PSI - Issue 40

Aleksey Antimonov et al. / Procedia Structural Integrity 40 (2022) 17–26 Aleksey Antimonov, Nadezhda Pushkareva / Structural Integrity Procedia 00 (2019) 000 – 000

20 4

0,6

1 ln 1 16 1  

 

0,6

N 

    

.

(2)

p

 

5. Experimental research Investigations of the materials strength to cyclic elastoplastic deformations, as a rule, are carried out under a homogeneous stress condition which is practiced at uniaxial tension and compression. The conditions for the bars breaking process by cyclic bending with notch for stress concentration significantly differ from samples laboratory tests at homogeneous stress condition; so, the possibility of using the formula (2) for practical calculations requires experimental verifications. This test was carried out by breaking thick-walled pipes made of low-carbon steel with carbon content of 0,2% by Sokolovskij, Poljanskij and Antimonov (1988). Experimental data for strain rate dependence  for pipes with different mechanical properties, diameter and wall thickness as a function of cycles number N before fracture are shown in Fig. 3. The comparison of calculated results to the formula (2) and the experimental data is shown in Fig.4. As it is shown in the figure, there is a significant difference between the calculated and experimental data. For a small number of cycles, the calculation results are six or more times higher than the experimental data.

Figure 3. Experimental data of the dependence of deformation range vs. number of cycles to fracture.

Figure 4. Comparison of calculated results by the formula (2) with experimental data:

Pipes sizes, mm: 1 – 9,6x2,95,  = 60; 2 – 13x3,32,  = 60; 3 – 13x3,32,  = 40.

1 and 2 - calculations, 3 and 4 – experiment Pipes sizes, mm: 1 and 3 – 9,6x2,95,  = 60; 2 and 4 – 13x3,32,  = 40.

6. Assessment of the rolled steel breaking process productivity Similarity of the curves in Fig. 3 and Fig. 4 for the calculated and experimental data shows that the qualitative of the deformation range dependence on the number of cycles before failure is the same in both cases. The formula which gives a good agreement between the calculated and experimental data is as follows

0,26

1 ln 1 16 1  

 

0,26

N 

  

(3)

 

The comparison of calculations by this formula with the experiment is shown in Fig. 5. The difference between the calculated and experimental data does not exceed 7%.