Crack Paths 2012
Ic f m m K N a Y N K ' ' ' ) ( ) ( S V allowing us to express the final crack size ratio,
m f w f a a , , / as:
1
2
VV
(5)
, mw f a
¸¸¹·¨¨©§''wm
!
.
Assuming a constant geometry factor Y during the crack propagation, the current
crack size ratio r
2
(6)
) ( ¸ ¸ ¹ · ¨ ¨ © § ' ' w m V
mw r N a
remains constant along the whole propagation process (see Fig. 3(b)), in particular, from
the beginning of the fatigue process, i.e., for the initial crack sizes
m i m a a ,
up to the final failure state, proving that on a logarithmic scale, the vertical
and
w i w a a , ,
distance c=log r between the curves related to
wia, and
mia, remains constant all along
the propagation process (see Fig. 3(b)).
(a)
(b)
Figure 3. Crack growth curves for the best and worst initial defect cases for a given
material (a) plotted in natural scale, and (b) plotted in logarithmic scale.
While the initial worst crack size a , related to p=0, is determined from Eq. (2), w, i
another crack size, a m, i , related to a given probability p (particularly the best one bia,
related to p=1, remaining, in principle, unknown) can be obtained by integration of the
) ( / K G d' N,ddue tao its uniqueness, for a generic number
crack growth rate law,
L N of cycles to failure , i.e.:
a
da
L N d N
N
m if
,
L
N a a a
³
) , , ( , m L m i m f V' . (7)
o
) ) ( ( a K G ' ³ 0
m f
, ,
,
Since Eq. (6) applies irrespective of the number of cycles considered, the above
quantities can be used to write a system of two equations:
1045
Made with FlippingBook Ebook Creator