Chloride-induced corrosion introduces significant uncertainty in assessing the seismic resilience of reinforced concrete structures, particularly in coastal environments. Current probabilistic methods, often reliant on predefined probability density functions (PDFs) and vulnerability models, struggle to accurately reflect structural conditions or control computational costs. This study proposes a fully probabilistic algorithm that leverages Bayesian inference and structural inspection data to update critical deterioration parameters, enhancing model accuracy. The algorithm employs a non-parametric maximum likelihood method to estimate seismic resilience and its associated uncertainty without relying on vulnerability models. By integrating inspection data, the proposed approach addresses the limitation of relying on predefined PDFs, which may not accurately reflect the actual condition of the structure. Additionally, by employing efficient computational techniques, it may address the computational inefficiency of conventional methods that use vulnerability models. Demonstrated through a case study of a corroded reinforced concrete school building in Ganaveh, Iran, the method provided more conservative resilience estimates and reduced uncertainty by up to 39 % compared to non-updated approaches. The proposed framework, which highlights the importance of probabilistic methods in lifecycle management and seismic design, might be adaptable to various structural systems.
