PSI - Issue 54

Jenny Köckritz et al. / Procedia Structural Integrity 54 (2024) 423–430 J. Köckritz / Structural Integrity Procedia 00 (2019) 000 – 000

424

2

Keywords: Weld Assessment, Fatigue, Aluminum, Multiaxial Load, Load Collective, Effective Notch Method, Hot Spot Stress Method

1. Introduction In the present use case, a cargo bicycle frame welded from EN AW 7020 tubes was assessed according to its fatigue behavior during the design phase. Fig. 1 (a) shows the cargo bicycle “CityPed®” . The maximum load is 190 kg , while its frame weighs 4.2 kg . It was developed for the project “SteigtUM” , which aims to implement and test an alternative mobility strategy for small city environments (SteigtUM, 2023). Similar to other cargo bicycles, the frames are produced in a small batch. Therefore, extensive fatigue testing is not possible. A newly introduced cargo bicycle must pass fatigue tests according to DIN 79010 (2020) for market introduction, so developers require an applicable and relatively reliable weld assessment method to achieve a high lifetime and lightweight design. The fatigue assessment of welded structures is the topic of several books and international standards such as the CSA W59.2 (2018), EN 1999-1 (2007) or the International Institute for Welding (IIW) recommendations by Hobbacher (2016) and numerous research papers (Möller et al., 2021; Mei et al., 2021; Bolchoun et al., 2014). The investigations aim to provide a reliable method for the service life assessment of welded structures throughout design phase and use. Recognized standard methods are normal stress method, the effective notch approach, hot spot stress approach and fracture mechanics approach (Hobbacher, 2016). Most of the applied methods are stress-based assessments. However, strain-based assessments are successfully applied for example in Rudorffer et al. (2021). The effective notch (EN) method includes both structural and notch stress effect by modelling the whole weld with an idealized effective toe and root radius, see Fig. 2 (b). The weld assessment is performed at the maximum stress concentration with so-called FAT classes (Hobbacher, 2016), which are available for aluminum and steel and correspond to a survival certainty of 97.5%. The EN method allows assessment of toe and root failure and is not dependent on the use case geometry, but it requires a high modelling and computing effort. Less modelling effort is needed for the hot spot stress (HSS) method, which can be applied for solid and shell modelling. In the HSS method, only structural stress peaks are evaluated, while notch stresses are disregarded. The structural stress peak is extrapolated from defined points perpendicular to the weld, see Fig. 2 (a). The extrapolated structural stress peak is evaluated with FAT classes dependent on specific weld details. If significant shear stresses exist, an additional assessment of the shear stresses with shear FAT classes can be performed. Multiaxial stresses can be considered for assessment (Bolchard et al., 2014). The HSS method is limited to the assessment of the weld toe (Hobbacher, 2016). In a variation of the HSS method, HSS are calculated from grid line forces perpendicular to the weld line and grid line moments along the modelled weld line (Fermer et al., 1998). Furthermore, two S-N curves are utilized for the fatigue assessment. A differentiation is made between welded material under primarily bending loading (bending S N curve) and primarily normal loading (membrane S-N curve). Other research such as Mei et al. (2021) also conclude this to be a sensible approach for multiaxial loaded welds. The “Seam Weld Fatigue” method (SWF) utilizes this approach and is implemented in the FEA software Altair®, where the fatigue life is assessed based on a S-N curve that is interpolated between bending S-N curve and membrane S-N curve. The interpolation is performed for each finite element according to its loading. Until a threshold of present bending moment, the bending ratio r , the membrane SN curve is used to calculate the expected lifetime under this loading. If a higher proportion of bending in an element exists at the weld, the two SN curves are interpolated, cp. Fermer et al. (1998) and Altair (2023). In Hobbacher (2016), most FAT classes are defined with a comparatively steep fatigue exponent of =3 for aluminum. Other standards apply fatigue exponents for welded aluminum that vary from =3…7 (EN 1999-1, 2007) to =8.4 (CSA W59.2, 2018). Experimental results conclude less steep fatigue exponents, for example in Mensinger et al. (2018) =5 for cross joints. Comparisons with experimental tests are usually conducted on simple butt joints or cross joints with not load carrying or load carrying (Ye et al, 2008) welds. Investigations on more complex geometries are less commonly conducted, for example in Möller et al. (2021) for aluminum or in Brahami et al. (2019) for steel. Some researchers conclude conservative results when applying the IIW recommendations (Bolchoun et al., 2014) while others apply the IIW standards with a good correlation to experiments, especially on standard joints (Stötzel, 2005; Möller et al., 2021). The HSS method was also shown to lead to non-conservative results in Ye et al. (2008). The IIW recommendations have limited applicability when weld quality, weld thickness or loading situations vary