Discussions and Closures

The writer presents a parametric study devoted to investigate the behavior of nonprismatic members having T-sections by using three-dimensional ~3D! finite-element models ~FEMs!. Fixed-end moments and stiffness coefficients were computed from the 3DFEMs considering thrust effects. As a result of the parametric study, the writer proposed a practical modeling technique based on two-noded beam elements with an effective length concept to use in frame analysis. In the model proposed by the writer, the effective length of the two-dimensional ~2D! beam element representing the haunch evolves as 75% of its actual length. The writer is well known for his previous studies on haunched beams using elastic finite elements and is to be complimented for the interesting study he presented. In particular, his discussion on the general behavior of nonprismatic members is very insightful for most readers, as he notes that the discontinuity in the location of the centroidal axis on nonprismatic members favors a nonlinear stress distribution over the cross section, and coupling attributable to the arching action. The writer also notes that, for T-sections, the shifting in the location of the centroidal axis produces a relatively small arch depth that results in smaller axial forces than those of