PSI - Issue 46

T.J. Gschwandl et al. / Procedia Structural Integrity 46 (2023) 17–23 T.J. Gschwandl et al. / Structural Integrity Procedia 00 (2021) 000–000

21 5

measured with XRD at the indicated positions (P1 – P4) are displayed. All XRD stress profiles exhibit only compressive stresses in a range between -60 MPa and -260 MPa along the 3 mm path from the rail surface. The contour method provides a global stress distribution in the damaged rail, which is qualitatively comparable to literature – see Fig. 2, whereas the X-ray diffraction method offers a high measurement resolution in the near-surface area. Due to the varying length scales, a comparison of the results is limited to two measurement points, as highlighted in Fig. 4. The compressive stresses obtained with CM (-205 MPa) on the surface tend to be higher than the ones measured with XRD (-127 MPa at 0.1 mm depth). In a depth of 2 mm -146.2 and -140.0 MPa are obtained with CM and XRD, respectively. The discrepancies at the surface are attributed to measurement variations in the CM. Due to the finer resolution the X-ray diffraction is more suitable for the near surface area and the contour method is applicable for the overall distribution. Consequently, the best predictions are expected from a combination of both methods.

Fig. 4 (a) Out-of-plane longitudinal stress distribution obtained with CM and close up area of evaluated position 1. (b) XRD stress profile of the four positions on the rail head for a depth up to 3 mm. The comparable values from CM for position 1 are marked with a cross.

3.2. Simulation Fig. 5(a) displays the out-of-plane longitudinal stresses evaluated from the FE simulation after 100 cycles. Fig. 5(b) depicts the stress distribution calculated with the contour method. The numerical and experimental results are qualitatively comparable. However, quantitatively significant deviations occur. Therefore, this comparison is regarded as a proof of concept. Note that for the simulation a virgin rail and simplified contact conditions are used. Hence, neither the residual stresses induced by production – see Fig. 2 – nor real contact conditions are considered.