Probabilistic code-based shear capacity model for I-shaped steel reinforced concrete (SRC) beams

Abstract

A code-based probabilistic prediction model is developed for structural design and reliability analysis of the shear capacity of shear-critical steel reinforced concrete (SRC) beams. First, an experimental database of 83 shear tests was compiled from the literature, including steel-reinforced normal-strength concrete, steel-reinforced high-strength concrete, and steel reinforced lightweight aggregate concrete. Then, three deterministic formulations from the most commonly-used specifications JGJ, AIJ, and ACI-AISC were presented and compared against the database. Based on the different shear mechanism and formula considerations in the code-based deterministic models, this study developed 12 probabilistic models that account for both epistemic and aleatory uncertainty to examine the impact of concrete strength, shear span ratio, number of parameters, and their interactions. The Bayesian theory and Markov Chain Monte Carlo (MCMC) simulation method were utilized to calibrate the models by generating posterior distributions of the model parameters and the standard deviation (SD) of the model error. Optimal model selection hinges upon balancing precision and intricacy. An analysis was then conducted to evaluate the accuracy of the selected model, and the probabilistic model was applied to estimate the shear reliability of a SRC beam. Results demonstrate that the model objectively represents the probabilistic properties of the shear capacity of SRC beams and has the capability to conduct reliability analysis.