Expert judgment in life-cycle degradation and maintenance modelling for steel bridges

Markov-based models for predicting deterioration for civil infrastructures are widely recognized as suitable tools addressing this mechanism. The objective of this paper is to provide insights regarding a network of orthotropic steel bridges in terms of degradation. Consequently, a model combining a dynamic Bayesian network and a Markov chain is first introduced that builds up the network in a concise way. In an attempt to represent a network composed of two general classes of orthotropic steel bridges, the classical method of structured expert judgment is carried out as a quantification procedure. The first objective is to indirectly elicit transition probabilities for a Markov chain that probabilistically describes how each type bridge temporally deteriorates. Second, experts are asked to provide estimates on required conditional probabilities related to the Bayesian network. An in-depth analysis of the results is presented so that remarks and observations are subsequently pointed out and, finally conclusions are drawn.