Data‐driven probabilistic curvature capacity modeling of circular RC columns facilitating seismic fragility analyses of highway bridges

• DOI: 10.1002/eer2.14
• Publication Date: 2022-08-08T09:46:12Z
• Authors: Xiaowei Wang; Xinzhe Yuan; Ruiwei Feng; You Dong
• Keywords: Bridge (graph theory)

Abstract The availability of reliable probabilistic capacity models reinforced concrete (RC) columns is a cornerstone for high‐confidence seismic fragility and risk analyses highway bridges. Existing studies often perform physics‐based pushover or moment–curvature the modeling RC columns, which may encounter nonconvergent problems under high levels nonlinearities in structural material constitutive elements, become computationally inefficient especially when analysis model contains plenty cases involving multisource uncertainties. To mitigate issues as well release computational burden column estimates, this study explores potency artificial neural network data‐driven curvature circular can facilitate assessment end, large database developed by fiber‐section‐based covering major ranges steel strengths, reinforcement ratios, vertical loads, geometries engineering practices. obtain an accurate model, fivefold cross‐validation training test process performed to optimize architecture. optimized leads estimating multilevel indices with percentage errors less than 15%. Finally, typical bridge taken case demonstrate applicability expediency analysis. For ease implementation, associated codes are available at https://bit.ly/3A1dh1V .