Risk-based bridge management: optimization and inspection scheduling

Abstract

Bridge management based on reliability methods permits the selection of the set of repair projects that minimizes the total network risk in a planning horizon. This approach, previously proposed by the authors, uses the Markovian model for the decay of structural components, evaluates the overall reliability of the bridge as a system, and computes risk by taking into consideration the consequences of reduced function. This paper presents important additional aspects of risk-based bridge management. First, genetic algorithms are proposed as an alternative optimization strategy; small and large bridge networks are solved, guiding the solution towards minimum risk. Second, Markovian transition matrices for deterioration and post-inspection Bayesian updating are combined to schedule inspections so that the overall reliability of the bridge network is maximized for a given inspection budget. Results from a simulation study using a real bridge network are presented and compared with current inspection practice. Key words: bridge management, reliability, risk assessment, optimization, inspection scheduling.