PSI - Issue 68

M. Zarazovskii et al. / Procedia Structural Integrity 68 (2025) 391–397 M. Zarazovskii et al./ Structural Integrity Procedia 00 (2025) 000–000

395

5

values, five of which were situated within the 77–293 K range. These values were achieved using a mixture of liquid nitrogen and concentrated ethyl alcohol in varying proportions. In the course of the FT tests, a Lutron TM-9017SD laboratory thermometer with a measurement accuracy of ±0.1°C was employed. The geometric characteristics of the half-size SE(B) and the compact C(T)0.16T specimens are shown in Fig. 2. Due to the unavailability of material for the three-point bending experiments, only five half-size SE(B) specimens of each material were manufactured. Subsequently, a further 20 specimens were produced from the broken halves, with a dimension of 0.16 CT. The experiments were conducted at temperatures ranging from -100°C to -196°C.

a b Fig. 2. Geometry of: (a) SE(B) specimen; (b) C(T)0.16T specimen.

2.3. Data processing and analysis The most challenging task for mini-CT testing is the crack growing. For such small dimensions a high accuracy to tolerances for specimen, grips and pins must be provided. Material selection for the Ø2 mm pins is also an issue, since pins can deform/fail leading to unsymmetric crack followed to stop cycling, rotate the specimen and return to cycling. Even if it initially defined that the crack is correct, post-experiment analysis can reveal that it is not (see Fig. 3). Based on the data obtained from the test machine a displacement versus force (or COD) graph is plotted. This curve is divided into at least two parts: a linear part and a curvilinear one (at that, non-linear part can be also divided into two functions). Each of these parts is approximated separately and then 'glued' together to the final piecewise smooth plot. When approximating a fracture diagram, an important consideration is the physicality of the choice of functions used to represent the different parts of the diagram. This means that the functions selected should adequately reflect the physical processes that occur in the material during loading – not only the mathematical accuracy of the approximation but also the logic of the model (i.e. function must be smooth without extremums inside). For example, in our cases the curved part of the graph is approximated by equation y=a·x+b·x 2 +c·x/ ( d+e·x+f·x 2 ). Once the final diagram had been constructed, analytical integration was performed in order to determine the plastic region for further calculations – all in accordance with the ASTM E1921 up to the T 0 value determination.

Fig. 3. Specimen with a not appropriate initial cracks.

3. Result and Discussion All results were obtained in fully accordance with the ASTM E1921 procedure for data treatment. The screened was carried out in accordance with the following procedure. At the beginning, the reference temperature was determined taking into account all the results of the experiment, the results that did not fall within the range of -50°C to T 0 to 50°C were discarded and a new value of T 0 was determined. The screened values that fell within the newly defined range of -50 °C to T 0 to 50°C were incorporated once more to redefine T 0 . This process was repeated until the calculated results fell within the specified range of -50 °C to T 0 to 50 °C, and the screened values did not.