Random field-based reliability updating framework for existing RC structures incorporating the effect of spatial steel corrosion distribution

In recent years, interest in the effect of the spatial variability associated with steel corrosion on the safety and reliability of reinforced concrete (RC) structures in harsh environments has increased. However, updating the reliability based on observations of existing RC structures remains a challenge. In this paper, an advanced framework for updating the time-dependent reliability incorporating the effect of the spatial steel corrosion distribution in existing RC members is presented. Probabilistic concepts are adopted to model random variables associated with material corrosion as random fields. With the application of a particle filter, spatially observed information obtained at discrete locations on RC members can be utilized to update the associated random variables. This reduces the epistemic uncertainty associated with structural performance assessments. In an illustrative example, time-dependent reliability updating is conducted on a RC bridge girder with the inspection results of the steel corrosion crack width. The effects of the number and space interval of observed locations on the updated structural reliability are investigated. The results demonstrate that the accuracy of the estimation of the time-dependent reliability of RC structures can be improved by using spatial observation information with a greater number of observed locations.