PSI - Issue 78

Ahmed Mabrouk et al. / Procedia Structural Integrity 78 (2026) 960–967

961

Keywords: Clay masonry infill walls; Weak infill walls; Strengthened infill walls; Combined in-plane/out-of-plane behaviour; Experimental calibration; Infilled RC frames

1. Introduction Clay masonry infill walls are widely used in reinforced concrete (RC) frame construction across many countries, including seismically active regions such as Italy. The adoption of infilled RC frames became increasingly common after World War II, with a notable rise in construction frequency after 1961, particularly after 1981, due to ease of installation and their thermal and acoustic benefits ("Istat (Italian National Institute of Statistics)," 2011; INSYSME, 2014). Despite their widespread use, infill walls are typically classified as non-structural elements and are often neglected in the design process (Di Ludovico et al., 2016; Dolce & Goretti, 2015). This oversight has contributed to the seismic vulnerability of many buildings, as infill panels exhibit brittle behaviour and are prone to significant damage during earthquakes — even under low-to-moderate seismic intensities — due to their coupled in-plane (IP) and out-of-plane (OOP) responses. Over the decades, research and post-earthquake observations have consistently highlighted the critical role of infill walls in structural response (Masi et al., 2017; Verderame et al., 2014). Thin clay masonry infills, in particular, are vulnerable to OOP failure triggered by prior IP damage (Furtado et al., 2021; Ricci et al., 2018). In response, numerous studies have explored strengthening techniques — such as FRP, TRM, and steel based retrofits — to enhance OOP resistance (Silva et al., 2016). Nevertheless, the degradation caused by prior IP deformation remains a key challenge, and current codes lack comprehensive guidance for modelling IP/OOP interaction. To address this gap, recent advancements have extended macro-models to account for IP/OOP coupling through incorporating OOP mass, non-linear degradation laws, and simplified truss/beam models that require stepwise updating of constitutive properties to simulate interaction effects. These models strike a balance between computational efficiency and realism, making them suitable for structural-scale simulations. In this context, the present study contributes to ongoing efforts by implementing and recalibrating the fiber-based macro-model proposed by Donà et al. (2021) within the STKO – OpenSees framework. The model is calibrated using experimental results on thin clay masonry panels, including both unreinforced and strengthened panels employing TRM and FRM systems. To evaluate the structural implications of infill behaviour, nonlinear Time-History (TH) analyses are carried out on a representative RC frame in three configurations: (1) a bare frame, (2) a frame with unreinforced infills, and (3) a frame with strengthened infills. The focus is placed on thin masonry panels ( ≤ 20cm thick), which remain prevalent in the European building stock despite a gradual shift toward more robust façade systems. 2. Modelling approach of infill wall and RC frame The masonry infill walls are modelled using the macro-model proposed by Donà et al. (2021), designed to simulate the coupled IP and OOP response of infill panels within RC frames. This model adopts a multi-strut formulation, with each panel represented by four diagonal struts — two per diagonal — connected to the RC frame through pinned ends. Each strut is modelled using two Beam With Hinges (BWH) elements, allowing a central nonlinear hinge zone where damage is concentrated. These hinges employ fiber sections aligned in the OOP direction, enabling the representation of combined IP – OOP interaction. The fibers are assigned a compression-only trilinear hysteretic material, consistent with the brittle behaviour of clay masonry, and allow simulation of strength degradation, stiffness loss, and residual capacity under cyclic and bidirectional demands. The inertial mass of each panel in the OOP direction is lumped at the mid-node of the struts, and EqualDOF constraints are used to synchronize their motion, ensuring a realistic representation of panel dynamics. The geometric positioning of struts within the panel is derived from the estimated contact length between the infill and frame, based on stiffness ratios between the columns and masonry (Donà et al., 2021). Analytical expressions are used to define the elastic properties of the struts, including the equivalent axial area , the moment of inertia , and the Young’s modulus in the diagonal direction . These parameters are initially estimated based on mechanics-based assumptions and later refined through calibration against experimental results. A schematic of the macro-model configuration is shown in Fig. 1. The RC frame is modelled using force-based beam column (FB-BC) elements with nonlinear fiber sections. This formulation captures the spread of plasticity and is particularly suited for simulating moment – curvature behaviour under seismic loading. Each member is subdivided to reflect the actual detailing of confined and unconfined regions, especially near beam – column joints. Concrete fibers are assigned the Concrete02 material model, which accounts for strength degradation, tension stiffening, and different behaviours in confined versus unconfined zones (Kent & Park, 1971; Mander et al., 1988). Reinforcement is modelled