PSI - Issue 23

Pavol Dlhý et al. / Procedia Structural Integrity 23 (2019) 185–190 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

188

4

Fig. 4 Tangential residual stress obtained from FEM simulations correlated with the slit ring method for different distribution types

evaluated from the calculated deflections using the approach based on the curved beam bending theory, see Poduška et al. (2014) and Poduška et al. (2016) for detail s. The evaluated values were compared to the original values of residual stress present in the model of non-slit ring. This comparison was done for three different variants of residual stress distribution – a simple linear distribution, a quadratic distribution and a complex distribution of an irregular shape, see results in Fig. 4. The results show agreement between tangential residual stress distribution in models with results determined by the slit ring method. Additionally, optimization of number of layers and number of rings for such limited axial segment length was done. The number of layers affects the precision of the residual stress determination. It also determines the minimum possible number of rings that need to be cut from the axle segment to successfully carry out the experiment. It was found that 17 layers is a suitable number to properly describe the residual stress distribution and that 22 rings are enough to successfully carry out the experiment. After a 280 mm long segment was cut out from the railway axle, a hole with diameter of 68 mm was drilled into the axle segment. The segment thus prepared was cut into the ring specimens of 9 mm thickness. Half of the rings were modified by removing one layer from inside of the first ring, two layers from the second ring and so on. Second half of the rings were modified in the opposite manner by removing layers from the outside. After the layers had been removed, marks were scribed on each ring to mark the position of the neutral surface (see Fig. 5) and their position was measured. Then, a segment of each ring corresponding approximately to an angle of 30° was cut off (see Fig. 6) by wire electrical discharge machining. After that, tangential residual stress was released and the whole ring deflected. The deflection of the ring was measured as a change in position of the marks, using a optical microscope. The change in diameter of the ring depends on the number of layers that were removed. The results of the slit ring method are shown in Fig. 7. The results show compression stress magnitude around -350 MPa from surface up to 5 mm depth. From that point, the values are ascending, reaching 300 MPa at the depth of 35 mm. The obtained data has been compared to the results of X-ray diffraction (XRD) and neutron diffraction (ND) (principles of both methods are described by Schajer (2013)) methods carried out on a ring specimen from the same railway axle. The comparison is shown in Fig. 8. The XRD and ND methods have been performed on the ring specimen without removing any layer. Comparison of results between slit ring method and XRD method shows general agreement between all three methods used. In subsurface region (for depth smaller than 10 mm) there is excellent agreement of all three methods and maximal compressive residual stresses was evaluated around -350MPa. In the 3. Experiment