PSI - Issue 39

Riccardo Caivano et al. / Procedia Structural Integrity 39 (2022) 81–88 Author name / Structural Integrity Procedia 00 (2019) 000–000

82 2

1. Introduction In the last few decades, the incredible developments and improvements of additive manufacturing (AM) processes, especially for metal alloys, permitted their introduction in the production framework, becoming key technologies for many industries.[1–4]. More important, AM changed radically the way structural components are thought, designed and produced with respect to traditional processes [5]. As a matter of fact, AM allows creating very complex and convoluted shapes with few manufacturing constraints, such as the optimised geometries provided by the topology optimisation (TO) algorithms [6–8], and even more complex features such as lattice structures and metamaterial cells [9,10]. Considering TO algorithms, they consent to find the best material distribution with the aim of maximising the structural stiffness or minimising the mass employment [11–13]. However, to exploit the TO as a design tool it is crucial to include constraints on the final structure depending on its final application. For example, in the case of components undergoing quasi-static loads, a limit on the maximum von Mises stress can be imposed [14]. Otherwise, in the case of critical compression loads, a buckling constraint may be included [15]. Recently, the inclusion of fatigue driven constraints within the TO framework has been addressed by the scientific community [13-24]. With this aim, classical fatigue theories based on cumulative damage or similar models are used and this solution is effective for many industrial problems. Anyway, AM processes, such as Electron Beam Melting (EBM) and Selective Laser Sintering (SLS), produce final parts affected by a non-negligible defect population, even if optimised process parameters are set. [28–31]. Pores, Lack of Fusions, and clusters of pores are almost always present in AM final part and they are the major responsible for the fatigue response of the component. Therefore, different and more appropriate theories, such as that of Murakami [32], must be employed to safely design AM parts in the fatigue regime. For this reason, recently the Authors in [33] developed an algorithm, named TopFat, able to include the defect population and its effect on the fatigue response in the TO, using the Murakami fatigue model. Nevertheless, in [33] only bidimensional test cases to highlight the effectiveness of the method are analysed. In the present paper, two main steps forward with respect to [33] are carried out. First, the TopFat algorithm, initially written as a proprietary code, is extended to Hypermesh commercial software, enabling a vastity of applications for industrial designers. Secondly, the TopFat criterion within the Hypermesh framework is used to safely re-design a real component for the aerospace industry to be produced by AM processes and therefore affected by a defect population. Overall, this study aims at providing a guideline for safely designing real AM parts considering the presence of defects by means of commercial TO software. 2. Method The Murakami model [32] states that the fatigue limit in the presence of defects can be expressed as a limit over the first principal alternate stress. In other words, the first principal alternate stress is linked to the positive traction load that a defect endures in the fatigue regime and may make a crack nucleate and propagate from it. Eq.(1), already employed in the TopFat formulation [33], shows the relationship between the limit of the first principal alternate stress ̄ and the material and load properties. ̄ = 1 ∙ ( + 120) �√ � 1 6 ∙ � 1 − 2 � 0 . 226+ ∙10 −4 (1) 1 is a constant parameter depending on the defect location, is the Vickers hardness of the material, is the stress ratio which is equal to the ratio between the minimum and the maximum stress during the cyclic load history and √ is the square root of the area of the defect measured in a plane perpendicular to the maximum applied stress. As reported in Eq.(1), the information about the defect population, namely √ , within the material are required to properly calculate the first principal alternate stress limit. Nevertheless, this information is unavailable before the effective component realization. Only after the final part is obtained, it is possible to access the defect population, for example using non-destructive techniques. To overcome this issue and evaluate in the design phase the fatigue limit,