PSI - Issue 62

5

Annarosa Lettieri et al. / Procedia Structural Integrity 62 (2024) 789–795 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

793

Fig. 2 compares all collected data and the current fatigue details suggested to quantify the fatigue strength of riveted shear splices. The fatigue details are denoted with an acronym indicating the assumed corresponding fatigue curves in terms of resistance and slope, i.e., 71m3 represents the fatigue curve recommended in the JRC-ECCS technical report (Kühn et al. , 2008), while 80m5 and 90m5 refer to the details proposed by Taras et al . (2010). The experimental fatigue results do not include tests in which failure was not observed. Additionally, aiming to perform a reliable comparison with the current suggested S - N curves, the beneficial effect of compressive stresses on the fatigue strength has been accounted for by reducing the stress range of an amount equal to 60% of the compressive portion (Eurocode 3 – Part 1.9). The experimental data have been statistically synthesised according to the methodology described in Section 2, with a fixed slope equal to 5 (based on the primary outcomes provided by literature, i.e., Taras et al . 2010). The results are reported in Fig. 2 through the mean curve, the 95% and 99% confidence intervals. The comparison highlights that detail 80, proposed by Taras et al . (2010), well represents the 95% fractile of the experimental collected data.

Fig. 2. Fatigue experimental data.

4. Influence parameters on the fatigue strength The collected experimental data have been filtered and separately grouped to evaluate some parameters' influence on the investigated details' fatigue resistance. Specifically, the influence of the joint configuration (single- or double shear), the distances from edges e 1 , e 2 and the plate width w normalised to the diameter d have been included in the present study. e 1 and e 2 represent the end distances from the centre of the rivet holes to the adjacent end, measured in the parallel and perpendicular directions of load transfer, respectively (Fig. 1). Fig. 3 shows the influence of the connection typology. Specifically, Fig. 3(a) and Fig. 3(b) report the results respectively for double- and single-shear joints. Fatigue S - N curves have been determined according to the methodology described in Section 2, by using a slope coefficient equal to 5. The figures highlight the 95% fractile net stress ranges corresponding to 2 million cycles, which, according to Eurocode 3 – Part 1.9, is conventionally indicated as fatigue resistance. As it is possible to observe, the fatigue resistance for double-shear details is equal to 81 MPa, a value close to the Taras et al . 2010 proposal. Conversely, the fatigue resistance decreases for the single-shear connections, as a value of 55 MPa is observed. The fatigue strength reduction can be mainly attributed to secondary actions induced by load eccentricity in unsymmetrical connections on joint fatigue resistance. It is noteworthy that, despite the number of experimental data referring to single-shear joints being lower than double-shear connections, the investigated net stresses cover a sufficient range, allowing the definition of the S - N curve in the fatigue regime investigated.