PSI - Issue 57

5

Elena Sidorov et al. / Procedia Structural Integrity 57 (2024) 316–326 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

320

Fig. 3. Technical contact: (a) surface roughness at stressless state, (b) plastic deformation of surface peaks under loading, (c) wheel load balanced by weld pressure p and contact pressure

After the non-destructive investigations, the test girders were cut into pieces to evaluate the contact conditions underneath the crane rail. Figure 4 shows typicalmicrosections of two test girders. It can be recognized that technical contact in vicinity of the rail welds seems to be achieved. This finding corresponds with the results of the aforementioned feeler gauge measurements. Greater gaps within the contact are visible in the non-fused region. The gaps in Figure 4b are slightly more pronounced in comparison with Figure 4a and are caused by irregularities (local round indents) of the rolled flange surface that (if they exist) typically concentrate above the girder’s web depending on the rolling process. Such irregularities are unavoidable up to a certain extent, but their depth and number are controlled by product standards (EN 10163-3, 2004).

Fig. 4. Microsections of rail fastening. Description of test girders: flat rail: 50  30 mm; crane runway beam: HEA 280, L = 3.50 m; chain intermittent fillet welds: a = 5 mm, h = 50 mm, g = 250 mm.

In the following, the gap size of the rail-flange contact is estimated by means of the surface roughness. The surface roughness is commonly described by two values: Rz and Ra (EN ISO 4287, 1998). For rolled surfaces, Rennert et al. (2012) specify a range of Rz  200 µm according to DIN 4768 (1990). Rz describes the average of the tallest peak to the depth of the deepest valley from each subsection of a surface measurement. In Figure 5a, a measurement with five subsections is considered. As only the most extreme instances of each surface subsection are taken into account, Rz can be used to estimate the local notch effect of a rolled surface, but it is not appropriate to describe the global properties of the technical contact. The roughness value Ra expresses the arithmetic roughness average of a whole surface. In detail, it averages the magnitudes of all profile height deviations from the mean line of a roughness profile. Therefore, this value seems to be able to quantify the technical contact in this paper. Extreme peaks and valleys (outliers) have a much smaller influence than for Rz. A value of Ra  50 µm corresponds with the aforementioned range of Rz (DIN 4768-1, Beiblatt 1). For two surfaces with identical roughness profiles, a maximum gap size of 2 ∙ Ra seems theoretically possible. Deviating from this theoretical value, the initial gap size between the surfaces of rail and flange is set to 1 ∙ Ra for the following reasons in this paper: (i) The value Ra neglects the direction (peak versus valley) of the profile height deviations. The valley depths, that also contribute to Ra, are considered to be less detrimental than the peak heights