PSI - Issue 41

Hendrik Baarssen et al. / Procedia Structural Integrity 41 (2022) 183–191

188

6

Baarsen et al. / Structural Integrity Procedia 00 (2022) 000–000 Table 1: Failure load measured for the notched plate specimens.

Nu , exp Nu , Rd

Specimen A net

N u

N u , exp

[mm 2 ]

[kN]

[kN]

B1 B2 B3 B4

726.5 282.1 326.2 1.16 746.4 289.8 329.2 1.14 731.7 284.1 327.7 1.15 728.2 282.7 328.7 1.16

0,95 1 1,05 1,1 1,15 1,2 1,25

0 100 200 300 400 500 600

Normal treated specimens B10 & B11

Stress [ MPa]

Normalised Max Clip [-]

Eng. stress - Eng. strain True stress - True strain

0

0,1 0,2 0,3 0,4 0,5

0

10

20

30

40

Cycles Normalized (225.000)

Strain [%]

(a)

(b)

Fig. 3: (a) Stress strain curves of parent material; (b) Normalized peak displacement read by the COD meter

3.2. Tensile tests on pre-cracked notched plate specimens

In order to determine when to terminate the cyclic loading, i.e when the pre-cracking procedure terminates, and the specimen is ready for the monotonic tensile test, the hole elongation is monitored with a COD-meter. The signal of the peak value, i.e. the displacement of the COD-meter when the load is maximum, has been further processed to increase consistency of the signal between the di ff erent specimens. Each signal has been normalized with respect to the value read at 10% of the predicted life as estimated by the S-N curve in Eurocode 3 part 1-9 CEN (2005c). This has been done to minimize the spread due to the positioning of the COD meter, the knife edges to which the COD-meter is attached, and to avoid the transient material behaviour due to cyclic hardening / softening occurring at the notch roots, see Figure 3b. The obtained signal is characterized, after 10% of the predicted life, by a steadily increasing trend which terminates with a rapidly increasing trend, which is the point where macro-crack formation is deemed to occur. Visual observations with a microscope confirmed the formation of a crack. Figure 3b shows the normalized data of the COD-meter during the fatigue loading of specimen B6 to B12. Here it can be seen that most of the specimen show a similar trend, until the point where a crack occurs. The di ff erences between the measurements are likely due to inconsistency of the placement of the COD-meter and geometrical im perfections, e.g. eccentricity inducing secondary bending. The di ff erent trend exhibited by specimens B10 and B11 has been associated to the surface roughness caused by rough sanding of the specimens. Cracks of relatively small size were correctly induced in specimen B5, B7,B10, and B12. Specimens B6 and B8 prematurely failed under cyclic loading and specimen B9 and B11 were tested too early in the attempt of obtaining smaller crack sizes. For these reasons, only the specimen, B5, B7, B10, and B12, are considered for further analyses. The fracture surfaces of these specimens are shown in Figure 4. It can be observed that no or little lateral contraction occurs in the presence of the