Comparing Testing Methods of Partially Composite Sandwich Wall Panel Glass Fiber–Reinforced Polymer Connectors

Abstract

Precast concrete insulated sandwich wall panels have been commonly used as exterior load bearing or cladding wall panels on building structures. These panels are usually designed as composite, partially composite, or noncomposite, depending on the strength of the shear connectors connecting the two exterior concrete wythes. Various proprietary available shear connectors are used to achieve the required degree of composite action. The design information of the shear connectors is provided by the connector supplier based on testing results. Currently, there is no universally accepted method of testing composite shear connectors in practice or in the literature, however, connector suppliers usually use one of two testing methodologies recommended by the International Code Council–Evaluation Service, namely, AC320 or AC422, respectively. It is not known which test provides a better measurement of design values such as strength and stiffness, and there is no such comparison in the literature. In this study, 68 pushoff experiments were presented to evaluate the difference between these two testing methodologies. The testing program consisted of 20 single shear specimens and 48 double shear specimens. Three different commercially available shear connectors were used as well as variable insulation wythe thicknesses. The results were statistically analyzed. Large differences in the measured connectors’ shear strengths and stiffness were observed between the two testing methods, with AC422 providing consistently larger ultimate shear and secant stiffness, but considerably more variability. Large-scale flexural specimens from the literature were modeled using AC320 and AC422 as-measured results, and AC422 tests were found to predict deflections, cracking, and strength more accurately. It was concluded that AC422 produces a more accurate measure of connector properties, but suffers from large variability, resulting in a recommendation that further investigation into changing testing parameters and/or methods is warranted to reduce this variability.