PSI - Issue 82

Chiara Bedon et al. / Procedia Structural Integrity 82 (2026) 65–71 Chiara Bedon et al. / Structural Integrity Procedia 00 (2026) 000–000

67 3

level, in order to maximize the possible unfavourable effects due to the presence of partial random cracks, and provide some insights on scale effects. 2. Materials, specimens and experimental methods The typical specimen consisted of a 2-ply LG section, composed of two AN glass plates and a bonding polymeric interlayer. The size of each specimen was set in 50 mm width × 100 mm length. Variations included modification of nominal glass thickness (t g = t g,nom = 4 mm and 10 mm), interlayer thickness (t int = 0.76 mm and 1.52 mm) and type of interlayer material (EVALAM VISUAL and SG5000 - EVA and SG respectively in the following). The governing parameters were combined to reproduce different configurations of technical interest. Most importantly, one glass layer for each specimen was preliminary fractured by means of a steel hammer, to facilitate the analysis of partially fractured configurations. This strategy resulted in randomly cracked glass elements. The attention was given, in the preliminary stage, to the definition of a through-the-thickness crack. Different test configurations were included in the experimental program. In particular, the attention of the investigation in Stage 2 was focused on the following schemes: Two different displacement rates, 25 mm/min (V1) and 250 mm/min (V2), were used to explore the effect of loading rate, which is known to strongly affect the behaviour of the viscoelastic polymeric interlayers (Centelles et al., 2020; Chen et al., 2022; Elkilani et al., 2025). Quasi-static tensile tests were first carried out on dog-bone specimens of EVA and SG interlayers. Quasi-static, displacement-controlled, cyclic tests were successively performed on a total of 57 specimens in a simply supported configuration (beam-like setup), based on 3-point-bending tests. The preliminary fractured LG samples were supported by steel blocks with rounded corners, and the distance of supports resulted in 70 mm. The vertical load was applied at the mid-span section of each LG sample through a Shimadzu AGS-X universal testing machine equipped with a rounded loading device, which also recorded the force-displacement response of the tested specimens. Most importantly, the imposed vertical load for the imposed protocol (with 20 cycles) was estimated according to Fig. 2, in order to avoid the fracture of the second glass layer and assess the mechanical performance of the partially cracked LG specimens under repeated loads. For convenience, the experimental load amplitude was quantified based on the “Scheme M” (i.e., single glass layer), which represents the lowest bound for the mechanical characterization of the cracked LG member in Stage 2, rather than the “Scheme I” (both glass layers intact). • Scheme II (SII): broken glass layer at the top (i.e., compressive side of the LG specimen); • Scheme III (SIII): broken glass layer at the bottom (i.e., tensile side of the LG specimen).

Fig. 2. Schematization of the approach for the definition of the imposed quasi-static cyclic protocol in the 3-point-bending setup.