum temperature reached and for the duration. High temperature treatments, such as SA and AN, were found to modify drastically tensile properties, in particular yield strength and ductility, and consequently fatigue. 2.2. Test protocol and IRT Methods Cyclic tests were carried out using a Rumul Mikrotron test rig, a resonance machine capable of operating at frequencies up to 250 Hz. In the present work the machine, equipped with a 20 kN load cell, operated at around 79 Hz and was controlled in force with a pulsating cycle (R = 0.01). A cyclic load was applied for 30000 cycles in a stepwise manner, progressively increasing the load level (20 MPa per step) until failure occurred. The various phases of a single test block are illustrated in Fig.1(a,b). Initially the tensile (static) mean load was applied, resulting in temperature decrease as a consequence of thermoelastic effect. After this temperature variation was recovered and a stable value