PSI - Issue 47

V. Giannella et al. / Procedia Structural Integrity 47 (2023) 892–900 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

893

2

demonstrate their applicability for the newest materials, extending the areas of application, or understanding and overcoming the inherent limitations of the process itself. Concerning the latter, the high temperature gradients induced during welding still represent one of the primary causes of issues that welding engineers have to deal with. When structures are manufactured by welding, non uniform temperature distributions are locally induced inside the material. These highly localized temperature gradients, caused by repeated heating-cooling transients, cause rapid thermal expansions-contractions in the weld (fusion zone) and the surrounding areas (heat affected zone). These deformations are proven to generate plastic deformations and residual stresses around the weldment that, in turn, can affect the fatigue strength and can jeopardize the quality and reliability of welded structures, but can also cause assembly issues. Many research efforts were devoted to the quantification of the fatigue resistance of welded structures, e.g., through the peak stress method (Meneghetti et al. (2017a-b); Giannella et al. (2022)). Other studies (Barsoum et al. (2009); Citarella et al. (2016)) were also dedicated to the fatigue propagation of cracks that can be triggered by welding. Although welding compressive residual stresses can be advantageous, slowing down the propagation of cracks, it is clear that more comprehension is required in order to satisfy the product requirements, e.g. to guarantee the structural safety and/or to avoid assembly issues (mitigating distortions). Many researchers focused their attention on the development of models for the prediction and the assessment of residual stresses. In general, it is rather unlikely to accurately obtain all the information about distribution of residual stresses in welded structures by means of experimental tests, mainly because of their high costs and measurement inaccuracies. These limitations can be overcome by means of the latest advances of the computational approaches ( Perić et al. ( 2014, 2019a-b)), which have been recently proposed with the aim of simulating the welding processes, quantifying the residual stresses and distortions, suggesting improvements of welding sequences or understanding the impact of welding parameters (pass velocity, energy, heat source distribution, etc.) on the residual displacement and stress fields. These numerical models, very often based on the Finite Element Method (FEM), typically involve transient elastic-plastic thermal-stress analyses of the welded joints. These numerical models can be very complicated due to the coexistence of high thermal gradients, material non-linearity, temperature-dependent properties, phase transformations. Therefore, some efforts were made with the aim of reducing the computational requirements for these numerical approaches, i.e. suggesting the usage of temperature-independent material properties (Bhatti et al. (2015); Sepe et al. (2021a)), performing parametric studies to have less computational expensive analyses (Sepe et al. (2015)) or proposing simple relationships between material parameters and residual stresses (Armentani et al. (2007)). Besides, it is a common practice to validate these advanced numerical approaches through cross-comparisons with the experimental evidence (Armentani et al. (2014)), though that many relevant data are hard to be retrieved experimentally. The investigation included within this document relies in this last framework. This document reports a numerical/experimental comparison of fields of temperature generated during welding of two low-carbon steel plates to obtain a two-passes butt-welded joint. The main goal of this investigation was to simulate the welding process; this was achieved by means of a transient uncoupled thermal-stress FEM analysis. Such numerical analysis leveraged on the innovative usage of th e “birth and death” technique, as p roposed by Armentani et al. (2007) and Sepe et al. (2017). The implementation of this advanced technique was recently refined and published by Sepe et al. (2021a-b c), hence demonstrating the performances of this technique in simulating welding processes. A similar implementation was developed here with reference to the experimental test presented by Murugan et al. (1998), used for validation. Nomenclature Welding current L bead Length of the weld bead Q w Energy supplied to the weld bead t weld Welding time for each pass Welding voltage v Welding speed η Welding efficiency