A practical framework for determining target reliability indices for the assessment of existing structures based on risk-informed decision-making

Abstract

Target reliability levels define structural safety requirements. Most current studies on target reliability indices (${\beta }_t$) have focused on reliability-based design for new structures. However, existing structures face significant safety challenges due to ongoing aging and financial constraints that limit maintenance and reinforcement efforts. Therefore, determining appropriate ${\beta }_t$ for the assessment of existing structures is crucial to balance the tradeoff between safety and economy. This study develops a practical, risk-informed framework to streamline the determination of ${\beta }_t$ for the reliability assessment of existing structures. It involves six critical steps including context definition, structural system modeling, failure statistics analysis, risk criteria establishment, and ${\beta }_t$ selection. The framework’s practical application is carefully demonstrated through a case study centered on the reliability assessment of existing medium- and small-span (MS) bridges in China. A database was compiled for failure statistics of MS bridges, documenting 241 bridge collapse incidents in China spanning from 1983 to 2024. The statistical analysis of lethality ratios and fatalities from these failure events is incorporated into individual risk criteria, group risk criteria, cost optimization, and the marginal lifesaving cost principle. Using these criteria, alongside a refined as low as reasonably practicable (ALARP) principle, informed decisions are made on selecting ${\beta }_t$ for reliability differentiation. Finally, three safety levels of ${\beta }_t$ are recommended for the bridge system as well as individual components. The proposed methodology framework, as demonstrated in the case study on MS bridges in China, can be readily applicable to the determination of ${\beta }_t$ for various other existing civil structures.