PSI - Issue 17

Eduardo A. Lima et al. / Procedia Structural Integrity 17 (2019) 246–253 Eduardo A. Lima/ Structural Integrity Procedia 00 (2019) 000 – 000

251

6

Table 1. Parameters used for Dang Van criterion (Moyar and Stone, 1991). Parameters Symbol Value

Units MPa MPa

σ uts

Ultimate tensile strength

1220

τ f

’

Fatigue shear strength coefficient Fatigue shear strength exponent

970

b’

-0.111

3. Results and discussion

The results for thermal treatment model, rolling models, and fatigue model were presented in this section.

3.1. Thermal Treatment Model

The FEM results for the thermal treatment model were analyzed two positions in the wheel, as defined in Fig. (1). Such results are showed in Fig. (6), in log scale. This graphic shows that the rim was gradually cooled during the quenching, differing from the tread behavior. Subsequently, these elements are brought to the same temperature ranges during the drawing back and in the slow cooling. In the same graphic is noted that the circumferential compression stresses are larger, in modulus, in the tread of the wheel than inside of the rim.

Fig. 6. Temperature-time and hoop stress evolution history of two elements in the wheel during the thermal treatment process.

3.2. Rolling Models

With loads of the wheel-rail model applied in the tread of the wheel model , it is possible to calculate the stress distribution, as showed in Fig. (7) for four conditions: using the complete wheel-rail model with thermal treatment process (TTP), the same model without the thermal treatment, the wheel model with and without the TTP. Fig. (7a) and Fig. (7b) show the residual stress obtained in the wheel model was similar to the wheel-rail model after the first rolling pass, both without residual thermal stress. The same happened in models with residual thermal stress, Fig. (7c) and Fig. (7d). After proving that both approaches lead to similar results, the load was applied during eight passes on the tread of the wheel model. Fig. (8) shows the comparison of the hysteresis cycles in the wheel models with and without residual stresses generated with the TTP. The results were extracted from the element with the maximum von Mises stress in the cross-section, Fig. (5), in the subsurface of the wheel. It can be noticed that when the wheel does not have the residual stresses from the TTP, the final deformation was 22.7% greater than in the model with it. Besides, the initial plastic deformation or hardening added to the material of the wheel model with TTP caused the wheel to reduce one cycle until the plastic stabilization.