PSI - Issue 35

Toros Arda Akşen et al. / Procedia Structural Integrity 35 (2022) 82 – 90 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

89

8

Punch stroke prediction for the TWIP940 steel was compatible with the experiment. It may result from that the hardening exponent of TWIP940 steel is very high in comparison to the conventional steels so that the localization of the deformation was postponed in UTT simulations. Besides, strain localization was not observed before macro cracking during UTT [1]. This situation provides a reliable prediction accuracy. On the other hand, a significant strain localization was observed for the TRIP590 steel in UTT. This instability hinders the TRIP590 steel to show his own potential in tensile test [1]. However, this instability is not seen in HET and thus the damage indicator obtained from UTT simulation led to a discrepancy in fracture strain prediction in HET. Fig.6 shows the potential fracture initiation locations.

a

b

Fig. 6. Locations of fracture initiation for (a) TRIP590 (b) TWIP940

Numerical results showed for both steels that, initial cracks were occurred near the hole edge. For TRIP940, failure was initiated in RD while, for TRIP590 steel, first crack initiation was seen in RD and TD. 8. Conclusions In the present study, the failure prediction performance of a numerical model with HomPol4 yield criterion was evaluated in HET. Generalized plastic work criterion was used as a damage model. Two AHSS steels namely TWIP940 and TRIP590, were studied. Initially, HERs were predicted by HomPol4 criterion at the failure strokes for both steels. Later, punch strokes which the failure occurs were estimated by HomPol4 criterion along with the generalized plastic work criterion. The conclusions were summarized below: • For both steels, HER predictions were in a good agreement with experimental results. • Numerical results obtained from HomPol4 criterion showed that the HER value was slightly underestimated for TRIP590 but overestimated for TWIP940 steel. It can be concluded that the yield criterion has a noticeable influence on HER prediction. • For TWIP940 steel, HomPol4 provided accurate result of prediction for punch stroke at which the failure occurs. However, failure stroke for TRIP590 steel was slightly underestimated by the numerical model. This may stem from the fact that, in UTT, the TRIP590 steel exhibited high strain localization but, in contrast, TWIP940 did not show significant strain localization. Since the damage indicator for both steels were obtained from UTT simulations, the difference in the amount of strain localization affected the success of the failure stroke predictions. • Generalized plastic work criterion is a practical damage prediction approach. This criterion provides accurate results for the materials showing limited amount of strain localization. However, for the materials showing high amount of strain localization, plastic work criterion may lead to inaccurate results for the fracture prediction in forming processes. Therefore, for this kind of materials, a method taking the effect of the instability into account should be included to the determination process of the damage indicator in order to compensate this effect.