|
Abstract
|
This study presents a sampling-based reliability sensitivity analysis to identify the random variables that are most influential on the seismic reliability of a corroded reinforced concrete (RC) bridge bent. To this end, a nonlinear finite element (FE) model of a scaled RC bent is developed, which accounts for the corrosion initiation and pitting corrosion models, as well impact of corrosion on the cover concrete, core concrete, reinforcing bars, and bond-slip behavior considering underlying uncertainties. Ground motion records are selected using conditional spectrum (CS) method. Pushover analysis is conducted on RC bent in discrete time points to evaluate the corrosion-dependent drift limit for different damage states. Seismic reliability analysis of the RC bent is conducted using Monte Carlo sampling (MCS) method in different damage states and time points. Results indicate the severe decrease in the seismic reliability index of the RC bent after 30 years of exposure to chloride, especially in the higher damage states. Finally, importance vectors are developed based on sensitivity analyses. The obtained results demonstrate that the importance of underlying random variables depends on the corrosion level and intended damage state. The obtained importance vectors show that the random variable related to the seismic input is the most influential variable on the seismic reliability of the corroded RC bent. It is also found that variables related to corrosion initiation time contribute the most to the seismic reliability of the RC bent in the first 15 years, while in the last 15 years the structural variables are the important ones.
|