PSI - Issue 79

Santi Marchetta et al. / Procedia Structural Integrity 79 (2026) 224–232

230

Fig. 6. T-joint parametric finite element model.

With regard to the control radius values R 0,I adopted in the SED approach, a clarification is required. For the steel welded joint, LAZZARIN et al., (2003) determined it to be equal to 0.28 mm and this value is now widely consolidated. For the titanium alloy joint, on the other hand, the reference control radius was taken from the work of Meneghetti et al., (2018) and it is equal to 0.071 mm; however, it should be noted that their study refers to the base material rather than the welded configuration. 4. Results and discussion 4.1. Hotspot stress and strain approaches validation Fig. 7 shows the result obtained from the hotspot approaches validation process of both steel and titanium T-joints. The validation of the hotspot stress approach on the steel joint (Fig. 7a) shows that, for low cycle counts, the experimental data are accurately captured by the FAT 80 curve, whereas at higher numbers of cycles the points tend to follow the FAT 125 curve. This behaviour is fully consistent with IIW recommendations, which indicate FAT 71 or FAT 80 as reference design curves for welded T-joints in steel. Consequently, using this method ensures a safe and conservative fatigue design. With regard to the titanium joints (Fig. 7b), the analysis reveals that the “hotspot strain” values fall between the FAT 100 and FAT 150 normalized curves. Although this preliminary approach of normalizing the FAT curves still requires further validation, it nevertheless provides a safe basis for fatigue design, in line with the considerations previously discussed. 4.2. SED approach validation Fig. 8 displays the results obtained from the SED approach validation process of both steel and titanium t-joints. In the case of steel joints (Fig. 8a), the SED values fall within the scatter band proposed by Livieri & Lazzarin, (2005). Hence, fatigue design based on the SED approach can be considered reliable for this class of joints, as long as the lower bound (probability of survivability of 97.7%) of the scatter band is adopted as the reference design curve. For titanium joints (Fig. 8b), a clarification similar to that made for the control radius is necessary: the scatter band proposed by Meneghetti et al., (2018) refers to the base material and not to the welded configuration. Indeed, the results show that the experimental points lie completely outside this band, suggesting that a new SED scatter band should be defined for welded Ti-6Al-4V. At the current state of the art, therefore, this method cannot be considered safe for the fatigue design of titanium joints.