Prognostics and Maintenance Optimization in Bridge Management

Abstract

This paper has been written in collaboration with the Norwegian Public Roads Administration (NPRA). In Norway, bridges are a vital part of the transportation infrastructure. With more than 18,000 road bridges across the country, an efficient bridge management system is of critical importance to avoid high costs from over expending, to ensure safety of the public and availability of the transportation system. In the bridge management system applied by NPRA, the inspections are mainly carried out periodically based on pre-defined rules and the decision about when to perform the maintenance is based on the findings of these inspections. The objective of this paper is to propose a modelling framework that makes it possible to challenge these pre-defined rules by doing degradation prognostic and maintenance optimization. We propose to use a Piecewise Deterministic Markov Process to encompass different modelling assumptions as non-negligible maintenance delays and time dependent inspections. State probabilities and performance indicators are assessed through Monte Carlo simulations and a numerical scheme. The experimental values provided at the end show that optimal maintenance and optimization strategies should be investigated and further developed.