Integrating real‐time hybrid simulation with multi‐fidelity Monte Carlo predictor for seismic fragility assessment

Abstract

Seismic fragility evaluates the performance of a structure regarding to its capability to withstand earthquake loads by describing the exceedance probability for selected engineering demand parameters. High‐fidelity (HF) models are essential for simulation‐based techniques to obtain accurate fragility assessment in the presence of uncertainties. Dynamic structural behavior however is often difficult to replicate realistically through numerical simulation, especially when parts of the structure are difficult for accurate modeling due to their inherent complex nonlinearity. This study presents the integration of real‐time hybrid simulation (RTHS) with multi‐fidelity Monte Carlo (MFMC) predictor for seismic fragility analysis. RTHS combines physical testing of experimental substructures and numerical modeling of analytical substructures thus presenting a cyber‐physical alternative for HF seismic performance evaluation. The MFMC predictor provides an optimized allocation algorithm that makes full use of data from both HF and low‐fidelity (LF) models to achieve unbiased estimation for response quantities of interests. Experimentally expensive HF data from RTHS is thus integrated with computationally cheap simulation of LF numerical models through MFMC to facilitate seismic fragility analysis. A two‐story steel moment resisting frame (MRF) with self‐centering viscous dampers (SC‐VDs) is utilized to demonstrate the effectiveness and efficiency of the proposed method. Computational simulation is first performed to numerically verify the feasibility of the proposed method for the MRF subjected to non‐stationary stochastic ground motions. RTHS is then conducted experimentally as HF model and integrated with linear elastic LF models through MFMC for fragility analysis of the MRF equipped with SC‐VDs. Both computational and experimental analysis demonstrate that the proposed method provides a promising technique for seismic fragility assessment especially when accurate modeling is not available due to complex structural behavior.

A real‐time hybrid simulation framework for reliability‐based design optimization of structures subjected to pulse‐like ground motions

“…Multifidelity modeling is integrated through Co-Kriging surrogate to render accurate response prediction over the entire sample space of structural uncertainties. More recently, Gao et al 32 integrated RTHS with Multifidelity Monte Carlo simulation predictor for fragility analysis through RTHS tests in the laboratory (considered as HF models) and LF computational simulation of corresponding structure (considered as LF models).…”