PSI - Issue 78

Nicola Di Battista et al. / Procedia Structural Integrity 78 (2026) 412–417

414

Fig. 1. Dataset description. (a): PGA of municipalities under investigation assumed as intensity measure of L’Aquila 2009 earthquake; (b): AeDES building usability pie chart for classes A, B, C, and E; (c): buildings simplified global vulnerability index count bar chart; (d): post-event granted reconstruction costs count bar chart, expressed in euros per square meter.

orV 3 (high). The vulnerability scores for masonry construction range as follows (in parentheses: low / medium / high contribution): Summing the scores yields the global index shown in Fig. 1 (c). Four reference amounts (700, 1 000, 1 100, 1 270 € / m 2 ) serve as anchors; the choice depends on the AeDES usability class, EMS-98 damage grade, and the computed vulnerability tier. That base cost may then be adjusted upward or downward to reflect, for example, (i) cultural-heritage constraints, (ii) atypical engineering di ffi culties, or (iii) special architectural finishes [25]. Nearly 70 % of the inspected buildings were labelled unusable (class E), yet Fig. 1 (d) shows that authorised grants span a wide spectrum: even class A structures, if burdened by extensive non-structural damage, can receive sizeable funding. The rich mix of usability ratings, vulnerability metrics, PGA values, and financial outcomes makes the dataset a fertile test bed for developing machine-learning models aimed at predicting reconstruction costs with greater fidelity than rule-based approaches.