PSI - Issue 57

Inge Lotsberg et al. / Procedia Structural Integrity 57 (2024) 569–580 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

571

3

Dynamic loading

σ s1

L1

Fillet weld

A

Hot spot

σ s2

σ d

Fillet welded plate to vertical stiffener

t s

t d

A

A - A

Fig. 1. Analysed structural detail.

2. Laboratory testing of fillet welded doubling plates

2.1. Laboratory fatigue testing

The laboratory fatigue testing of the specimen with the fillet welded doubling plates in Fig. 2 is described in Lotsberg et al. (2014) and in Lotsberg (2016). Fatigue testing was performed on four such specimens shown in Fig. 2; one in as-welded condition, two after weld toe grinding and one specimen after ultrasonic peening of the weld toes. On each test specimen there was two circular and one quadratic doubling plates of thickness 10 mm with diameter and length equal to 150 mm fillet welded to a main plate of thickness 20 mm. The corners of the square plate were rounded by a radius equal to 25 mm. The testing was performed to crack growth through the plate thickness and the number of test cycles was compared with design standards. The number of test cycles to failure for the as-welded test specimen corresponded to the mean S-N curve F1 which is the same as FAT 63 in IIW (Hobbacher, 2009). The measured stresses from strain gauges placed at position 10 mm, 20 mm, and 30 mm from the weld toes for specimen 1 with a = 7 mm are shown in Fig. 3. The single filament strain gauges were placed in a direction normal to the weld toes. A quadratic function is fitted to the measured strain values to derive the stress graphs. Young’s modulus equal to 2.1·10 5 MPa is used to calculate stresses from the measured strain values. Extrapolated stresses from positions 10 mm and 30 mm to the weld toe corresponds to that recommended for finite element analysis for Method A in DNV-RP-C203 (2019) with extrapolation from 0.5t and 1.5t back to the weld toe. The test data and the calculated stresses corresponds to a load level of a unit nominal stress in the main plate at the considered section. Measured strain data at the 6 hot spot locations shown in Fig. 2 and results from finite element analyses showed similar stresses at the hot spot locations for the circular plates as the quadratic plate. Thus, the results from all the 6 hot spot locations were expected to belong to a similar group. Some scatter in measured values may be expected due to some variability in geometry and in positioning of strain gauges. For specimen 1 a mean stress concentration factor equal to 1.48 is derived. The standard deviation in the 6 data points is equal to 0.036 and a Coefficient of Variation (CoV) equal to 0.024. For specimen 2 with ground weld toes a mean stress concentration factor equal to 1.39 is derived. The standard deviation in the 6 hot spot data points is equal to 0.069 and a CoV equal to 0.049.