PSI - Issue 78

Riccardo Raimondo Milanesi et al. / Procedia Structural Integrity 78 (2026) 1374–1381

1379

contact with the top beam. Identical testing protocols were applied to specimens T2, T3 and T5; their results are summarized here solely for T4 as a representative case and further information is reported by Morandi et al. (2025). Figure 7 presents hysteretic loops of acceleration versus displacement at T4 panel’s centre, with positive direction toward the plastered side. For T4, the final 2.75 g run is omitted because marker free -fall exceeded the sensor range and does not reflect true wall motion. T4 exhibited a near- linear acceleration increase to 13 g at 2.30 g PFA, then surged to 19.7 g before collapse. Figure 7 ’s 3D deflections, interpolated from marker data, confirm this symmetric out-of-plane bending in both axes. Identical measurement strategies and profile analyses were conducted for all specimens, with T4 serving as the illustrative example.

Figure 5. Force-displacement curve of T1 specimen (left) and comparison of the envelope curves of the infill only contribution of batch 1 specimens (T1, T2, T3) (right) (Morandi et al., 2025).

Figure 6. Sequence of pictures of the collapse mechanism of T4.

Figure 7. Out-of-plane hysteretic cyclic acceleration-displacement curves at the centre of the T4 (left) and 3D deformed shape of T4 (right). (Morandi et al., 2025). Dynamic properties were identified using accelerometers mounted on one face of each infill panel (Figure 8). To capture out‑of‑plane response during in‑plane tests, controlled transverse impulses were applied before each drift increment, whereas out‑of‑plane tests employed low‑amplitude, broadband white‑noise excitations via the shaking table. Frequency content was extracted by FFT, and for white‑noise records an Operational Modal Analysis (Peeters et al., 2012) ensured robust frequency and damping estimate s even in the presence of high damping or noisy data.