PSI - Issue 13

Reza H. Talemi et al. / Procedia Structural Integrity 13 (2018) 775–780 Author name / Structural Integrity Procedia 00 (2018) 000–000

779

5

4. Result and discussion 4.1. Experiment

Fig. 2(a) plots the absorbed energy, J DWTT , at different tested temperatures. The absorbed energies, for the tested steel slab material, are almost similar for both configurations at all temperatures expect for 300°C. It can be noted form the figure that the absorbed fracture energy increases from 25°C to 300°C. The maximum value of the J DWTT is around 4kJ at 300°C for config2 and decreases until 500°C. However, for config1 the energy value reaches a plateau at 200°C and slightly drops till 500°C. The fracture surfaces of the samples tested at 25°C have shown a 100% brittle fracture appearance. Results obtained for config1 at 100°C, show mainly transgranular brittle fracture with a very small amount ductile areas which occur at the edges; however, with a significant higher amount of ductile fracture (almost 100 % ductile fracture). Fig. 2(b) shows optical and SEM images of the fracture surfaces of the broken samples for config2 at 100°C and config1 at 300°C. The fracture surface of DWTT-config1 sample tested at 300°C, shows a mix of shiny and matt areas, in which always ductile fracture could be observed. The dimples in the matt areas are more elongated. Several shrinkage cavities (with a typical dendritic wall) were observed. Just below the notch, a few cracks were seen which have been grown parallel with the notch. The visual aspect of the fracture surfaces of the samples tested at 400°C and 500°C was hard to be interpreted concerning the ratio of brittle area and ductile area. 4.2. Simulation Fig. 3(a) compares the force versus displacement curves obtained from the FE model and the experimental observations for both configurations tested at 25°C. As shown, the FE model predicts the dynamic fracture initiation and propagation reasonably well. The good correlations between the numerical model and experiments were found for 25°C, 100°C and 200°C testing temperatures. However, for higher temperatures, i.e. 300°C, 400°C and 500°C, the agreement between the numerical estimation and experimental results were not as good as other temperatures. The main reason for this behavior is linked to the fact that at higher temperatures the fracture response of the steel slab material is mixed ductile and brittle and the proposed XFEMmodel underestimates the crack initiation and propagation energies. This is shown in Fig. 3(b) by comparing the absorbed energy for crack initiation between the numerical results and the experimental observations for both configurations at all tested temperatures. Nonetheless, the predicted energy for the crack initiation, for critical temperatures i.e. 25°C to 200°C, are in very good agreement with the experimentally measured ones. Using the predicted crack initiation energy, it is then possible to quantify a critical stress value which is used to assess the dynamic cleavage fracture of the steel slab when subjected to thermal stresses in the reheating furnace or mechanical stress due to transportation of the slabs. Fig. 3(c) suggests a design curve which plots the critical cleavage fracture stresses at different temperatures for both tested configurations.

Fig. 2. (a) absorbed energy, J DWTT , versus testing temperature; (b) fracture surfaces and SEM images of broken samples for config2 at 100°C (left) and config1 at 300°C (right).