Issue 33

D. G. Hattingh et alii, Frattura ed Integrità Strutturale, 33 (2015) 382-389; DOI: 10.3221/IGF-ESIS.33.42 387 ( ) , ( ) t eff k A ref eff f A a N N             (6) To conclude, it is worth observing that the MWCM has proven to be highly accurate in performing the multiaxial fatigue assessment of conventional welded joints when it is applied not only in terms of nominal [20-22] and hot-spot stresses [23, 24], but also along with the reference radius concept [24-27] as well as the Theory of Critical Distances [26-30]. 1000 10000 100000 1000000 10000000 100000000 1000 10000 100000 1000000 10000000100000000 N f [Cycles] N f,e [Cycles] Axial loading, R=-1 Axial loading, R=0.1 Torsion, R=-1 Torsion, R=0 =0°, =√3, R=-1 =0°, =√3, R=0 =90°, =√3, R=-1 =90°, =√3, R=0 =0°, =1, R=-1 =0°, =1, R=0 P S =90% P S =10% Non-Conservative Conservative Torsional Scatter Band Uniaxial Scatter Band FSW Aluminium Tubes  B R  B R  B R  B R  B R  B R Run out Figure 6 : Accuracy of the MWCM in estimating the fatigue lifetime of the tested Al 6082-T6 FS welded joints. V ALIDATION BY EXPERIMENTAL DATA n order to check the accuracy of the MWCM in estimating the fatigue lifetime of the tested FS welded joints, initially the calibration constants in Eqs (2) and (3) were determined, according to Eqs (4) and (5), by using the fatigue curves generated under fully-reversed uniaxial and torsional fatigue loading (see Tab. 1), i.e.:   4.3 10.8 eff eff k        (7)   Re 21.2 38.9 f eff eff        [MPa] (8) As proven by the uniaxial fatigue curves summarised in Tab. 1, the axial fatigue strength of the investigated FS welded joints was seen to be sensitive to presence of non-zero mean stresses, this holding true even though the specimens were tested in the as-welded condition. As to this aspect, it is interesting to observe that a similar behaviour has been observed also in conventional welded joints tested, in the as-welded condition, under uniaxial fatigue loading (see, for instance, Refs [31, 32] and references reported therein). According to this experimental evidence, in this initial investigation the mean stress sensitivity index, m, was simply taken equal to unity [10]. Further, owing to the fact that the MWCM was aimed to be applied in terms of nominal stresses, the limit value for stress ratio  eff was set equal to 1.3 (i.e.,  lim =1.3) [10, 20]. After calibrating the MWCM, multiaxial fatigue software Multi-FEAST (www.multi-feast.com) was used to systematically post-process all the experimental results that have been generated so far. The experimental, N f , vs. estimated, N f,e , number of cycles to failure diagram reported in Fig. 6 summarises the overall accuracy which was obtained by using the MWCM to predict the lifetime of the FS welded tubular samples being tested. I