Experimental Investigation of Composite Coupling Beam-to-Wall Connections in Coupled C-PSW/CF Systems.

Four different composite coupling beam-to-composite plate shear wall (CPSW/CF) connection details were developed and proposed. Six large-scale specimens representing the four connections were designed, fabricated, and tested. The connections were subassemblies of coupled composite plate shear walls/concrete filled (CC-PSW/CF) subjected to cyclic lateral loading. The coupling beams were designed to be flexure-critical with clear span-to-depth (Lb/d) ratios of 3.5 or 5.1. This paper presents the experimental program, capacities, and detailed behavioral observations of all six specimens. The effects of connection type and Lb/d ratio on the ultimate strength, stiffness, ductility, and failure modes are evaluated. Major limit states and events included yielding of the steel plates comprising the coupling beam, followed by local inelastic buckling, fracture initiation in the base metal (near the weld toes), and fracture propagation through the beam flange and web plates, leading to loss of flexural strength and failure. All the connections were able to develop and transfer the flexural capacity of the composite beam section. The composite coupling beams developed flexural capacities (10%–50%) greater than those calculated using the plastic stress distribution method. The underlying reasons for this overstrength are evaluated. The AISC flexural stiffness equation for filled composite sections could reasonably estimate the stiffness of the composite coupling beam sections. A fiber-based model of the cross-section was used to calculate the section moment–curvature response of the filled composite beam sections. The calculated flexural capacities were consistent with those calculated by the plastic stress distribution method but lower than the experimentally observed values. The flexural stiffness values were slightly higher than the experimental results. For more accurate comparisons with experimental results, a numerical model is needed to estimate the cyclic lateral load-deformation response while accounting for the effects of local buckling, low cycle fatigue, and fracture. [ABSTRACT FROM AUTHOR]