Abstract

Recently, mesh-objective (i.e., mesh insensitive) fiber-based numerical models have been developed to simulate RC walls with compression-controlled softening failure modes (i.e., concrete crushing and rebar buckling). These models regularize the post-peak stress-strain behavior of concrete and steel materials based on experimentallycalibrated failure energy equations. While the ultimate displacements of RC walls predicted by these models have been accurate, the local curvature and material strain demands are very sensitive to the selected mesh. Consequently, subsequent normalization of the section curvatures based on an assumed plastic hinge length has been proposed in the literature, resulting in a two-step analysis process. This normalization step uses approximate equations that are dependent on the structural configuration and that can be challenging to implement in cyclic analysis. In accordance with these limitations, this paper studies the effect of material regularization in forcebased beam-column elements with a plastic hinge integration method to accurately simulate, in a one-step process, the softening behavior of slender planar RC walls under lateral loading. Numerical results are evaluated using available experimental measurements of eight slender flexural walls with compression-controlled failure modes from the literature. Previous research has shown that one of the biggest limitations of using conventional material models (i.e., without regularization of the stress-strain curves) is that the predicted ultimate displacements are significantly sensitive to the length of the critical integration point over which the nonlinear strains concentrate. This paper demonstrates that this sensitivity of the analysis results can be significantly reduced by using regularized material models and a plastic hinge integration method over a wide range of assumed plastic hinge lengths. The combination of material regularization and plastic hinge integration analysis can also predict the section curvatures and strains of RC walls, with no need for a separate normalization step, though these local responses are more sensitive to the assumed plastic hinge length.