PSI - Issue 3

A. D’Aveni et al. / Procedia Structural Integrity 3 (2017) 432–440

437

6

Author name / Structural Integrity Procedia 00 (2017) 000–000

thermoelastic phase is to consider exhausted, are defined.

Fig. 4.  T vs time curve (red points) and stress vs time curve (blue line) for spot number 5 .

Fig. 5.  T vs time curve (red points) and stress vs time curve (blue line) for spot number 4.

The Figure 5 refers to the spot located on a corner of the same specimen. In it is reported the trend of the temperature that, unlike what occurred for the central spot, decreases with continuity up to a value in which it is possible to define, even if it is in a less evident than occurs in the previous central spot the change of the slope. The figure 6a, that is related to the thermal image immediately before the breakage of the specimen, shows the distribution of temperatures near a corner which, as it is known, is completely different from the one found on the center spot. This is due to the plate of the testing machine that activates a distorted tensional state. Despite the plate machine effect detectable in the corner spot, the value of the critical load defined as the intersections point of two interpolating straight lines, is equal to about 1000 kN and so practically comparable with the one recorded on the central spot (spot 5). In all the examined specimens, it was observed a trend of temperature data similar to the one already described. In all specimens, it can be clearly identified the value of the load of the first change of slope of the interpolating line the points detected by the thermography camera. From these points, it was easy to read on the load diagram, the critical load and the related "critical stress".

Fig. 6. (a) thermal image of the specimen at the failure; (b) hourglass shape of the specimen at the failure. The assessment of the critical load was performed with reference to the spot positioned at the centre (spot 5) of all specimens where the plate machine effect was almost absent and the temperature variations were better comparable due to their better homogeneity. However, the temperature data detected on the corner that first showed signs of failure, were used as result comparison. The load data, the variation of temperature and the machine time, shown in the diagrams, were coordinated among