PSI - Issue 57

Mathias Euler et al. / Procedia Structural Integrity 57 (2024) 298–306 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

301

4

The effective loaded length eff is derived from a calculation model consisting of an elastic shell with load distributing top chord ( Bernoulli beam). According to Section 5.7.2 of EN 1993-6 (2006), the peak magnitude of the local shear stress  ||local at the top edge of the web, that occur slightly shifted to the loading section, can be assumed by  0.2 min  ⊥ .

min

/  ⊥ = p t

(1)

max w

max eff / = p F

(2)

3 rf w 3.25 / = I t

(3)

eff

The stresses  || ,  ⊥ and  || in the web with thickness t w can be transformed by Eq. (4) into corresponding weld stresses of the flange-to-web connection (double fillet weld) in order to determine the relevant stress ranges for the fatigue verification:

w w / (2 )   ⊥ ⊥ = t a

w|| ||   =

(4)

w|| || w / (2 )   = t a

The localstress ranges   w ⊥ of the welds are checked with a fatigue resistance   C = 36 N/mm² for welded flange to-web connections with partialpenetration and with   C = 71 N/mm² for full penetration accordingto EN 1993-1-9 (2005). These fatigue resistances are identical with those of corresponding welded T-joints under cyclic tensile loading. Thus, the current classification neglects that the details are exclusively exposed to compressive stresses. The local shear stresses in the welds are superposed with the global shear stresses. The stress ranges of the total shear stresses  w|| in the weld are verified using a fatigue resistance   C = 80 N/mm² for both, partial and full penetration welded connections. Finally, the interaction of the stress components is taken into account by EN 1993 1-9 (2005) using Miner ’s rule that adds the fatigue damages due to   w ⊥ and   w|| . Up to now EN 1993-6 (2006) has not provided particular guidance on how to determine the local stresses in rail welds due to the wheel load introduction. Moreover, EN 1993-1-9 (2005) lacks a detail classification of continuous rail welds. Therefore, a fatigue verification of wheel-loaded continuous rail welds is practically impossible today. 3. Experimental investigations Two cyclic test series had been performed to determine the fatigue resistance of continuous rail welds and partial penetration welded flange-to-web connections (Kuhlmann et al., 2015, 2016). These constructional details are denoted as detail #1 and #2 in the following. In the first test series, ten test girders for detail #1 and six test girders for detail #2 were tested under a travelling wheel load at the Material testing institute (MPA) of the University of Stuttgart as shown in Figure 3a. These tests were complemented by 12 girders for detail #1 and seven girders for detail #2 that were tested under a stationarily pulsating wheel load. The test girders for detail #1 consisted of a hot-rolled I section HEA280 of structural steel grade S355J2+AR according to EN 10025-2 (2004) with a hot-rolled flat 50  30 mm according to EN 10058 (2005) of structural steel grade S355J2+AR that served as crane rail. The crane rail was fastened to the top flange by single-run fillet welds (fully mechanized submerged arc welding process). The plate girders, that were used for the investigation of detail #2, were made of structural steel grade S355J2+N and had an actualweb thickness of 9.8 mm. Flat material50  30 mm of steel grade S355J2+AR was used as crane rail that was not mechanically fastened to the top flange. Both tested constructional details represent partial penetration welded connections as exemplified by Figure 3b. Thus, it is expected that the fatigue failure is initiated at the weld root under the wheel loading with subsequent crack growth in the weld. 2.2. Continuous rail welds