Experimental and Analytical Investigation of Seismic Performance and Retrofit Techniques for Non-Ductile RC Structural Walls

Abstract

Reinforced concrete (RC) walls are widely used as lateral-force-resisting systems. Despite the interest in retrofitting techniques, critical knowledge gaps remain regarding the reliability and validity of current retrofit design guidelines and procedures, particularly for barbell walls with single sided retrofit strategies, and barbell walls with inadequate lap splices. This dissertation is aimed at addressing this need through a combination of experimental and numerical investigations focused on non-ductile RC structural walls. Six non-ductile barbell experimental specimens were built and designed with varying retrofit strategies including single sided fiber reinforced polymer (FRP) overlay, single sided RC overlay, end wrapped bar buckling and lap splice confinement. Brittle shear deformation was precluded in the web through shear strengthening strategies, while bar buckling and lap splice slip were effectively prevented through the current ACI-440 design equations. The end wrapping strategies also had the unintended consequence of localizing strains to the base of their respective specimens, which resulted in a large number of fractured rebar. The experimental results were compared to numerical models through the nonlinear truss analogy, which was adapted for this work to incorporate lap splice slip and overlay considerations. The experimental results were well estimated by this numerical modeling approach, achieving within 6% of the peak strength of all experimentally obtained values, as well as accurately predicting the failure modes. The nonlinear truss model is a computationally efficient and user-friendly implementation of a nonlinear finite element strategy. A parametric study was undertaken on a set of 200 rectangular and barbell walls to assess the sensitivity of several parameters typical in older (pre-1970s) structural walls, as well as to compare two assessment methodologies: a set of strain-based criteria and the current ASCE-41 guidelines. The results of the parametric study indicate a reasonable ability of the strain-based methodology to consistently capture key performance metrics of the backbone curve while maintaining a similar statistical spread and trend to the performance metrics obtained through the ASCE-41 guidelines for most cases of flexurally-controlled walls. This indicates that the strain-based failure criteria can be an equally effective method used to capture seismic behavior of older RC walls and may even be preferred as an alternative assessment procedure, since it relies on a local assessment framework which can be uniquely adapted for any structure or retrofit configuration.