ABSTRACT
This paper presents a methodology for life-cycle structural assessment and performance prediction of existing bridges and bridge networks based on Artificial Neural Networks (ANNs). Multilayer ANNs are trained and applied to the prediction of the overall condition state of bridges taking as input the available information about the time-variant structural performance over limited time periods. The proposed approach is applied to life-cycle assessment of a reinforced concrete arch bridge under chloride-induced corrosion, as well as to the prioritization of maintenance and repair interventions of a group of bridges included in the California National Bridge Inventory.
