Article Dans Une Revue Engineering Structures Année : 2025

Bayesian updating of the seismic behavior of nuclear reinforced concrete structures: Methodology and Application

Résumé

The blind numerical assessment of the nonlinear behavior of structures under seismic loads, especially beyond design range, is a complex task given the strong aleatoric and epistemic uncertainties on properties and behavior of soil, reinforced concrete, and soil–structure-interaction. Hence, there might exist some gaps between the in-situ measurements and the blind simulation. To reduce such gaps, one can either improve the physical background of the model or, improve the knowledge of the model’s physical inputs to best fit the real response of the structure. This work aims at exploring a practical method to improve the knowledge, for a given verified and assumed valid model, of inputs based on available data and measurements that can be achieved at the structural scale. More precisely, this work explores the added value of metamodeling coupled to Bayesian updating techniques to improve our modeling and predictive analysis of reinforced concrete structures under seismic loads. For the sake of application, we consider here the SMART2013 mock-up representing at a scale 1:4 of an asymmetric auxiliary nuclear building subjected to seismic loads generated by a shaking table. Through this example, we demonstrate that the use of metamodeling and Bayesian updating techniques allows achieving better quantification of the median properties of concrete at the structural level and a clear reduction of epistemic uncertainties based on the experimental measurements. An application of the same approach at the full and industrial scale shall lead to the same observations as soon as reliable data are available in a sufficient amount.

Fichier non déposé

Dates et versions

irsn-04904919 , version 1 (21-01-2025)

Identifiants

Citer

Try Meng, David Bouhjiti, Benjamin Richard. Bayesian updating of the seismic behavior of nuclear reinforced concrete structures: Methodology and Application. Engineering Structures, 2025, 328, pp.119703. ⟨10.1016/j.engstruct.2025.119703⟩. ⟨irsn-04904919⟩
176 Consultations
0 Téléchargements

Altmetric

Partager

  • More