PSI - Issue 78

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

650

To incorporate both stochastic and epistemic uncertainties related to key parameters defining each wall class, a Monte Carlo simulation was carried out. This probabilistic framework ensures the applicability of the results to large-scale risk assessments, especially in contexts with limited data on building stock and exposure. To characterize the OOP response of the wall elements, point clouds representing capacity at key performance points were generated for each class and load configuration. Given the extensive dataset produced, surrogate capacity models were formulated and calibrated to directly estimate wall capacity. Specifically, three second-order polynomial equations were derived − one for each performance point − tailored to each infill class and type of lateral load. 3.1. Taxonomy of structural masonry elements The non-structural category comprises infill panels, which are further subdivided based on geometric dimensions and their placement in either existing or newly constructed buildings. Additional distinctions are made according to the size and mechanical properties of the masonry units composing the panels (Boschi et al., n.d.; “Istruzioni per l’applicazione Dell’«Aggiornamento Delle ‘Norme Tecniche per Le Costruzioni’» Di Cui al Decreto Ministeriale 17 Gennaio 2018.,” 2019; “Laterizi comuni, mattoni e forati,” n.d. ). As an example, Tab. 1 provides geometric and mechanical properties (for the sub-class of infill panel made with hollow blocks with horizontal holes).

Table 1: Main characteristics of the infill panel selected class.

Value range of key parameters Wall geometry Masonry unit size Mechanical properties

Existing constructions

f v = (5.0÷8.0)/4MPa E v = (3500÷5600)/4MPa f h = 5.0÷8.0MPa E h = 3500÷5600MPa

h b = 0.24m l b = 0.30m t m = 0.01m

h p = 2.2÷3.3m l p = 2.5÷7.0m t p = 0.12m

Hollow blocks with horizontal holes

3.2. Monte Carlo simulations The analytical model described in the previous section was embedded within a Monte Carlo simulation framework to generate point clouds representing the OOP capacity of different wall/infill classes at defined performance points. These datasets provide the basis for constructing capacity models across varying configurations. The Monte Carlo approach is essential for capturing both aleatory and epistemic uncertainties associated with key model parameters, including: • geometric features (e.g., aspect ratio, slenderness, masonry unit dimensions) • mechanical properties (e.g., material strength, stiffness, stiffness of the external bond). The simulation relies on random sampling techniques to propagate uncertainties through the model and assess their impact on structural response. Each parameter is assigned a probability distribution reflecting its variability: geometric parameters were sampled from uniform distributions bounded by the ranges reported in Tab. 1, while mechanical properties followed lognormal distributions, with mean ( μ ) and standard deviation ( σ ) values derived in accordance with national technical guidelines (“Istruzioni per l’applicazione Dell’«Aggiornamento Delle ‘Norme Tecniche per Le Costruzioni’» Di Cui al Decreto Ministeriale 17 Gennaio 2018.” (2019)). By running a large number of iterations, the simulation produces statistically meaningful outcomes, enabling the development of surrogate models that approximate wall capacity under uncertain conditions. 3.3. Surrogate vulnerability models ’ coefficients of load-bearing walls class After generating the point clouds for each performance point − considering both rectangular and triangular