Abstract Extensive experimental studies have been conducted to investigate the flexural cracking behavior of reinforced concrete (RC) members (e.g., beams and slabs), while relevant finite element (FE) modeling and analyses of the crack width and crack spacing of them are quite limited. In this study, a two‐dimensional FE modeling approach for predicting the crack width and crack spacing of RC members was developed based on the “smeared crack approach,” and then in‐depth FE analyses were conducted. The developed FE modeling approach was verified against an experimental database consisting of 28 RC beams/slabs under four‐point bending reported by three independent research groups. The FE analyses show that the perfect bond assumption leads to mesh‐dependent calculated crack widths, while by modeling the bond–slip behavior between steel bars and concrete using cohesive elements, mesh convergence can be achieved. The reason why the stirrups in the pure bending zone serve as the initiators of flexural cracks is evidently explained via FE analyses for the first time. In addition, the performance of the bond–slip model in Model Code 90 (MC90) on predicting the bond–slip behavior between steel bars and concrete in flexural RC members was carefully assessed. The assessment shows that the MC90 bond–slip model tends to underestimate the bond stress between steel bars and concrete in flexural RC beams and slabs at the serviceability limit state, and thus leads to overestimation of mean crack width and crack spacing, indicating that in future studies, more accurate bond–slip models may need to be developed for the FE modeling.
Huang et al. (Mon,) studied this question.