PSI - Issue 78

Merani Margherita Gabriella Bruna et al. / Procedia Structural Integrity 78 (2026) 785–792

787

2.1. The baseline predictive scenario

Immediately following a seismic event, a predictive damage scenario is generated through a structured workflow. First, a seismic hazard shakemap is created using GMPEs, which are adjusted with site amplification factors to account for local geology. Next, the building stock's exposure and vulnerability are defined; exposure is determined from census data, while vulnerability is modeled by assigning fragility curves to different building types. Finally, these components are convolved: the predicted shaking intensity is applied to the fragility curves to produce a probabilistic map with quantitative estimates of expected damage, providing a crucial "first picture" for emergency managers. 2.2. Refining ground shaking maps to enhance the scenario using dynamic monitoring data and site-specific hazard While essential for first response, initial shakemaps are estimations based on generalized models (GMPEs) and simplified site amplification factors. These models average past earthquakes and cannot capture an event's unique characteristics, while generic site factors oversimplify the complex mosaic of urban geology. Consequently, initial maps can be inaccurate, either by systematically over- or under-predicting shaking or, more critically, by missing localized pockets of severe damage. Integrating real-time monitoring data is transformative. Data from sensors at the base of buildings provide direct "ground truth" measurements of the motion, including key parameters like PGA, Peak Ground Velocity (PGV), and Spectral Accelerations (Sa). These direct measurements are invaluable for correcting and refining the initial shakemap. Furthermore, the building's own dynamic response can refine the hazard assessment locally. Standard site amplification assumes a building's linear- elastic period (T₁), but strong shaking causes non -linear behavior, elongating this period. By pre-calculating the relationship between shaking intensity PGA and this elongated period using Non-Linear Dynamic Analyses (NLDA), we can use the real-time PGA from the shakemap to instantly estimate the new period (T'₁). This updated period yields a more precise site amplification factor and a more accurate estimate of the Spectral Acceleration the structure experienced. To assess a building's condition, we first characterize its dynamic properties (natural frequencies, mode shapes) in an undamaged state using ambient noise data. During an earthquake, the same sensors record the building's response. Comparing post-event dynamic properties to the initial baseline allows for a direct, quantitative damage assessment. A significant decrease in the fundamental frequency (increase in period, T₁) is a robust indicator of structural damage, shifting the assessment from a forecast to an evidence-based evaluation. However, the knowledge of the building's fundamental period (T₁) unlocks a second, powerful advantage for refining the entire urban damage scenario. Traditional damage scenarios often rely on fragility curves that correlate damage states with a generic ground motion parameter like PGA. While practical, PGA is a poor indicator of a structure’s specific response, as it does not account for how the building’s dynamic characteristics interact with the frequency content of the earthquake. The integration of monitoring data overcomes this limitation. By identifying the fundamental period T₁ for an instrumented building (which can be considered representative of its typological class), we can select or develop fragility curves that are expressed as a function of Spectral Acceleration at that period, Sa(T₁). These spectral -based fragility curves are physically more robust because they link structural damage directly to the seismic demand at the period where the structure is most sensitive. This provides a dual benefit: • more accurate hazard: as described in Section 2.2, T₁ is used to calculate a more precise, structure -specific site amplification factor and thus a better estimate of the shaking demand (Sa(T₁)). • more accurate vulnerability: T₁ is simultaneously used to select the appropriate fragility curve, ensuring that the vulnerability model is consistent with the hazard parameter being used. This synergy between hazard and vulnerability refinement is a core strength of the proposed framework. The 2.3. Direct damage assessment and period-specific vulnerability refinement