Fatigue Crack Paths 2003

22.0

20.0

18.0

16.0

)

m

14.0

a ( m

12.0

CT1(CA)Experiment

10.0

CT1(CA)Predicted

CT2(CA)Experiment

68.0 0

100000

200000

300000

400000

CT2(CA)Predicted

N (cycles)

Figure. 6. Predicted and measured F C Gfor modified C(T) specimens under C Aloading.

Several crack retardation models were calibrated based on the standard C(T) data

under V Aloading, including the Constant Closure model [11] (where the crack opening

load Kop was calibrated as 26% of the maximumoverload SIF, Kol,max), a modified

Wheeler model [12-13] (where the model’s exponent was estimated as 0.51), and N e w

man’s closure model [14] (generalized for the V Aloading case, where the stress-state

constraint factor was fitted as α =1.07, suggesting dominant plane-stress F C Gcondi

tions). The fitted load interaction parameters were then used to predict in the ViDapro

gram the crack growth behavior under V A loading of the hole-modified CT1(VA)

specimen, see Fig. 7. The significant retardation effects of the CT1(VA)specimen were

very well predicted using these three load interaction models.

14.0 ()m12m680.0

12.0

Without Retardation

a

Experiment

Modified Wheeler (0.51) C nstant Closur (26 %).

10.0

N e w m a(n1.07)

8.0

0

100000 200000 300000 400000 500000 600000 700000

N (cycles)

Figure 7. Crack growth predictions (based on straight-crack calibrations) on a modified

C(T) specimen under V Aloading.