Sources of uncertainties for total chloride profile measurements in concrete: quantization and impact on probability assessment of corrosion initiation
Résumé
Reliability methods have proved in the past that they were rational aid-tools for the safety assessment of existing structures, within which some uncertainties occurred. Condition assessment is usually carried out using on-site measurements, which are assumed perfect. However, it is now accepted that some significant uncertainties may affect the assessment of material properties using semi-destructive methods. The purpose of this paper is to present a method for the identification and evaluation of measurement uncertainties using a bias and a zero mean error modeled by a random variable. These uncertainties obtained are then modeled using a probabilistic model. In a marine environment, the main cause of reinforced concrete structure degradation is the corrosion due to chloride ingress. The chloride profiles are determined using a destructive method involving many steps where the experimenter plays a key role. In order to identify sources of errors, four researchers have performed repeatability tests. The total chloride content is expected to be the same for all the samples. The heterogeneity has been studied using statistical analysis. A value of the bias is provided and the model results are consistent with the original results. Finally, the impact of measurement errors on reliability and life-cycle assessment is discussed.
Fichier principal
BonnetSchoefsSalta_EJECE_6avril2017-fs_sb.pdf (600.87 Ko)
Télécharger le fichier
BonnetSchoefsSalta_EJECE_hal.pdf (599.13 Ko)
Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)