Bayesian updating of the residual structural performance of existing concrete structures based on experimental tests

Reliability analysis of existing concrete structures requires performance estimations which are affected by uncertainties related to system geometry, material characteristics, loading scenarios, and exposure conditions. A framework for Bayesian model updating with structural reliability methods of concrete structures is presented and validated based on the results of an experimental campaign on prestressed concrete (PC) bridge deck beams extracted from a dismantled 50-year-old viaduct. Subset simulation techniques are applied for structural reliability analysis. The prior knowledge of concrete mechanical properties and residual prestressing levels is updated based on the results of laboratory tests. The statistical residual performance of the PC deck beams is compared with the experimental results of a full-scale load test up to collapse. The structural modelling is developed with bi-dimensional finite elements for plane-stress analysis formulated in accordance with the Modified Compression Field Theory (MCFT). The comparison between experimental and numerical results allows to validate the modelling strategies, the outcomes of the experimental tests, and the Bayesian updating framework.