PSI - Issue 13

Andrzej Neimitz et al. / Procedia Structural Integrity 13 (2018) 862–867 Andrzej Neimitz et al. / Structural Integrity Procedia 00 (2018) 000 – 000

864

3

point to the large-strain domain. Typically, the extrapolation is preceded by the approximation of the TS – LS curves. The approximation is performed using either a power function (i.e., the Ludwik constitutive equation) over the plastic part of the curve or a linear function of the final part of this curve before the maximum is reached. The constitutive curves were implemented into ABAQUS finite element code. After numerical simulation of the experimental tests on the specimens shown in Fig. 1, the force – elongation curves were recorded and compared with those obtained experimentally. Selected results are shown in Fig. 2.

a)

b)

Fig. 2. Experimental and numerical curves before and after calibration: a) PN specimen, b) PR specimen.

Several conclusions can be drawn from the obtained results: 1) In most analysed specimens, the experimental curves lie below the numerical curves.

2) The distance between experimental and numerical curves decreases as the temperature of the tests decreases. In some cases, at low temperatures, the numerical and experimental curves partially overlap, and at -50ºC and -20ºC rare numerical curves are located below the experimental ones. 3) The method of approximation of the TS – LS experimental curves is not meaningless. In the case of notched cylindrical specimens, when the linear function is used for approximation, the constitutive equation can be calibrated more easily to fit the experimental curve after numerical computation. At lower temperatures, calibration of the constitutive equation is often not successful, using the calibration procedure discussed above, when the power function is used during approximation (see Fig. 3).

a)

b)

Fig. 3. Force – elongation curves obtained numerically and experimentally: a) C1 specimen, C04 specimen.