Fatigue Crack Paths 2003

Table 2. Experimental data and calculation results of angle calαˆ

) W ( ˆ 1 c a l α

) W ( ˆ 2 c a l α

) W ( ˆ 3 c a l α

) W ( ˆ 4 c a l α

No

σr -

σ λ -

expˆα minexp,ˆα maxexp,ˆα

deg

deg

deg

deg

deg

deg

degI

1 -

0

1.5

0.38

2.57

44.5

0.0

0.0

0.1

2 -

∞ 43.6

39.2

47.9

44.4

45.0

45.0

0.0

3 1.00 0.50 20.5

15.8

25.2

22.3

22.5

22.5

22.5

4 1.00 1.00 31.4

27.5

37.3

31.4

31.7

31.7

13.3

5 0.50 0.54 16.2

12.3

20.0

24.8

19.3

20.4

18.6

6 0.50 0.97 28.2

24.2

32.3

29.9

32.5

33.6

8.1

7 -0.01 0.56 4.8

0.8

8.7

25.5

15.5

17.1

13.9

8 -0.01 0.97 31.2

26.3

36.2

30.2

32.9

35.9

4.1

cal α α ˆ ˆexp −

In Fig. 4 the absolute values of difference

between experimental and

calculated values of fatigue crack lines angles using the weight function method are

shown. For pure torsion, pure bending and loading with the cross correlation coefficient

rσ= 1 and rσ= 0.5 the weight function W2 and W3 correlate the experimental data very

well. In the case of loading with the cross correlation coefficient equal to

rσ= - 0.01, a choice of proper weight function depends on the maximumstresses ratio

λσ. For λ σ =0.56 the weight function W4 based on parameter of shear strain energy

density correlates experimental data in the best way. For λ σ =0.97 the best results

obtained using weight functions W1and W2.

α α ˆ ˆ exp −

Figure 4. Absolute values of difference

between experimental and calculated cal

values of fatigue crack lines angles using the weight function method