Crack Paths 2012

The introduction of a new conservative locus, as proposed by Desimone et al in [7],

seems to better reproduce the experimental results. For the smaller defect size, see Fig.

6a, the fatigue limit in out of phase is correctly predicted. This tendence is confirmed by

the second defect size, also if the prediction seems to be a little more conservative.

The introduction of the conservative locus seems to be a necessary condition for the

prevision of mode III failure: the out-of-phase tests have clearly shown that in fact this

is failure mode completely different from the ‘usual’ Mode I that defines the Dang

Van’s locus (which is defined by fatigue limits under tension-compression and torsion),

so the superposition of a more conservative limit appears to be correct.

BEARINGSTEEL —area = 220 P m

1.50

'KIII/'KI,th,LC

= 0.62

'KIII/'KI,th,LC

= 0.38

'KIII/'K

= 0.30

I,th,LC

1.25

Coplanar crack depth = 85 P m

N = 1.2x105 cycles

1.00

0.75

Coplanarcrack depth = 70 P m

Original DangVan

N = 5x105 cycles

Proposed conservative

0.50

locus

locus

0.25

Coplanar crack depth = 0 P m

V

N = 106 cycles

1.0

-1.5

-1.0

-0.5

0.0

0.5

(a)

h /V

w

(b)

BEARINGSTEEL —area = 630 P m

'KIII/'KI,th,LC

= 0.68

'KIII/'KI,th,LC

= 0.53

'KIII/'KI,th,LC

= 0.49

crack

1.2550

Vh/V

Coplanar N = 1 0 6cycles

depth = 947 P m

Coplanar crack depth = 1 0 0 P m N = 5x105 cycles

Coplanar crack depth = 0 P m N = 5x105 cy les

1.00

0.75

Proposed conservative locus

Original DangVan

0.2550

locus

1.0

-1.5

-1.0

-0.5

0.0

0.5

w

Figure 6. Application of Dang Van’s criterion. Bearing steel- Load Path 1. a) defect size

¥area = 220 Pm, b) defect size ¥area = 630 Pm.

This idea apparently looks to be supported by the analysis of LP2 and torsional tests.

Since failure under these tests is controlled by Mode I, drawing all these load paths in

the Dang Van’s plane, it is possible to observed that in all cases the mode I failure is

22