PSI - Issue 75

Elena Sidorov et al. / Procedia Structural Integrity 75 (2025) 276–288 Elena Sidorov et al. / Structural Integrity Procedia (2025)

285

10

5.3. Top flange uniformly supported at the bottom surface Figure 9 compares the experimental and the numerical flange strains of the test specimen type 3 with chain intermittent rails welds, whose flange is supported at its whole bottom surface. The small strains underneath the rail centre, that are calculated by the FE model, indicate that the gap under the rail does not close completely in the numerical calculation. This is a direct consequence of the restrained flange bending. The experimental data also show a comparable tendency to lower strains below the rail centre. In contrast to Figure 8, the flange strains underneath the rail welds (highlighted in grey) are tendentially greater than those underneath the rail centre. Obviously, the transverse stiffening of the flange results in a relatively higher stressing of the rail welds. If only the FE strains in the weld region are compared, the strains of Figure 9 are 80 % higher than those of Figure 8. As a consequence, the applicability of Eq. (1) and Eq. (2) to calculate the nominal stress of intermittent rail welds in the case of transversely stiffened flanges seems to be doubtful the view of the investigated small-scale specimens.

Fig. 9. Experimentally and numerically determined vertical flange strains of the flange at a depth of 0,33 t f under the flange’s top surface when the flange is supported along the whole bottom surface in case of chain intermittent rail welds (specimen type 3). Note: The grey highlighted zones represent the flange region underneath a rail weld. 5.4. Recommended value of gap size The scatter between the numerical and experimental flange strains in Figure 8 and 9 is caused by the randomly varying geometry of the contact surfaces in the tested small-scale specimens. Nevertheless, the implementation of a wedge-shaped gap of size g con = 50 µm under the rail in the FE model appears to simulate the technical contact between the rail and the top flange with an acceptable accuracy. The overall picture of the strain distribution pattern seems physically reasonable. It is important to note that the investigations concentrate on the nominal stressing of the rail welds and not on the local notch stresses of the weld root or weld toes. Therefore, a gap size of g con = 50 µm is recommended for the further investigations on the nominal stresses of the rail welds.