PSI - Issue 39

10

Hannes Panwitt et al. / Procedia Structural Integrity 39 (2022) 20–33 Author ame / Structural Integrity Procedia 00 (2019) 000–000

29

40

a

b

21

Potential th c w

Potential th ε 1

= 0.3%

= 0.4% const.

th

= 0.3% w/o outliers

30

18

c w

= 1.0% const.

th

ε 1

= 2,

n

40

t h

20

15

th

= 0.4%,

6.5

6.75

7.0

30

ε 1 , s t a r t

a [mm]

= 1.0%

th

ε 1 , e n d

20

10

a [mm]

10

0

0

0

2

4

6

8

0

1

2

3

4

5

6

7

8

N [-]

N [-]

10 5

10 5

0.00 1.46 2.92 4.38 5.84 7.30

c

Surface Distances

Surface Distances

0.0mm

8.3mm

0.0mm

8.3mm

th cw = 0.4 % = const.

th ε 1 = 1.0 % = const.

N ×10 5 [-]

Fig. 6: Calibration of the a - N -measurement methods with a force-controlled test ( F max = const.). a) ε1 -method; b) Crack width method; c) Detected crack paths with different th ε1 .

4.3. Validation for in-phase mixed mode loading conditions To validate both methods on mixed mode loading conditions the direct current potential drop technique cannot be used because of the complex crack geometry inside the specimen. Instead, the marker load technique is utilized as a reference. For this purpose, a cyclic test was carried out with an in-phase superimposed tension-torsional loading with R axial = 0 and R torsion = 0. The maximum load ( F max = 27.5 kN) and the maximum torsional moment ( M t,max =105 Nm) were kept constant. This corresponds to a load level of σ N,max / R p0.2 = 0.07 and a load ratio of τ N,max / σ N,max = 1.31 at the beginning of the experiment. Every 10 5 load cycles, a single overload was introduced with an overload ratio of 2 ( F ol = 2 ⋅ F max and M t,ol = 2 ⋅ M t,max ). After the experiment was stopped after 1.6×10 6 cycles, the final fracture was initiated quasi statically by means of a tensile testing machine. The resulting bottom half of the specimen is shown in Fig. 7a and the top half in Fig. 7c. Both kinking and twisting of the crack are visible on the fracture surfaces. Fig. 7b shows the fracture surface in more detail from the back side. Arrest marks are clearly visible, from which a - N -curve can be obtained until the end of the test for this side of the specimen. However, on the front side of the specimen (Fig. 7d) the main crack branches similar to the example in Fig. 3. Thus, the fatigue fracture surface of the main crack is not completely exposed and as a result arrest marks can only be evaluated up to 1.2×10 6 cycles.