METAMODELING CHOICES FOR SEISMIC VULNERABILITY ASSESSMENT OF BRB-RETROFITTED LOW-DUCTILITY RC FRAMES

Abstract

Damage incurred in low-ductility reinforced concrete (RC) buildings during recent earthquakes continues to underline their structural vulnerability under seismic shaking. Among the viable seismic retrofitting procedures, passive control systems such as buckling-restrained braces (BRBs) have emerged as an efficient strategy for structural damage mitigation through stable energy dissipation while providing additional strength and stiffness to low-ductility buildings. Although quantifying the beneficial effects of BRBs for vulnerability reduction through seismic fragility curves has been suitably investigated in literature, almost all such studies consider a deterministic description of the BRB device. This study illustrates a metamodeling framework rooted in statistical learning techniques for efficient seismic vulnerability assessment of BRB-retrofitted low-ductility RC frames. The framework develops multidimensional probabilistic seismic demand models for response prediction of a case study retrofitted frames as a function of ground motion characteristics as well as the design parameters of the BRB device. These demand models when compared against damage states capacity estimates subsequently yields vector-based seismic fragility functions that provide notable advantages over unidimensional fragility curves in terms of efficiency as well as generality. Additionally, uncertainties stemming from a multitude of sources can also be conveniently captured and propagated through the different stages of statistical model development. The proposed study aims to help researchers, stakeholders, and even device manufactures by providing a conven-ient tool for vulnerability evaluation of retrofitted structures with reasonable accuracy and enhanced efficiency of computation. Concrete Frames, Seismic Retrofit, Seismic Devices Uncertainty, Metamodeling.