Abstract
The development of a methodology to model network level bridge deterioration using a Bayesian Belief Network (BBN) is presented in this paper. BBN's are capable of handling complex relationships between elements (e.g. beams, columns, etc.) and the system (e.g. stock of bridges) by means of conditional probabilities specified on a fixed model structure. The advantages and limitations of BBN's for such applications are discussed. The application of the methodology is presented through a case study on a group of UK railway masonry arch bridges. The condition of elements within a selected sample of bridges is used as input in the BBN, together with a set of conditional probabilities based on inspection experience, to yield, in probabilistic terms, the overall condition of the bridge group. Sensitivity of various input parameters, as well as underlying assumptions, on group performance is investigated, which can help with the prioritization of assessment and maintenance intervention activities. © 2010 Taylor & Francis Group, London.