Probabilistic evaluation of the axis distance’s influence on the flexural strength deterioration of reinforced concrete beams under ISO 834 fire

The distance from the nearest concrete exposed surface to the centroidal axis of main longitudinal steel re-inforcing bars, so-called axis distance, plays a critical role in ensuring the safety of reinforced concrete (RC) structures under fire, as it helps the rebars not being directly exposed to heating in a fire incident. However, a large axis distance value could reduce the effective height as well as the beam’s flexural strength at ambient condition. In order to determine the appropriate values of axis distance, this article developes a data-driven method for predicting the flexural strength deterioration (FSD) of RC beams under ISO 834 standard fire based on the material and geometrical inputs. This method consists of two main stages: (i) Establishing a theoretical/experimental database by collecting experimental data from the literature; and (ii) Engineering a probabilis-tic model based on the Bayesian Neural Network. The results obtained show that the proposed approach is a practical tool that is capable of performing quick and reasonably accurate analysis such as degradation curves of FSD against exposure time. In addition, the uncertainty related to the prediction results is also evaluated, providing useful information for structural fire engineers to achieve conservative designs.