A comparative study of models for shear strength of reinforced concrete T-beams. (January 2022)

• Record Type: Journal Article
• Authors: Liu, Jian; Gao, Ziyue
• Is Part Of: Structures. Volume 35(2022)
• Page Start: 922
• Page End: 930
• Publication Date: 2022-01
• Subjects: T-beams -- Shear strength -- Prediction models -- Statistical comparison -- Parametric study -- Finite element analysis; Structural engineering -- Periodicals; 624.1
• DOI: 10.1016/j.istruc.2021.11.044
• Languages: English
• ISSNs: 2352-0124

Abstract

Abstract: Reinforced concrete T-beams are widely used in bridges where beams are cast integrally with deck sections to either side at the beam tops. Although it has been recognized by numerous experimental studies that flanges may have a significant contribution to the shear strength of reinforced concrete T-beams, they are neglected in most of the design codes. In this paper, the load paths to transfer shear force in T-beams are firstly investigated with the help of existing experimental tests. Compared with rectangular beams, an additional load path diverts the diagonal compression from the loading point and the end support, and it carries a portion of the shear force through the flange to the end support. In order to find a relatively reliable model to predict the shear strength of T-beams, a comparative study is carried out among five representative models selected from the literature by using a database with 233 reported T-beam tests. Parametric studies with individual test series are also performed for a detailed evaluation of the five models. It is found that a model proposed by Cladera et al. (i.e., Model [3] is one of the models producing the least scattered predictions. The average shear strength experimental-to-predicted ratio Vexp / Vpred is 1.06 with the coefficient of variation (COV) of 19.6%. At the same time, it is shown that Model [3] exhibits uniform results across the entire range of experimental data without obvious bias. At the same time, Model [3] is one of the models requiring the least computational effort but with the largest range of applicability.