Propagation of RC Beam-Column Joint Modeling Uncertainty to the Seismic Performance of RC Buildings Using Incremental Record-Wise LHS

Abstract

Beam-column joints are critical elements for the structural integrity of Reinforced Concrete (RC) framed buildings under earthquake loading. Many aspects of joint modeling can have a significant impact on the estimate of collapse performance, while non-negligible parameter uncertainties are present in their strength, stiffness, deformation capacity, and cyclic deterioration. An accurate test-informed probabilistic beam-column joint model is of paramount importance as it can improve cost and time efficiency of new building designs and retrofit measures of existing buildings. Based on a recently collected cyclic-test database on 771 RC beam-column joints, probabilistic joint macro models have been proposed (Elmorsy 2020). Models were proposed for both ductile RC joints and seismically deficient (shear deficient) RC joints, without shear reinforcement, which is typical in existing buildings in the US that were built prior to the 1980s. On this basis, incremental record-wise Latin Hypercube Sampling (iLHS) (Vamvatsikos 2014; Kazantzi et al. 2014) accompanied by Multiple Stripe Analysis (MSA) is applied to quantify the resulting seismic performance uncertainty due to joint modeling uncertainty. Two frames are used as case studies; one designed according to modern codes (new) and the other is deliberately designed to contain joint shear deficiency (existing). Structural collapse risk sensitivity due to the RC beam-column joint modeling uncertainty is assessed for both new and existing (with joint shear deficiency) frames, quantifying its effect on structural performance.