PSI - Issue 75

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

284

9

Underneath the welds (highlighted in grey), the numerical strains show a steep gradient. In contrast to the FE model that assumes a fillet weld with a = 5 mm and without penetration, the test specimens showed a greater weld size with significant penetration as shown in Tab. 1, Therefore, the observed deviation between the FE results and test data is explainable. Despite these quantitative differences, Figure 8 indicates that the greatest portion of load tends to be directly transmitted from the rail into the flange. The small gap of the FE model, that simulates the technical contact, obviously closes under the loading. Only a small amount of the load is transferred by the rail welds. The flange strains in the weld region are significantly smaller as those underneath the rail. The difference of the flange strains in the weld region due to the different rail weld configuration in Figure 8a and Figure 8b is not significant. As the flange strains, in vicinity of the contact region of rail and flange, are considered as an appropriate indicator for the nominal stress in the rail welds, the nominal stress will be independent of the rail weld configuration under comparable boundary conditions.

Fig. 8. Experimentally and numerically determined vertical flange strains at a depth of 0,33 t f under the flange’s top surface when the flange is only supported by the web in case of a) chain intermittent rail welds (specimen type 1) and b) staggered intermittent rail welds (specimen type 2). Note: The grey highlighted zones represent the flange region underneath a rail weld.