Page 375 - Biaxial Multiaxial Fatigue and Fracture

Three-Dimensional Crack Growth: Numerical Evaluations and Experimental Tests 359

Fig. 22. Experimental and numerical crack shapes for through cracks at different stages.

CONCLUSIONS
With reference to two-dimensional MSD crack propagation, a satisfactory agreement was
obtained between numerical and experimental crack propagation rates on specimen 1 when
using the Paris formula, with the related constants provided by in house made experimental
tests. Such formula was not anymore accurate for variable amplitude load cycles, as applied to
specimen 2, because unable to keep in account the load ratio variability. That is why a more
complex correlation, based on an enriched set of experimental data and on information from
NASGRO database (without the need to model crack closure effect), was attempted getting a
satisfactory agreement between numerical and experimental results. The later approach could
be improved by increasing the experimental data by cycling some simple notched specimen
with different R values.
For that concern the three-dimensional crack propagation again a very interesting correlation
between numerical and experimental results was obtained even if some further refinement are
necessary with regards to the crack propagation times. It is to point out the extreme flexibility