PSI - Issue 78

Carpanese Pietro et al. / Procedia Structural Integrity 78 (2026) 536–543

541

of the losses associated with each damage state. In PERIL, damage-loss matrices are derived as the 75th percentile from the distributions proposed by Di Ludovico et al. (2022) for masonry and reinforced concrete residential buildings, while for industrial buildings the model proposed by Gurpinar et al. (1978) was used as default. The economic loss thus calculated refers only to direct losses related to the building damage, and doesn’t take into account the damage suffered by contents. The content loss component is calculated similarly to the structural one: the probability of exceedance of each damage state and the percentage of value associated with them are the same as those calculated for the building, but the reference reconstruction cost is the value of the contents, defined as 150% of the value of the building (FEMA 2022a), i.e., 1275 €/m 2 . The economic loss is expressed by the platform in different ways: absolute terms [€], loss per square meter of the building [€/m 2 ], and percentage of loss to the total value of the building [%]. Another way to assess the impact of an earthquake on the building is to determine the probability that the damage suffered by the building will compromise its usability. The percentage of each category of usability (usable, unusable for a short period or partially usable, unusable) is calculated by summing the rates of usability/unusability associated with each damage state. In PERIL, the damage-usability matrices proposed by Di Ludovico et al. (2022) are selected for residential buildings. In the case of industrial buildings, the focus is usually not directly on the usability conditions of the building, but more on the business interruption (downtime) to which the production activities may be subjected. To assess the business interruption, the model proposed by Hazus (FEMA 2022b) was adopted, which suggests a recovery time for each damage state, i.e., the period (in terms of days from the date of the earthquake) required for the business to return fully operational. Again, the default usability and the business interruption models are customizable by the user. The possible flood analyses performed by the PERIL platform are the following. • Conditional analysis: the analysis is carried out for the three flood scenarios provided by ISPRA, corresponding to the three return periods of 30 years (HPH), 150 years (MPH), and 400 years (LPH). The map representing the extension of the floodable area is used to determine whether a building is subject to flood risk, while the characteristics of the event, required to perform the analysis, are provided by the corresponding maps of water height and velocity. • Unconditional analysis: as for the seismic case, this analysis involves the combination of multiple flood events and consists of the convolution of all possible seismic scenarios that may occur (for all values of TR) considering their probability of occurrence in an observation time window (1 year, 10 years or 50 years). The model used by PERIL for flood risk calculation is the INSYDE multivariable model proposed by Dottori et al. (2016). The model was originally developed for residential buildings, but it has also been adapted to industrial buildings. INSYDE is based on two types of input parameters: 6 parameters related to the flood event with which the hazard component of the model is defined and 17 parameters related to the building that correspond to its main exposure components to the flood event. Due to the large number of parameters considered, the model allows for very detailed analyses, but it is also easily adaptable to large-scale assessments through some simplification and the assumption of default values (Dottori et al. 2016, Molinari and Scorzini 2017, Galliani et al. 2022). INSYDE analyzes the possible damage components of the building, defining for each of them the extent and expected percentage of damage, through formulas or deterministic and probabilistic damage functions. The damage components are 33 in total, and can be grouped into 7 categories: clean-up, removal of damaged elements, replacement of non-structural components, repair of structural components, restoration of finishing, replacement of doors and windows, replacement of damaged building system. The calculation of the percentage of expected damage is the same for residential and industrial buildings, while the extent of such damage may differ, as these two building types are very different, especially in terms of finishing elements. Once the probability of damage and its extent is calculated for each component, the model applies the respective unit costs to obtain the economic loss of each component. The economic loss related to each component is then calculated by multiplying the respective extent and percentage of expected damage by the unit cost just defined. 4.2. Flood risk analysis