A Generalized Framework for Reliability Assessment of Bridges using Adjusted Partial Safety and Combination Factors

Abstract

Assessment of degrading infrastructure, particularly bridges, represents a growing challenge for structural engineers. While traditional strategies for bridge assessment, such as a code-based rating factor assessment, are still frequently utilized by practitioners, the advantages of a probabilistic approach are now well-recognized. However, the applicability of probabilistic assessments for large infrastructure is impeded by many factors including high dimensionality, high computational costs, and the difficulty of incorporating advanced structural analysis models for estimation of resistance. In this paper, the authors propose a generalized framework for reliability assessment of bridges using adjusted partial and combination factors. The framework is conceptualized using six indepent modules, operating in mutual interaction and cooperation: structural properties, linear elastic analysis, reliability analysis, Bayesian data assimilation, partial factor adjustment and calibration, and non-linear analysis module for a final performance check. The framework addresses the shortcomings of typical semi-probabilistic assessments by leveraging the strengths of each analysis module and enables the application of probabilistic assessment methods on large infrastructure. The main challenge for the practical applications of probabilistic assessments relates to uncertainty quantification and data collection, that would facilitate establishing the probabilistic models for resistance (including degradation effects) and load effect parameters well reflecting bridge-specific conditions.