Bayesian uncertainty assessment for mechanical properties of steel reinforcing bars in Colombia

The variability in the mechanical properties of reinforcing steel bars used in reinforced concrete elements affects the performance of the different structural systems. This variability may lead to struggling to comply with the requirements of local building codes. This study presents and discusses the uncertainty quantification of the mechanical parameters of the stress-strain curve of steel reinforcing bars in Colombia. The database comprised results from experimental programs, including monotonic axial tests on steel bars from different manufacturers in the country. A model was calibrated using the Bayesian estimation framework, and the Monte Carlo Markov Chain method. Bayesian statistical analysis procedures were used to update the Raynor model of the monotonic stress-strain curve of the steel. Distributions for parameters of the steel are then presented and predictive check tests are performed for validation. The Bayesian statistical analysis allowed contrasting the mechanical properties obtained experimentally with nominal properties specified by the manufacturers. The distributions and associated variability for the main mechanical parameters of the rebars are presented and grouped by manufacturer and by bar size.