Fatigue Crack Paths 2003

side approximately in the ± 36°direction from the symmetry plane of the specimen, see

Fig. 1 top. On the backside the crack initiation point was first in the symmetry plane, as

expected, see Fig. 1 bottom. During the first part of the test, both cracks on the front

side propagated simultaneously but not continuously. The crack on the backside

propagated with an expected crack growth rate of approximately 2.0 nm/cycle. In the

second part of the test, the crack on the back side in the 0°-direction stopped and

suddenly the cracks in the ± 36°direction, first only observed on the front side, broke

through and were seen also on the back side. The loading was changed three times from

Δ K increasing to Δ K decreasing and to Δ K increasing again. The Δ K value calculated

based on the projected crack length on the symmetry plane varied between 4.5 and

8.0 M P a √ mand the crack growth rates varied between 1.0 nm/cycle and 100 nm/cycle.

In the last phase of the test a poor resonance characteristic of the system dominated and

only the upper crack propagated. The test was stopped, the specimen was photographed

and the crack paths were measured under the microscope. The results are shown in Fig.

1 right.

The interpretation of the test history is as follows. At the beginning of the test, two

cracks initiated in the ± 36°direction on the front side and one crack in the symmetry

plane on the backside. For the loading condition of Δ K approximately 5.0 M P a √ mthis

coexistence of three crack fronts was in a metastable condition. When Δ K decreased

further, the cracks in the ± 36°direction propagated faster than the crack in the

symmetry plane. The crack in the symmetry plane stopped, when it was shadowed

completely by the two cracks in the ± 36°direction. This happened, when Δ K was below

5.00 MPa√m.Increasing

Δ K could not reverse this crack branching.

For the second specimen the load level at the beginning of the test was higher than

for the first one, such that a Δ K value of 10.0 M P a √ mresulted for the crack at the notch.

After crack initiation a load shedding test was performed. The expected crack growth

path in the symmetry plane was observed on both sides, until the

Δ K value reached

5.0 MPa√m.The crack growth rate was stable at approximately 1.5 nm/cycle. The crack

length had reached approximately 28.5 m mon both sides, when below and above the

crack in the symmetry plane additional cracks appeared. The position of the top crack

tip was measured on the backside and the position of the lower crack tip was measured

on the front side of the specimen. Both cracks were propagating more or less

simultaneously. The load was further decreased until a crack length of approximately

33 m mwas reached, when the test was stopped. The specimen was photographed and

the crack paths were measured under the microscope. The results are shown in Fig. 2.

The interpretation of the test history with the second specimen is as follows. At the

beginning of the test, one crack initiated in the symmetry plane as expected. The crack

propagated to a length of approximately 25 mm, when Δ K reached 6.5 MPa√m.

Probably at this stage two additional cracks initiated in the ± 40° direction, but only in

the bulk of the specimen. As in the first specimen the situation was that three crack

fronts were in a metastable condition. On both surfaces the crack in the symmetry plane

was propagating with 1.5nm/cycle. When Δ K decreased further, the cracks in the

± 40°direction propagated also towards the free surfaces, until they were visible at