PSI - Issue 78

Sara Mozzon et al. / Procedia Structural Integrity 78 (2026) 646–653

651

loading profiles − surrogate capacity models were developed and calibrated accordingly. These models, which also incorporate key features of the flood action such as load shape and height, may be interpreted as surrogate vulnerability models. Beyond summarizing the extensive numerical simulations, they allow for the direct estimation of wall capacity across different performance levels using simple second-degree polynomial expressions. The selection of a polynomial form reflects a balance between model simplicity and predictive reliability. A preliminary sensitivity analysis was performed to identify the most influential parameters derived from the Monte Carlo simulations. For infill panels, five variables were found to govern capacity: panel height ( h p [m] ), inverse slenderness ( t p /h p [-] ), aspect ratio ( l p /h p [-] ), relative impact height ( h w /h p [-] ), and mean masonry compressive strength ( f m [MPa] ). These parameters influence capacity either individually (through linear or quadratic relationships) or in combination. Individually, variables such as height, inverse slenderness, aspect ratio, mean strength, and floor load exhibited predominantly linear effects on capacity, while the relative impact height showed a clear nonlinear (quadratic) influence. Moreover, interactions between the impacted height and other parameters − like height, aspect ratio, or floor load − were adequately captured through linear approximations. Based on these relationships, the general structure of the surrogate model for load-bearing masonry walls is defined as follows: (1) Tab. 2 provides the calibration coefficients of the surrogate vulnerability model − for a rectangular-shaped load − for infill panels made of horizontally perforated hollow blocks.

Table 2: Coefficients of the surrogate vulnerability model of the selected infill class, for rectangular-shaped load.

Coef. values (rectangular-shaped load) First cracking

Coef. name

Variables

Moderate damage Maximum capacity

Intercept [-]

c 0 c 1 c 2 c 3 c 4 c 5 c 7 c 8 c 9 c 6

-1.81E+01 3.25E+00 4.17E+02 -1.88E+00 6.91E-02 6.31E+00 8.58E+00 -2.73E+00 -3.38E+02 1.41E+00

-3.24E+01 4.32E+00 6.88E+02 -5.23E-01 9.80E-01 9.68E+00 1.37E+01 -2.97E+00 -5.10E+02 -5.13E-01

-4.66E+01 5.39E+00 9.59E+02 8.33E-01 1.89E+00 1.31E+01 1.89E+01 -3.21E+00 -6.81E+02 -2.44E+00

h p [m] t p /h p [-] l p /h p [-]

f m [MPa] h w /h p [-]

(h w /h p )

2 [-]

h p (h w /h p ) [m] (t p /h p ) (h w /h p ) [-] (l p /h p ) (h w /h p ) [-]

4. Proposal of fragility models Once the capacity values were established, simplified demand models − hydrostatic and hydrodynamic − were derived based on literature sources. By comparing these capacities with incrementally increasing demands, calculated according to established models, a procedure analogous to Incremental Dynamic Analysis (IDA) was applied to develop fragility functions. These functions were generated for different fluid densities and impacted wall ratios ( h w /h p ). The fragility curves presented in this section refer to 12 cm-thick infill panels composed of hollow blocks with horizontal perforations, subjected to a rectangular load profile. As such, they are conditioned on flow velocity, adopted as the intensity measure. Fig. 4 displays the fragility curves for all damage states (DS) under water flow conditions (density of 1000 kg/m 3 ), with results shown separately for five impacted wall ratios. Tab. 3 reports the corresponding lognormal parameters − mean and standard deviation of velocity − for each damage state and h w /h p ratio.