PSI - Issue 18

Benjamin Möller et al. / Procedia Structural Integrity 18 (2019) 556–569 Author name / Structural Integrity Procedia 00 (2019) 000–000

565

10

2.6. Fatigue test results under variable amplitude loading A fatigue assessment of maritime structures under variable amplitude loading (VAL) is related to a linear distributed spectrum, as was already applied in previous research activities, e.g. QuInLas project (2014). Therefore, a linear load spectrum is used here for variable amplitude fatigue testing. In Fig. 10(a) a linear load spectrum with a maximum force amplitude of � � = 14.0 kN is shown. For a reduction of the testing time, load levels with a very low damage content have been removed from the load spectrum using omission. This is the case for force amplitudes equal to or smaller than 1 kN in terms of fatigue testing of adapters V2, which is the only specimen type investigated under VAL. A random load-time history has been derived from the linear spectrum incl. omission, as shown in Fig. 10(b).

Fig. 10. (a) Linear load spectrum for a maximum force amplitude of � � = 14.0 kN and (b) corresponding random load-time history Fatigue test results under variable amplitude loading of the random load-time history for the linear spectrum and the evaluated Gaßner curve have been added to the Wöhler curve of adapters V2 (cf. Fig. 12). The Gaßner curve is characterised by a force amplitude of � � ( � � = 2ꞏ10 6 ) = 12.86 kN, a slope � = 8.5 and a scatter � � = 1:1.45. Compared to results under CAL, the slope flattens ( � > k ) and the scatter is increased. The increased scatter might be traced back to the fact that welded adapters without and with control of the penetration depth have been investigated, which show a difference in the weld seam characteristics, and that the failure location and behaviour change due to partial high load amplitudes of the spectrum, as documented in Fig. 11. The failure location changed from the first weld seam in the steel sheets and crack propagation through the aluminium under CAL to the third weld seam (higher welds in fatigue testing in Fig. 11) and crack propagation through the steel sheet. The explanation for the variation of the failure location is given by the different load levels between CAL and VAL. Under high load levels of VAL, the load path through the specimen changes due to partial or full fracturing of weld seams from the aluminium sheet, so that the joint type changes from mainly metallic continuity to a domination of form closure. This results in a modified stress field with different stress concentrations. Due to the change in the failure behaviour from CAL to VAL and the investigation of welded adapters without control of the penetration depth, a linear damage accumulation on the basis of the experimental Wöhler curve is not applicable. Evaluated Wöhler and Gaßner curves of adapters V2 are shown in Fig. 12.

Fig. 11. Fatigue failure of LBW adapters V2 joined (a) without and (b) with control of the penetration depth under VAL at � � = 12.0 kN