PSI - Issue 78

Fadel Ramadan et al. / Procedia Structural Integrity 78 (2026) 859–866

864

a)

b)

Figure 4.2 – ITA18FAS spectra for both horizontal and vertical components under a strike-slip scenario with Vs30 = 300 m/s. Spectra are shown for Mw = 5.5 and varying distances (a: Rjb = 1 km, 20 km, 50 km), and for Rjb = 20 km with varying magnitudes (b: Mw = 4.5, 5.5, 7.0).

The standard deviation of both components, horizontal and vertical, and for the different terms, is comparable.

5. Application of VH model in 3D spectrum compatibility selection As a case-study example, the selection and spectral compatibility of real ground motion suites are evaluated for the city of L’Aquila (Italy), under a seismic scenario corresponding to a 475 year return period, with specific attention to three-dimensional (3D) compatibility, i.e., simultaneously matching both horizontal and vertical response spectra. Two different target spectra were considered: the NTC18 design spectrum, as prescribed by Italian seismic code, and a physics-based horizontal and vertical spectrum derived from the ITA18 ground motion model. All ground motion selections are performed using the RexelWEB platform (Sgobba et al., 2019; Iervolino et al., 2010), an open-access tool designed for the selection of spectrum-compatible accelerogram suites. RexelWEB is integrated with the Engineering Strong Motion (ESM) database (Luzi et al., 2020), enabling selection based on user-defined seismological and site parameters, as well as constraints on spectral compatibility and scaling. An initial record selection is carried out using the NTC18 design spectrum (Fig. 5.1, a). The selection considered rock site conditions (soil classes A and B), with classifications derived primarily from geophysical investigations, and supplemented by geological and topographic inference where required. Records from events with any focal mechanism are accepted. Selection parameters included magnitude range (M w or M l ): 5.7 – 7.5; epicentral distance: 0 – 30 km; horizontal spectral period range: 0.1 – 2.0 s, vertical spectral period range: 0 – 0.5 s; compatibility tolerance: – 10% / +30%; maximum scale factor applied to the mean spectrum (SF): 3.0. Despite the application of maximum scaling (SF = 3) and substantial relaxation of the vertical tolerance bounds (up to +90%), the resulting suite showed only partial success in reproducing the target spectra, particularly for the vertical component. The average horizontal spectral score was 0.062, indicating acceptable performance, whereas the vertical score increased to 0.205, clearly revealing a poor spectral fit at intermediate and long periods (see Figure 5.1, a). This result highlights a key limitation in using conventional code spectra, such as NTC18, for the selection of 3D spectrum-compatible records. Even when high tolerances and high scale factors are allowed, compatibility, especially for the vertical component, is difficult to achieve. This is due to the substantial shape mismatch between the NTC18 vertical target and the median response spectrum of pre-selected real ground motions at medium-to-long periods. While the scale factor adjusts for amplitude at short periods (e.g., PGA), it does not reconcile the spectral shape, thus

6