Modeling error analysis of shear predicting models for RC beams

Abstract

Analyses of the modeling error associated with several commonly referred models for predicting shear capacity of reinforced concrete beams were carried out using a database containing results of test beams that was established based on test specimens found in the literature. The database contains a total of 1146 test beams, each has sufficient information for calculating the shear capacity using the shear predicting model in the design code suggested by the American Concrete Institute (ACI), the simplified method recommended by the Canadian Standard Association (CSA), the modified compression field theory, the shear friction method and Zsutty’s method. Statistics of the modeling error suggest that the performance of the shear predicting model in the ACI code and the CSA simplified method is similar but significantly biased and associated with large coefficients of variation of the modeling error. Overall, Zsutty’s method may be considered to be the best predicting model among the models included in this study. However, for beams with stirrups, use of the modified compression field theory provides less biased prediction with less scatter. This is followed by Zsutty’s method and by one of the shear friction models.