PSI - Issue 57

Gloria Hofmann et al. / Procedia Structural Integrity 57 (2024) 452–460 Hofmann, G.; Bartsch, H.; Kuhlmann, U.; Feldmann, M.

453

2

1. Motivation The designs of many constructions, such as bridges, crane runways and wind turbines are largely determined by the fatigue design. The standard is usually based on the detail catalogue according to (EN 1993-1-9, 2005). This current standard is based on test results dating partially as far back as 1920. The course the revision of the standard now offers the, it is possibility to incorporate new research findings and thus to enable a more economicaldesign. In this paper, both geometric influences and imperfections are considered in more detail. Currently, the detail category for butt-welded joints with plate thicknesses higher than 25 mm must be reduced by the factor k s , where k s is calculated according to the formula (1), with t as the thickness of the plates. k s = (25 ⁄) 0.2 (1) In the upcoming Eurocode 3 Part 1-9 (prEN 1993-1-9, 2023) the reduction has already been somewhat reduced for ground flushed butt-welded joints by lowering the exponent to 0.1. For butt-welded joints without post-weld treatment, however, nothing has changed so far. Based on the existing data from experiments on butt-welded joints, it is difficult to make a valid statement about the influence of the plate thickness for two reasons. On one side, there are many test results, which show a large scatter, on the other side, there are only a few test results for large plate thicknesses. All test data on butt-welded joints taken together, do not show a statistical influence of the thickness on the fatigue strength. Therefore, experiments were carried out on butt-welded joints up to 80 mm thickness, which are described in more detail in the following. Another geometric influence concerns the longitudinal attachments. While the fatigue strength for longitudinal attachments without post-weld treatments must be reduced from detail category (DC) 71 to DC 56 with increasing length, longitudinal attachments with a radius at the end larger than or equal to 150 mm are classified in a better DC of 80, regardless of the length. However, experiments with beveled longitudinal attachments showed no improvement in fatigue strength. Therefore, the question arose whether in fact, the geometric transition moderated by the radius increases the fatigue strength. Alternatively, the additionally required flushing of the weld transition at the end of the attachments and thus the moderation of the weld toe transition could be responsible for the improvement in fatigue strength. Since the radius of 150 mm costs a lot of materialand effort,the influence of the radius was investigated by testing longitudinal attachments with radii of 80 mm and 150 mm. These experiments are also explained in more detail later on. A further pending update of (EN 1993-1-9, 2005) concerns the inclusion of weld quality influences. The fatigue strength of various welded details depends on the size of possible weld imperfections, which are divided into evaluation groups in (EN ISO 5817, 2014) depending on their size. However, these evaluation groups have not been defined on the basis of scientific investigations, and thus a correlation to fatigue strength values has not yet been made. The investigation described here conducted at RWTH Aachen University is intended to remedy the situation and provide new findings on the influence of internal weld imperfections on the fatigue strength of typical welded details. 2. Investigation of the thickness effect of transverse butt-welded joints The large data set on fatigue tests with butt-welded joints was first filtered accordingto various criteria to perform a valid evaluation. In this paper, only flat butt-welded joints welded on both sides are considered, since the root pass of butt-welded joints welded only from one side leads to a sizable notch so these details are classified without further inspection in DC 36. For better comparability, only test results were considered of flat specimens under axial loading, with a stress ratio of R ≥ 0, constant amplitude and cycles between 2 ∙ 10 4 and 5 ∙ 10 6 . Apart from grinding the weld reinforcement flush, no other post-weld treatments such as stress relieved annealing were considered. Since the removal of the weld reinforcement has a major influence on the fatigue resistance and the Eurocode also assigns ground flushed and as welded butt-welded joints to different DC, these are evaluated separately. 2.1. General information