PSI - Issue 34

Harry O. Psihoyos et al. / Procedia Structural Integrity 34 (2021) 253–258 Harry O.Psihoyos et al. / Structural Integrity Procedia 00 (2021) 000 – 000

254

2

1. Introduction Selective Laser Melting (SLM) is an Additive Manufacturing (AM) process which enables the fabrication of components with complex geometry directly from 3D models. These components have been usua lly designed employing topology optimization techniques and would be very difficult to be manufactured with conventiona l manufacturing methods (Jiménez et a l., 2019). Despite its benefits and the integration of advanced design techniques, SLM has some inherent disadvantages that prevent its wide adoption from the industry. The most important disadvantages include the formation of process-induced defects that have a significant detrimenta l influence on mechanica l properties and especia lly on the fatigue behavior of SLM parts, as indicated by Sanaei and Fa temi (2020) and Yadollahi andShamsaei (2017). Defects in SLM process can either result from insufficient combination of process parameters, the most important being laser power, scanning speed, layer thickness and hatch spacing or by process instabilities of stochastic nature, as described by Snow et a l. (2020). The main defects that can be formed in SLM parts include surface roughness, lack-of-fusion, keyhole, ba llingand gas porosities (Sanaei and Fatemi, 2020). These defects act as loca l stress ra isers on SLM parts, from which cracks can initiate and propagate. Despite the process parameters optimization attempts and post-processing methods, such as surface machining and heat treatment , a iming to the elimination of defects, a certa in amount of defects still exists and cracks can initiate from their site, as it has been shown in the works of Gong et a l. (2015) and Hu et a l. (2020). In order to move towards a wide successful adoption of SLM process for part fabrication, the effect of defects must be considered in design and ana lysis of parts. For this reason, reliable methods for the defect-based prediction of fatigue lives must be developed. Yadollahi et a l. (2018) and (2020) proposed a plasticity-induced crack closure model for fatigue life prediction of AM Inconel 718 and 17 -4 PH SS meta ls, respectively. Characteristics of interna l defects, such as size, shape and orientation, as well as surface roughness and heat treatment post-processing were considered and the estimation of fatigue life were fitted with experimenta l data quite well. More recently, Sanaei and Fa temi (2021) presented an ana lytica l framework for the prediction of fatigue life of AM Ti-6Al-4V and 17-4 PH SS meta l specimens implementing the Paris equation and the Hartman-Schijve variant of NASGRO equation. In their work, they considered ava ilable sma ll fatigue crack growth properties of AM meta llic a lloys, as long fatigue crack growth may lead to non-conservative predictions of fa tigue lives of AMmetals. In the present work, a modelling framework for the estimation of fatigue life of SLM AM Ti-6Al-4V a lloy is presented. Although, the current ana lysis is limited to prediction of fa tigue life of AM Ti-6Al-4V it can be expanded for other AM meta l a lloys as well. The ma in assumption of the present model is that the fatigue fa ilure is controlled by the largest defect which serves as the initia l crack that propagates until its critica l size. In contrast to the aforementioned litera ture studies, the susceptible areas on the which the critica l defects can be formed were predicted performing therma l history ana lysis in ANSYS 2020R2 Additive Suite and applying well known criteria for the formation of ma in defects. The information about the susceptible areas determined the model for fatigue life estimation in AFGROW fracture mechanics software. The results of current ana lysis were compared with ava ilable experimental litera ture data and presented good correlation further verifying the modellingprocedure. 2. Experimental data The modelling results will be compared with ava ilable experimenta l fatigue data of SLM Ti-6Al-4V for model va lidation purposes. Du et a l. (2021) investiga ted the effect of process parameters on the high-cycle (HCF) and very-high-cycle fatigue (VHCF) of SLM Ti-6Al-4V. Nine groups of specimens were fabricated with different combina tion of process parameters in HBD-200 SLM machine. The amount of resulting porosities on each group was different as a result of the process parameters combination. The specimens of Group 4 were considered in this study and present the third level of minimum amount of porosity which was about 1.2% of the nomina l density. The process parameters of Group 4 are presented in the Table 1, after Du et a l. (2021). The specimens were built vertica lly and the scan strategy for each group of specimens was a combination inner hatch/filling and a single contour scan line, rotated 60˚ on each subsequent layer. All the groups of specimens were heat annea led a t 600˚C for 2 h after their fabrication to relieve the developed residua l stresses and they were polished to remove the surface defects.