Robustness Assessment of RC Frame Buildings under Column Loss Scenarios

Abstract

A computational assessment of the robustness of reinforced concrete (RC) building structures under column loss scenarios is presented. A reduced-order modeling approach is presented for three-dimensional RC framing systems, including the floor slab, and comparisons with high-fidelity finite element model results are presented to verify the approach. Pushdown analyses of prototype buildings under column loss scenarios are performed using reduced numerical models, and an energy-based procedure is employed to account for the dynamic effects associated with sudden column loss. The load-displacement curve obtained using the energy-based approach is found to be in good agreement with results from direct dynamic analysis of sudden column loss. A measure of structural robustness is defined by normalizing the ultimate capacity under sudden column loss by the applicable service-level gravity loading. The procedure is applied to two prototype 10-story RC buildings, one employing intermediate moment frames (IMFs) and the other employing special moment frames (SMFs), and the SMF building, with its more stringent seismic design and detailing, is found to have greater robustness. INTRODUCTION Although computational and experimental studies on collapse resistance of RC beamcolumn subassemblies or planar frames were reported by researchers in recent years (Bao et al. 2008, Yi et al. 2008, Bao et al. 2012, and Lew et al. 2011), limited studies have been done on RC floor systems or realistic buildings. Previous studies of steel frame buildings have found that the floor slab contributes significantly to the collapse resistance of structures (Alashker et al. 2010). Experimental investigations on floor systems or realistic building structures are usually costly and [unrepeatable]. Numerical simulation provides an alternative approach. However, the challenge for numerical investigation is to develop [reliable model] that can be used in the analyses of large-scale structures with [affordable computational costs]. In this study, [experimentally calibrated models] for [beam-column frames] are used to develop reduced models for floor systems. The developed models are verified by comparing analysis results with high-fidelity finite element models. Using the [developed] reduced models, a collapse assessment approach is proposed and [descripted in great details]. An energy-based approximate procedure for analysis of sudden column loss, previously proposed by Powell (2003) and Izzuddin et al. (2008), is also considered and verified computationally, which enables the structural capacity under sudden column loss to be evaluated using the results of a single pushdown analysis. A metric for structural robustness is defined by normalizing the ultimate capacity under sudden column loss by the applicable service-level gravity loading. Two 10-story prototype buildings, which were designed for different seismic design categories, are evaluated for the potential loss of a first story column based on the proposed assessment approach. One building was designed for Seismic Design Category C (SDC C) and employs intermediate moment frames (IMFs), and the other was designed for Seismic Design Category D (SDC D) and employs special moment frames (SMFs). Full-scale beam-column assemblies from the prototype buildings have been tested to characterize the beam-to-column joint behavior (Lew et al. 2013) and to provide experimental data for validation of detailed and reduced numerical models (Bao et al. 2012). The results of the robustness assessment procedure show that the SMF building, with its more stringent seismic design and detailing, has greater robustness. FLOOR SYSTEM MODELING Two finite element models were developed to study the response characteristics of a two-bay by two-day prototype floor system. One is a detailed model with a total number of about 217,000 elements, including beam elements representing reinforcing bars and solid elements representing concrete. The other is a reduced model which consists of around 1700 shell elements representing slabs and 230 elements for beams. The plan view of the prototype floor system is shown in Fig. 1 and its reinforcement details are listed in Table 1. A bilinear stress-strain relationship is assumed for reinforcing bars with yield strength of 400 MPa and ultimate strength of 520 MPa. The corresponding plastic fracture strain is 15.6 %. Concrete compressive strength is assumed to be 25 MPa.