A computational model for probabilistic analysis using the pseudo-dynamic method: a case study of the effects of seismicity on the structural behavior of the Jucazinho Dam in Pernambuco
Abstract The study of the behavior of gravity dams under seismic actions has, over time, been little explored in engineering projects in Brazil. In Brazil, although the occurrence of large magnitude earthquakes is uncommon, seismicity is not zero, especially in passive margin areas with intraplate seismicity. This study presents a method for evaluating the probability of seismic fragility of a dam's structural system. The method is based on computational simulation procedures involving pseudo-dynamic analyses and kriging modeling to establish the probabilistic characterization of the structure's safety factors, taking into account uncertainties in parameters related both to material properties and loading conditions, as well as ground motion characteristics. Stress analysis revealed critical concentration near the gallery, with maximum tensile stresses of 2.69 MPa and compressive stresses of -2.39 MPa. In stability assessment, all sliding safety factors exceeded the minimum design threshold (1.0), with the most critical case observed in the intermediate section under full reservoir conditions (1.05). Probabilistic analysis indicated failure probabilities ranging from 10−3 to 3.74 × 10−2 and reliability indices between 1.60 and 3.09, identifying the dam-foundation section as the most vulnerable and the intermediate section as the most resilient. The results highlight that foundation stiffness plays a decisive role in dam response, amplifying dynamic effects through multidomain interaction. Overall, the integration of numerical modeling with probabilistic methods proved effective for identifying critical sections, quantifying uncertainties, and supporting monitoring and reinforcement strategies. This approach strengthens dam safety management, contributing to failure prevention and mitigation of potential socio-economic impacts.